Biogenesis of Mitochondrial Ubiquino1:Cytochrome c Reductase (Cytochrome bcl Complex) PRECURSOR PROTEINS AND THEIR TRANSFER INTO MITOCHONDRIA*

The precursor proteins to the subunits of ubiquinol:cytochrome c reductase (cytochrome bc1 complex) of Neurospora crassa were synthesized in a reticulocyte lysate. These precursors were immunoprecipitated with antibodies prepared against the individual subunits and compared to the mature subunits immunoprecipitated or isolated from mitochondria. Most subunits were synthesized as precursors with larger apparent molecular weights (subunits I, 51,500 versus 50,000; subunit II, 47,500 versus 45,000; subunit IV (cytochrome c1), 38,000 versus 31,000; subunit V (Fe-S protein), 28,000 versus 25,000; subunit VII, 12,000 versus 11,500; subunit VIII, 11,600 versus 11,200). Subunit VI (14,000) was synthesized with the same apparent molecular weight. The post-translational transfer of subunits I, IV, V, and VII was studied in an in vitro system employing reticulocyte lysate and isolated mitochondria. The transfer and proteolytic processing of these precursors was found to be dependent on the mitochondrial membrane potential. In the transfer of cytochrome c1, the proteolytic processing appears to take place in two separate steps via an intermediate both in vivo and in vitro. In vivo, the intermediate form accumulated when cells were kept at 8 degrees C and was chased into mature cytochrome c1 at 25 degrees C. Both processing steps were energy-dependent.

In Neurospora and in all other organisms studied so far, the cytochrome b is synthesized on mitochondrial ribosomes, whereas all the other subunits of the cytochrome bcl complex are synthesized on cytoplasmic ribosomes (16)(17)(18)(19).
We have studied the biogenesis of the cytochrome bcl complex in Neurospora mitochondria. Precursors to the cytoplasmically translated subunits were synthesized in a heterologous cell-free system and immunoprecipitated using antibodies prepared against individual subunits. These were compared to the mature subunits isolated or immunoprecipitated from mitochondria. Most, but not all, of the cytochrome bcl complex proteins in Neurospora mitochondria appear to be synthesized as larger molecules with additional sequences which are cleaved during or after transfer into the mitochondria.
Four of the precursor proteins were transferred into isolated Neurospora mitochondria in an in vitro system. This transfer was post-translational, like that of the other mitochondrial proteins whose import has been studied (20)(21)(22)(23)(24)(25)(26)(27), and in contrast to the co-translational transfer that takes place in the endoplasmic reticulum (28,29). We investigated the energy dependence of precursor processing and import into mitochondria of these four cytochrome bcl complex proteins. In each case, a membrane potential appears to be required for transfer.
The transfer of cytochrome c1 is unique in that an intermediate form is observed during the conversion from precursor to mature protein both in vitro and in vivo.

EXPERIMENTAL PROCEDURES
Growth of Neurospora Cells-Neurospora crassa wild type 74A was cultivated i n Vogel's minimal medium (30), supplemented with 2% sucrose, at 25 "C for 15-17 h under vigorous aeration and illumination with fluorescent lights. When indicated, cells were labeled with 35S by growing them with a reduced concentration of sulfate (0.08 mM) in th4 presence of 0.5 mCi/liter of sodium [35S]sulfate  Ci/mol, Nirw England Nuclear), or cells were labeled with 3H by adding 1 mCi/liter of C3H]leucine (40-60 Ci/mmol, New England Nuclear) to the culture 1 h before harvesting. Cells were harvested by filtration.
