Sequence requirements for mitochondrial import of yeast cytochrome c.

Apocytochrome c is synthesized in the cytoplasm, transported to the mitochondrial intermembrane space, and subsequently covalently attached to heme in a reaction catalyzed by the enzyme cytochrome c heme lyase. We have investigated the amino acid sequences in cytochrome c which are required for mitochondrial import, using a systematic series of site-directed alterations of the CYC7-H3 gene which encodes iso-2-cytochrome c in the yeast Saccharomyces cerevisiae. Import of the altered apocytochromes c was assayed in yeast strains that overexpressed cytochrome c heme lyase. Under these conditions, there was efficient mitochondrial accumulation of forms of apocytochrome c which are incapable of having heme covalently attached. In fact, all apocytochromes c containing deletions located to the carboxyl-terminal side of His27 efficiently accumulated in the mitochondria of strains overexpressing heme lyase, even though all but one of these deletion-containing proteins were incapable of heme attachment. A minimum length of polypeptide chain at the extreme amino terminus of cytochrome c, rather than any specific sequence element in this region, appears to be required for efficient mitochondrial import. Certain amino acid substitutions in the region extending from Gly15 to Leu18, at residue Phe19 and at residue His27, lead to reduced mitochondrial import of apocytochrome c, resulting from stalling of the altered apocytochrome c in partially imported states.

Most mitochondrial proteins, including cytochrome c, are encoded in the nucleus, synthesized in the cytoplasm, and then transported across one or more mitochondrial membranes. For several of these imported mitochondrial proteins, the sequences that direct them to mitochondria, and to the correct subcompartment of mitochondria, have been identified. Such sequences are generally located near the amino terminus of the imported protein, often as part of a presequence that is cleaved during or after import.
Cytochrome c is translocated into the mitochondrial intermembrane space from its site of synthesis in the cytoplasm along a pathway that differs in several respects from pathways followed by other, more extensively studied, mitochondrial proteins that are targeted to the same mitochondrial subcompart-ment. Cytochrome c is initially synthesized as apocytochrome c, lacking heme. During or after import into mitochondria, heme is covalently attached via stereospecific thioether linkages to two cysteine residues in the protein. This reaction is catalyzed by the enzyme cytochrome c heme lyase (CCHL), 1 which is located in the mitochondrial intermembrane space, predominantly associated with the inner membrane. Heme attachment is accompanied by a transition from a partially extended conformation to the native compact structure. Cytochrome c does not contain any cleaved targeting sequence. No protease-sensitive receptor on the external surface of mitochondria mediates the initial binding and import of apocytochrome c. Cytochrome c import does not depend on the presence of an electrochemical potential across the inner mitochondrial membrane or on the presence of ATP (for review see Hartl et al., 1989;Stuart and Neupert, 1990;Kiebler et al., 1993;Hannavy et al., 1993;Dumont, 1995).
The signals in the cytochrome c sequence which are responsible for the mitochondrial import have not been well characterized previously. Since cytochrome c does not contain a cleavable signal presequence at the amino terminus of the protein, the import signal must be a part of the mature sequence. Furthermore, sequences important for import are unlikely to reside at the extreme amino terminus of cytochrome c, since deletions and replacements of up to 11 residues at the amino terminus of iso-1-cytochrome c of yeast (Sherman and Stewart, 1973;Baim et al., 1985;Hampsey et al., 1988) do not prevent normal targeting and maturation.
Sequences responsible for subcellular targeting of many proteins have been identified through mutational alteration of the relevant structural genes followed by analysis of the subcellular distribution of the altered protein. Application of this approach to cytochrome c has been problematic because of the difficulty of separating effects on heme attachment from effects on import. Failure to attach heme to cytochrome c leads to accumulation of the precursor, apocytochrome c, in the cytoplasm. Heme attachment can be blocked by competition with a heme analog that cannot be incorporated into holocytochrome c, by mutational alteration of cytochrome c, or by a lack of CCHL, the enzyme catalyzing the attachment (Hennig and Neupert, 1981;Dumont et al., 1988Dumont et al., , 1991Nargang et al., 1988). Such coupling between heme attachment and import is consistent with a mechanism in which apocytochrome c is reversibly transported across the mitochondrial outer membrane and subsequently trapped in the intermembrane space by the protein folding reaction accompanying CCHL-mediated heme attachment. However, direct participation of CCHL in import has not been completely ruled out (see Dumont, 1995;Mayer et al., 1995). * This work was supported by National Institutes of Health Grant R01 GM12702. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. Two previous mutational studies of sequences responsible for targeting cytochrome c to mitochondria were conducted in systems where mitochondrial localization could be assayed independent of heme attachment. The first of these was carried out in vitro, assaying association of Drosophila melanogaster apocytochrome c with isolated mouse liver mitochondria in the presence of a cytoplasmic factor from wheat germ Sprinkle et al., 1990). For unknown reasons, mitochondrial association was not accompanied by a significant level of heme attachment in this system. The second of the mutational studies was carried out in vivo, using gene fusions between yeast iso-1-cytochrome c sequences and the chloramphenicol acetyltransferase sequence Scarpulla, 1990a, 1990b). In this system, the presence of a large folded protein fused to cytochrome c prevented correct localization to the intermembrane space, which, in turn, prevented efficient heme attachment. Both mutational studies concluded that cytochrome c contains two targeting sequences that are at least partially redundant, one in the first 60 residues of the protein, and the other in a more carboxyl-terminal region.
We have investigated the sequences involved in targeting cytochrome c to mitochondria using an in vivo system that is closely related to the authentic pathway of import in the yeast Saccharomyces cerevisiae. To separate the effects on heme attachment from effects on import, we have capitalized on a previous observation that high level expression of CCHL allows apocytochrome c to be efficiently imported into mitochondria in the absence of heme attachment (Dumont et al., 1991). In this system, import of apocytochrome c appears to be driven by direct binding to the heme lyase rather than by conversion to holocytochrome c (Dumont et al., 1991;Mayer et al., 1995), thus allowing evaluation of the import competence of altered cytochromes c that are not capable of heme attachment.
We report the effects of a systematic set of deletions, as well as selected point mutations, on the import of yeast iso-2-cytochrome c into mitochondria. Levels of mitochondrial apocytochrome c were assayed by immunoblotting subcellular fractions of strains expressing high levels of CCHL. In addition, examination of heterozygous diploid strains revealed that expression of certain altered apocytochromes c prevented the maturation and targeting of coexpressed normal cytochrome c. Some of the altered cytochromes c appeared to be stalled at specific stages of translocation into mitochondria. Taken together, these analyses have led to the identification of amino acid residues required for cytochrome c import into mitochondria and to the detection of partially translocated forms of cytochrome c which accumulate when certain sequence elements are missing.
Construction of Plasmids-A series of mutations in the translated region of the CYC7-H3 gene was generated by the method of Kunkel et al. (1987), using the plasmid pAB595 (Dumont et al., 1991) essentially as described previously (Das et al., 1989). The sequences of the mutagenic oligonucleotides are listed in Table I. The names of the plasmids  and the corresponding mutagenic oligonucleotides are listed in Table II. Plasmid pAB1385, containing the 12CA5 epitope (Kolodziej and Young, 1991) YPYDVPDYA at the extreme amino terminus of iso-2-cytochrome c, was also derived from plasmid pAB595, using the same procedures described above with the oligonucleotide OL91.228. A 2.3-kilobase EcoRI-SalI fragment from plasmid pAB1385 containing the CYC7-H3 gene with the epitope was ligated into the EcoRI-SalI site of plasmid

Import of Yeast Cytochrome c
YCp50 (Rose et al., 1987) to construct a plasmid pAB1386. For construction of the cyc7-H3-87 allele (see Fig. 1) with the 12CA5 epitope replacing the first 14 amino acids of the protein, oligonucleotides OL91.072 and OL91.038 were used with plasmid pAB1386 in the polymerase chain reaction (PCR) to amplify the epitope-containing CYC7-H3 gene as described by Scharf (1990). The sample was cycled 30 times with denaturation at 94°C for 1 min, annealing at 50°C for 2 min and 72°C for 4 min. The PCR product was treated with restriction endonuclease SalI and self-ligated. Reverse PCR using this same protocol was then performed on the circular product with oligonucleotides OL91.300 and OL91.298 as primers, thereby creating the deletion. The PCR fragment was treated with restriction endonuclease BamHI and self-ligated. The circularized product was subjected to an additional round of PCR using oligonucleotides OL91.335 and OL91.257 as primers. Plasmid YCp50 was digested with restriction endonuclease NruI, and a single T was added to each end of the DNA molecule with deoxy-TTP through the action of terminal deoxynucleotidyl transferase. Finally, the PCR fragment containing the CYC7-H3 allele with the epitope and the deletion was ligated to the modified YCp50 vector to construct a plasmid pAB1326.