Isolation of Mitochondria-Mitochondria were prepared from 35slabeled cells by grinding with sand. One g of cells (wet weight) was mixed with 1.5 g of sterile quartz sand (Riedel de Haen, Hannover) and 2 ml of SET medium (0.44 M sucrose, 0.002 M EDTA, 0.03 M Tris-HCl, pH 7.2) and ground in a mortar for 5 min at 4 "C. The resulting slurry was centrifuged twice for 5 min at 3,500 X g and the pellets were discarded. The supernatant was centrifuged for 12 rnin at 17,400 x g. The mitochondrial pellet was resuspended in SET medium and recentrifuged for 12 min at 17,400 X g. Unlabeled  prepared from Neurospora spheroplasts as described previously (31). Protein Synthesis in Reticulocyte Lysates-Reticulocyte lysates were prepared according to Hunt and Jackson (32) from rabbits which had been made anemic by injection of phenylhydrazine as described by Allen and Schweet (33). Cell-free protein synthesis was carried out according to Pelham and Jackson (34). One ml of incubation mixture contained 1 mg of Neurospora RNA (35) and 0.5 mCi of ["SS]methionine (New England Nuclear, specific activity 500 Ci/mmol). After the samples had been incubated for 30 min a t 30 "C, a postribosomal supernatant was prepared by centrifuging for 1 h a t 166,000 X g.
Transfer in Vitro of Cytochrome bc, Complex Subunits-Following incubation of a reticulocyte lysate with Neurospora RNA and [:' "SI methionine and preparation of the postribosomal supernatant, methionine and sucrose were added to the supernatant to final concentrations of 50 p~ and 0.2 M, respectively. Mitochondria isolated from spheroplasts were gently resuspended in this mixture (0.5-1 mg of mitochondrial protein/ml) and the suspension was incubated at 25 "C for 20-60 min with gentle shaking. The reaction was terminated by cooling to 0 "C and separating mitochondria from supernatant by centrifuging for 12 rnin at 17,400 X g. The mitochondrial pellet was Precursor was synthesized in a reticulocyte lysate, which was then incubated with Neurospora mitochondria for 1 h. The mitochondria were reisolated by centrifugation and lysed. Cytochrome CI was immunoprecipitated from both the supernatant and the lysed mitochondria, followed by SDS-gel electrophoresis and fluorography. washed once by resuspending in SET medium and recentrifuging. When indicated, proteinase K treatment was performed by adding 50 pg/ml of proteinase K (Boehringer Mannheim) to the mitochondrial pellets suspended in SET medium or to the supernatant. Samples were incubated for 30 min at 4 "C and then proteolysis was stopped by adding 0.5 pmol/ml of phenylmethylsulfonyl fluoride. Transfer in Vivo of Cytochrome cl-Neurospora hyphae were grown for 15 h at 25 "C as described above. The cell suspension (50-100 m l ) was then cooled to 8 "C and 5 pCi/ml of [3H]leucine (40-60 Ci/mmol, New England Nuclear) was added. After 6-12 min a t 8 "C, unlabeled leucine and cycloheximide were added to tinal concentrations of 0.1 mg/ml and 0.05 mg/ml, respectively. The cells were then incubated at 25 "C for another 10-15 min. During both the pulse and the chase periods, aliquots were removed and rapidly filtered at 0 "C. The hyphae were homogenized in 1 ml of boiling 3% SDS,' 1 mM ethylene glycol bis(P-aminoethyl ether)-N,N,N',N'-tetraacetic acid, 1 mM EDTA, pH 7.5 (36) using an Ultra-Turrax (IKA, Staufen, FRG). After boiling for 5 min, the samples were cooled to 4 "C, diluted with 10 volumes of Triton buffer (1% Triton, 0.3 M NaCl, 5 m EDTA, 10 m Tris, pH 7.5) containing 0.1 mM phenylmethylsulfonyl fluoride and 0.1 m o-phenanthroline, and centrifuged a t 39,000 X g for 15 min. Cytochrome cI was immunoprecipitated from the supernatant.
Preparation of Antibodies to Cytochrome bcl Complex Subunits-Ubiquino1:cytochrome c reductase was isolated as described elsewhere (7). The polypeptides of the enzyme were separated by preparative SDS-polyacrylamide gel electrophoresis. Approximately 6 mg of protein dissolved in 2 ml of 2 m dithioerythritol, 5 n m Tris/acetate, pH 7.0, 0.5% SDS was applied to a vertical 18% acrylamide gel (2 X 300 X 250 mm). The running buffer was 0.1% SDS, 50 mM Tris/acetate, solution containing 5 volumes of water, 1 volume of acetic acid. After destaining the gel for 24 h, the polypeptide bands were cut out of the gel and the proteins were recovered by electroelution of the gel slices.