Subcellular Fractionation-For preparation of subcellular fractions, all strains except B-9152 were cultured in S.D. medium containing 0.1% casamino acids plus histidine and leucine  to an A 600 of 3.0 -4.0. Strain B-9152 was grown in uracil omission medium  to an A 600 of 3.0 -4.0. Subcellular fractions were prepared as described by Daum et al. (1982) and Dumont et al. (1991Dumont et al. ( , 1993, with the following modifications.: (i) Zymolyase 100T (ICN Biomedicals, Inc.) was used to make spheroplasts. (ii) Cells were lysed in 0.5 M mannitol, 50 mM KCl, 20 mM HEPES/KOH, 1 mM EDTA, 0.1% bovine serum albumin, 1 mM phenylmethylsulfonyl fluoride, pH 7.4. (iii) Proteinase K digestions of mitochondria were carried out at 1-2 mg of mitochondrial protein and 100 g of enzyme per ml for 30 min at 0°C.
Western Blot and Immunologic Procedures-Samples of each fraction were separated by electrophoresis on a 10% acrylamide-sodium dodecyl sulfate gel (Schagger and von Jagow, 1987), transferred to 0.2-m nitrocellulose (Hoefer Scientific Instruments), and probed with affinitypurified anti-cytochrome c polyclonal antibodies (Dumont et al., 1993) followed by horseradish peroxidase-conjugated goat anti-rabbit immunoglobulin G (Bio-Rad). Anti-cytochrome c antibodies were diluted in 5% fetal calf serum in phosphate-buffered saline (Maniatis et al., 1989) and incubated with the nitrocellulose membranes for 3 h. Secondary antibodies were used at 1:1,000 for 2 h. The membranes were then washed with phosphate-buffered saline and equilibrated in 10 mM sodium citrate, 10 mM EDTA, pH 5.0, for 5 min. After an additional 10-min incubation with 10 mM sodium citrate, 10 mM EDTA, 1% dextran sulfate, pH 5.0, the membranes were again washed, three times for 5 min each, with 10 mM sodium citrate, 10 mM EDTA, pH 5.0. Finally the membranes were incubated in 10 mM sodium citrate, 10 mM EDTA, 1% dextran sulfate, pH 5.0, with 0.1 mg/ml tetramethylbenzidine and 0.003% hydrogen peroxide until colors were developed (McKimm-Breschkin, 1990). Quantitative comparisons of cytochrome c levels in different subcellular fractions were performed by comparing the intensities of a series of dilutions of the different fractions on immunoblots (Dumont et al., 1991).
Determination of Cytochrome c Levels-The strains, described in Fig.  2 and Tables II and III, were grown on 1% sucrose medium (Sherman et al., 1974) at 30°C for 3 days, and the absorption spectra were recorded as described previously (Hickey et al., 1991). A low temperature The haploid series 1 contains single copies of the cyc7-H3-xx alleles and multiple copies of the CYC3 gene (cyc1-⌬ cyc7-H3-xx p[CYC3] N ) (see Table  III). Accumulation of iso-2-cytochrome c, even without heme attachment, occurs because of overexpression of CCHL by p[CYC3] N , unless the cyc7-H3-xx allele causes a specific defect. The diploid series was generated by crosses of strain B-6945, containing CYC1, with haploid series 2, which contains single copies of the cyc7-H3-xx alleles and a single copy of CYC3, (cyc1-⌬ cyc7-H3-xx CYC3 ϫ CYC1 cyc7-⌬ CYC3) (see Table III). The diminution of holo-iso-1-cytochrome c, encoded by CYC1, is caused by the altered apo-iso-2-cytochrome c, encoded by cyc7-H3-xx, due to the competition between apo-iso-1-cytochrome c and apo-iso-2-cytochrome c at any one of the steps leading to the normal accumulation of holo-iso-1cytochrome c.  Dumont et al. (1991). b Amino-terminal fusion of iso-2-cytochrome c or ⌬(Ala 1 -Lys 14 ) with the 12CA5 epitope YPYDVPDYA. c See "Materials and Methods" for details.