Antibodies were raised in rabbits which were bled prior to immunization to obtain preimmune sera. Aliquots of 0.5 mg of protein in 1 ml of phosphate-buffered saline, mixed with 1 ml of incomplete Freund's adjuvant, were injected intradermally in several sites along the backs of the rabbits on the lst, 4th, and 7th week. The rabbits were bled on the 5th week and weekly thereafter.
IgG was isolated as follows. Ten ml of antisera was passed through a protein A-Sepharose column (1 X 3 cm) (Pharmacia, Uppsala, Sweden). The column was washed with 30 ml of phosphate-buffered saline and the IgG was eluted with 5 ml of 1 M acetic acid. The eluent was dialyzed for 24 h against phosphate-buffered saline, lyophilized, and stored a t 4 "C.
Immunoprecipitation-For immunoprecipitation of cytochrome bcl complex subunits from mitochondria, mitochondrial pellets were lysed in 50 pl of 2% SDS, 5% P-mercaptoethanol, 0.06 M Tris, pH 6.8, 5 mM phenylmethylsulfonyl fluoride, 5 m~ o-phenanthroline for 3 min at 95 "C. Then, 1 ml of Triton buffer was added and the solution was clarified by centrifugation for 15 min a t 35,000 X g. Twenty to fifty pl of immunoglobulin solution (1 mg/ml) was added and the mixture was shaken for 15 min a t 4 "C. Sepharose-bound protein A (5 mg dry weight preswollen in 50 p1 of Triton buffer) was then added and the mixture was shaken for another 15 min at 4 "C. The Sephawith 1 ml of Triton buffer and twice with 1 ml of 10 m Tris, pH 7.5. rose beads were collected by centrifugation and washed three times Immunoprecipitates from reticulocyte lysates were obtained by adjusting the postribosomal supernatants or the in vitro transfer supernatants to 0.3 M NaCl and 1% Triton, followed by addition of antibodies to the cytochrome bcl complex subunits and Sepharosebound protein A, as described above.
Immunoprecipitates from in vivo experiments were obtained by adding 50 pl of immunoglobulin solution to the 39,000 X g supernatant and incubating overnight at 4 "C. Then, 250 pl of a 10% suspension of Staphylococcus aureus cells (37) was added and the mixture was shaken at room temperature for 30 min. Cells were collected by centrifugation and washed as described for protein A-Sepharose beads above.

Synthesis of Precursors to Neurospora Cytochrome bcl
Complex Subunits in Rabbit Reticulocyte Lysates-Neurospora RNA stimulated the incorporation of [35S]methionine into proteins in rabbit reticulocyte lysates by a factor of about 10. Using antibodies directed against each of the cytochrome bel complex subunits, immunoprecipitates were obtained from postribosomal supernatants of the lysates for seven of the nine subunits. Each of these immunoprecipitates migrated as a single band upon gel electrophoresis with an apparent molecular weight that was, in most cases, greater than that of the isolated subunit of the mature protein precipitated from radiolabeled mitochondria (Fig. 1). Subunits I, 11, IV (cytochrome el), V (Fe/S protein), VII, and VI11 with apparent molecular weights of 50,000,45,000,31,OOO,25,000,11,500, and 11,200, respectively, were synthesized as larger molecular weight precursors with molecular weights of 51,500, 47,500, 38,000,28,000,12,000, and 11,600, respectively. Subunit VI was synthesized with the same apparent molecular weight (14,000) as the authentic protein (Fig. 1).
A translation with N-f~rmyl-[~S]Met-tRNA, which would have confirmed the absence of an additional NH2-terminal sequence, was not done because not enough label can be expected in the very small quantity of this protein that we could immunoprecipitate.
In the case of subunit 11, relatively little radioactivity was incorporated into the mature protein when cells were grown on [%3]sulfate. Much better labeling was obtained with ['HI leucine; the immunoprecipitated protein could also be visualized by staining the gel (Fig. 1). This indicates that the mature protein may have relatively few methionine and cysteine residues.
Subunit I11 (cytochrome b) is known to be coded by mitochondrial DNA and synthesized on mitochondrial ribosomes (10) and was not synthesized in the reticulocyte lysates, as expected. Antibodies which precipitated isolated subunit IX (M, = 9,OOO) did not precipitate any protein from the reticulocyte lysates nor were they able to precipitate the mature subunit from mitochondria labeled with ["%]sulfate or [3H] leucine in uiuo.
Transfer of Cytochrome bel Complex Subunits from Postribosomal Supernatants of Reticulocyte Lysates into Isolated Mitochondria-Mitochondria isolated from Neurospora spheroplasts were suspended in postribosomal supernatants from reticulocyte lysates which had previously been incubated with ["'S]methionine and Neurospora RNA. Transfer in vitro of the precursors of subunits I, IV, V, and VI1 into the mitochondria was investigated (Figs. 2-4). The mitochondrial fraction contained both bound precursors, which were susceptible to digestion with proteinase K, as well as subunits processed to the mature size which had been transferred into a protease-resistant position. This is shown for cytochrome cl valinomycin, 10 p~ oligomycin, 1 mM ATP; 6, 2.5 p~ carboxyatractyloside, 10 p~ oligomycin. and subunit VI1 in Figs. 2 and 3. Similar results were obtained with subunit V, but not with subunit I, where the bound precursor could not be completely digested by proteinase K.
In the case of subunit IV (cytochrome el), precursor (Mr = 38,000) was found bound to the mitochondria and two species were found transferred into proteinase K-resistant locations in the mitochondria: the mature subunit (MI = 31,000) and an intermediate form with an apparent molecular weight of 35,000 (Fig. 2). This suggests that cytochrome c1 may be processed in two distinct steps during its transfer into the mitochondria. Energy Requirements for the Transfer into Mitochondria-The energy dependence of the transfer of subunit I, subunit IV (cytochrome cl), subunit V (Fe-S protein), and subunit VI1 was investigated in the in vitro system. Both the uncoupler CCCP and the ionophore valinomycin inhibited the transfer of the four proteins studied (Fig. 4). The precursors still bound to the mitochondria under these conditions, but they were not processed to their mature forms or translocated to protease-resistant locations. The latter was verified by digesting the bound precursors with proteinase K (not shown).
CCCP acts as a protonophore and breaks down the membrane potential; valinomycin in the presence of a high concentration of K' , such as that present in the reconstituted medium (approximately 90 mM), also causes a breakdown of the membrane potential. The inhibition of transfer by these two compounds could, however, also be caused by a secondary phenomenon. One major consequence of the dissipation of the membrane potential could be a decrease in the level of intramitochondrial ATP, due to the oligomycin-sensitive ATPase pumping out protons to compensate for the influx of H' or K' (40,41). This ATPase activity can be inhibited by oligomycin (42).
In order to determine whether it is the mitochondrial membrane potential which is directly required for the transfer of the cytochrome bcl complex proteins or whether it is the intramitochondrial ATP, conditions were used under which the mitochondrial ATP level is high, whereas the membrane potential is broken down. When mitochondria are incubated in a combination of CCCP, oligomycin, and ATP, the absence of a membrane potential allows the ADP/ATP carrier to transport the external ATP into the matrix space and oligomycin inhibits the hydrolysis of this ATP, causing the matrix ATP level to rise (43). Under these conditions, transfer of cytochrome bcl complex proteins was inhibited (Fig. 4). Furthermore, when the membrane potential was kept intact but the mitochondria were depleted of ATP, transfer still occurred. This was accomplished by adding carboxyatractyloside to block the ADP/ATP carrier and prevent the import of ATP, as well as oligomycin to prevent the synthesis of ATP in the mitochondria (Fig. 4). These results show that it is the  mitochondrial membrane potential and not the intramitochondrial ATP that provides the energy for the import and/ or processing of the precursor proteins.