(Ϫ196°C) spectrophotometer (model 14DS, Aviv, Lakewood, NJ) was used to scan the intact yeast cells. The ␣-peaks of cytochromes a⅐a 3 , b, c 1 , and c are located, respectively, at 602.5, 558.5, 553.3, and 547.3 nm. The heights of the C ␣ -peaks and comparisons with strains having known amounts of cytochrome c were used to determine the levels of cytochrome c (see Fig. 5).

RESULTS
Experimental Design: Isocytochromes c System-Normal strains of S. cerevisiae contain two isocytochromes c, iso-1cytochrome c and iso-2-cytochrome c, encoded, respectively, by the CYC1 and CYC7 chromosomal genes (Sherman et al., 1966;Downie et al., 1977). These two isozymes constitute, respectively, 95 and 5% of the total cellular complement of cytochrome c (Sherman et al., 1965). Although both are approximately equivalent in their function in electron transport, the properties of each of the apo-iso-cytochromes c differ. In particular, studies of mutants defective in heme attachment revealed that apo-iso-1-cytochrome c is rapidly degraded if it is not converted to holocytochrome c, but that apo-iso-2-cytochrome c is stable under the same conditions (Dumont et al., 1990). As a result, cyc3-⌬ strains, containing a deletion of the structural gene that encodes CCHL, are deficient in apo-iso-1cytochrome c but contain apo-iso-2-cytochrome c at a level that is approximately the same as the level of holoiso-2-cytochrome c in a related CYC3 strain containing CCHL. Thus, to allow analysis of the subcellular distribution of apocytochrome c, all of our import studies were conducted with altered iso-2-cytochrome c.
Because iso-2-cytochrome c normally constitutes only about 5% of the total complement of cytochrome c, making it difficult to detect, we have used the CYC7-H3 allele, which constitutively overproduces iso-2-cytochrome c to a level that is approximately the same as the normal level of iso-1-cytochrome c (McKnight et al., 1981). This increased expression is due to a deletion that removes the normal promoter and fuses the translated region of iso-2-cytochrome c to an upstream promoter (McKnight et al., 1981;Melnick and Sherman, 1993).
Haploid Series 1: Analysis of the Subcellular Distributions of Altered Apocytochromes c-Oligonucleotide-directed mutagenesis of the CYC7-H3 allele on plasmid pAB595 was used to generate a series of deletions and amino acid replacements of iso-2-cytochrome c (Table II). Strain B-9077 (cyc1-⌬ cyc7-⌬ CYC3 p[CYC3] N ) was transformed with these p[cyc7-H3-xx] plasmids containing the series of cyc7-H3-xx alleles that encode the altered forms of iso-2-cytochrome c. The resulting series of haploid strains, containing an integrated copy of each of the p[cyc7-H3-xx] plasmids, are referred to as haploid series 1 and are abbreviated as follows (Table III): Each strain of haploid series 1 has the following relevant genetic properties: a single copy of the cyc7-H3-xx allele integrated into the chromosome; a cyc1-⌬ deletion causing a total lack of iso-1-cytochrome c; a normal copy of the CYC3 gene encoding CCHL; and a multicopy plasmid p[CYC3] N leading to overproduction of CCHL driven by the actin promoter. As described previously (Dumont et al., 1991) and discussed above, overproduction of CCHL leads to efficient accumulation of apocytochrome c inside mitochondria when heme attachment does not occur. Such accumulation appears to be mediated by a direct interaction between apocytochrome c and CCHL in a nonproductive enzyme-substrate complex located in the mitochondrial intermembrane space (Dumont et al., 1991;Mayer et al., 1995;Dumont, 1995). The combination of genetic elements in haploid series 1 allows examination of the effects of cyc7-FIG. 1. Scheme for making the deletion mutation cyc7-H3-87 with PCR. Two arrows inside the plasmid pAB1386 denote two oligonucleotides that are used to PCR amplify the Ept ϩ iso-2 gene containing the 12CA5 epitope inserted upstream of the CYC7 gene. The directions of the arrows indicate the 5Ј 3 3Ј direction of the oligonucleotides. The arrow that is parallel to the circle represents the oligonucleotide that is fully complementary to the plasmid sequence. The tail at the beginning of the arrow represents the part of the oligonucleotide sequence which is not complementary to the plasmid sequence. The Ept ϩ iso-2 gene is represented by a solid box. The restriction endonuclease SalI (S) sites are marked by short lines. The PCR product was digested with SalI and self-ligated at 16°C for 12 h. Reverse PCR was performed on the ligated plasmid with two oligonucleotides that incorporated a BamHI (B) site at the 5Ј ends of their sequences. The PCR product was digested with BamHI and self-ligated. A third round of PCR was then used to amplify the mutated gene. Plasmid YCp50 was digested with restriction endonuclease NruI (N). The blunt ended DNA fragment was incubated with dideoxy-TTP (ddTTP) and terminal deoxynucleotidyl transferase (TDTase) at 37°C for 1 h. A single T was added to two 3Ј ends of the molecule. The modified plasmid was then ligated with the mutated gene, resulting in plasmid pAB1326, which contained cyc7-H3-87.