Competition Experiments with Apocytochrome c-There is evidence to suggest that the transfer of cytochrome c into Neurospora mitochondria involves specific receptor sites (44, 45). The precursor, apocytochrome c, was shown to bind to specific sites on the mitochondria before its transfer and conversion to holocytochrome c by attachment of the heme group. Unlabeled apocytochrome c was able to compete with radioactive apocytochrome c for binding to these sites and addition of excess unlabeled apocytochrome c resulted in reduced amounts of radioactive holocytochrome c appearing in the mitochondria during transfer in vitro (44). In order to investigate whether cytochrome c1 might be imported by the same pathway, we studied the effect of excess unlabeled apocytochrome c on the transfer of cytochrome c1 in uitro. Fig. 2 shows that 1 p~ apocytochrome c (lane 4) did not inhibit the transfer of cytochrome c1 when compared to the control (lane 1). This concentration of apocytochrome c, however, significantly reduces the amount of radioactive'holocytochrome c found in the mitochondria after transfer in uitro (44). This was confmed by immunoprecipitating holocytochrome c from the samples used for Fig. 2, lanes 1 and 4. Cytochrome c1 must therefore be imported by a different pathway from that of cytochrome c.
Transfer in Vivo of Cytochrome cl-In order to show that the two-step processing observed for cytochrome c1 was not an in vitro artifact, the transfer of cytochrome cI was also studied by pulse-labeling with [3H]leucine in uiuo. Since the transfer in vivo at 25 "C was too rapid, the cells were cooled to 8 "C before the ["Hlleucine was added. As can be seen in Fig. 5, both labeled precursor and labeled intermediate form appeared while the cells were at 8 "C, but conversion to the mature form did not take place. After 12 min, unlabeled leucine and cycloheximide were added to prevent any further incorporation of label and the temperature was raised to 25 "C. This caused the accumulated intermediate form to be quantitatively converted to the mature protein (Fig. 5). A second experiment was carried out where the intermediate form was allowed to accumulate during a ["Hlleucine pulse at 8 "C and then CCCP was added to the cell suspension along with the unlabeled leucine and cycloheximide. When the temperature was raised to 25 "C, no conversion to the mature cytochrome c was observed (Fig. 6). Under these conditions, the transfer of cytochrome c, which is not energy-dependent (49), was not inhibited by CCCP. This shows that the second proteolytic processing step in the transfer of cytochrome c1 is also energy-dependent.

Transfer of Neurospora Precursor Proteins into Yeast
Mitochondria-When %-labeled Neurospora precursors in a reticulocyte lysate were incubated with mitochondria isolated from yeast cells, import of labeled proteins took place. Using antibodies against subunit V (Fe-S protein) of the Neurospora cytochrome bcl complex, both bound precursor (Mr = 28,000) and a protein with the apparent molecular weight of the mature subunit (Mr = 25,000) could be immunoprecipitated from the yeast mitochondria following transfer in uitro. The proteolytically processed mature subunit had been translocated into the mitochondria, as judged by its resistance to added protease (Fig. 7). The transfer was dependent on the mitochondrial membrane potential, as shown by the inhibition in the presence of valinomycin or antimycin A + oligomycin.

DISCUSSION
We have shown in earlier publications that a number of mitochondrial proteins synthesized on cytoplasmic ribosomes are translocated into Neurospora mitochondria by a posttranslational mechanism (20, 21,49). Some of these proteins, such as subunit 9 of the oligomycin-sensitive ATPase or citrate synthase, are synthesized as larger precursors with additional sequences that are removed during or after transfer into mitochondria. Other mitochondrial proteins, such as cytochrome c (22), the ADP/ATP carrier (31), and the mitochondrial porin2 are transferred without proteolytic processing. In the experiments reported here, both of these modes of translocation were observed among the various subunits of the cytochrome bcl complex. Subunits I, 11, IV, V, VII, and VI11 were synthesized as precursors with apparent molecular weights larger than those of the respective mature subunits. Subunit VI, however, appeared not to have a precursor with an additional sequence.
Subunit IV (cytochrome cl) is unusual in that it seems to be processed in two discrete steps. In the first, the precursor (M, = 38,000) is processed to an intermediate form (Mr = 35,000), which is already in a location or a form that makes it resistant to proteolysis by proteinase K added to intact mitochondria. In the second processing step, the intermediate form is converted to the size of the mature protein (M, = 31,000); this species is also in a protease-resistant location, presumably the functional site.
Subunits VI1 and VI11 were also unusual because the apparent molecular weights of their precursors were only about 500 larger than those of the respective mature subunits. This would correspond to an additional sequence of only about five amino acids in each case, which is much shorter than those observed on other proteins that are proteolytically processed upon translocation across a membrane. The precursors to * H. Freitag and W. Neupert, unpublished observations. cytochrome cI and the Fe-S protein have also been studied in yeast (36.50). Interestingly, the extensions appear to have the same size as in Neurospora. In addition, our finding that a Neurospora precursor could be imported into yeast mitochondria and was processed to the correct apparent molecular weight shows that the components of the transfer pathway must be very similar in these two species. The transfer of Neurospora precursors into rat liver mitochondria has also been observed (42), indicating that proteins involved in import and processing of precursors must be highly conserved in evolution.
For all of the cytochrome bcl complex subunits whose transfer was studied, the import and proteolytic processing of the precursors in vitro required a membrane potential across the inner mitochondrial membrane. We discriminated between the membrane potential and the closely inter-related (40) matrix ATP level by employing conditions under which the level of ATP in the matrix was high and the membrane potential was dissipated. The precursors were able to bind to the mitochondria under these conditions, but they were neither imported nor proteolytically processed. On the other hand, when the matrix ATP was depleted but the membrane potential was not destroyed, the transfer of the subunit proteins was not inhibited. Identical results have recently been obtained with two other integral inner membrane proteins in Neurospora: ADP/ATP carrier and subunit 9 of the oligomycin-sensitive ATPase (42). Nelson and Schatz (50) concluded from experiments with whole yeast cells that processing of precursors to the cytochrome bcl complex proteins, cytochrome cl, and the Fe-S protein required ATP in the mitochondrial matrix. Their results showed that rho-mutants, which should not be able to generate a membrane potential via electron transport or the reversed action of ATP synthase (51, 52)) are still able to import these proteins into their mitochondria. It cannot be ruled out, however, that rhomitochondria can still generate a small membrane potential sufficient to drive protein transfer, by import of ATP or some other process.
In the case of cytochrome cl, where the proteolytic processing of the precursor takes place in two separate steps, our experiments in vivo showed that both the processing of the When the mutant was given 8-aminolevulinic acid, the intermediate form was converted to mature cytochrome CI. In order to eliminate the possibility that the intermediate form is simply an apocytochrome cl, missiig the heme group but having an apparently larger molecular weight, we prepared apocytochrome c1 by a method established for cytochrome c (54). In our gel system, this apocytochrome c1 had an apparent molecular weight equal to or slightly smaller than that of mature cytochrome cI (not shown) and therefore could not be the intermediate form.
The transfer of cytochrome c into Neurospora mitochondria has been studied more extensively (22,44). The precursor, apocytochrome c, binds to specific sites on the mitochondria and is imported when the heme group is attached to form holocytochrome c. This process does not require energization of the mitochondria. Excess unlabeled apocytochrome c was able to compete with radioactive precursor and inhibit the appearance of radioactive holocytochrome c in the mitochondria, but it had no effect on the import and processing of cytochrome c1 precursor. This observation, together with the difference in energy requirements, lead us to conclude that the pathway by which cytochrome CI is transferred is different from that of cytochrome c.