( Table II H3-xx alleles on mitochondrial accumulation of apocytochrome c under conditions where heme is not attached to the protein.
Mitochondrial accumulation of mutant cytochromes c was assayed by immunoblotting of subcellular fractions derived from the relevant yeast strains. The antibodies used recognize both holocytochrome c and apocytochrome c.
Initially, the studies described below were performed with the cyc7-H3-xx alleles to which a 12CA5 epitope tag from influenza hemagglutinin had been added (Kolodziej and Young, 1991). However, in some cases, the presence of this tag appeared to alter the intracellular distribution of cytochrome c (see below). Comparison of the immunologic reactivity of antiepitope-tag antibodies with that of anti-cytochrome c antibodies demonstrated that the affinity-purified polyclonal anticytochrome c antibodies efficiently recognized all of the mutant cytochromes c that we tested. Thus, anti-cytochrome c antibodies rather than anti-epitope-tag antibodies were used in the assays of intracellular cytochrome c distributions presented below.
Haploid Series 2: Analysis of Heme Attachment and Function of Altered Iso-2-cytochromes c-The function and degree of heme attachment of the cyc7-H3-xx altered iso-2-cytochromes c were determined with the following haploid series 2 (Table III) which did not overproduce CCHL: The level of holo-iso-2-cytochrome c in each member of this series was estimated by low temperature (Ϫ196°C) spectroscopic examination of intact yeast cells. The ability of each altered iso-2-cytochrome c to function in electron transport was estimated from the ability of each strain to grow on media containing nonfermentable carbon sources, such as glycerol or ethanol. The levels and functions were assigned values in comparison with the control strain cyc1-⌬ CYC7-H3 CYC3 ϩ .
Diploid Series: Assay of Competition between Normal Cytochrome c and Altered Iso-2-cytochromes c in the Mitochondrial Import Pathway-Heterozygous diploid strains were used to determine if the altered apo-iso-2-cytochromes c retained sufficient integrity to compete with normal cytochrome c or if the altered forms even had abnormally high affinities for any components of the mitochondrial import pathway which could lead to displacement of normal cytochrome c. The extent of the displacement was estimated from the levels of holocytochrome c in the following series of diploid strains (Table III): Thus, each of the diploid strains contains the normal level of CCHL, a single copy of the cyc7-H3-xx encoding one of the altered iso-2-cytochromes c, and a single copy of CYC1 ϩ encoding normal iso-1-cytochrome c.
Using low temperature spectroscopy of intact cells, the level of holocytochrome c in each of the strains of this series was compared with the level in two control strains. One control strain contains two normal cytochrome c alleles. The other contains one normal cytochrome c allele accompanied by a deletion of the other allele. The level of holocytochrome c in the double hemizygous CYC7-H3/cyc7-⌬ CYC1 ϩ /cyc1-⌬ control strain was defined as 100%. The level of holocytochrome c in the heterozygous cyc7-⌬/cyc7-⌬ CYC1 ϩ /cyc1-⌬ control strain corresponded to 50% of the previous one. We used diploid series in which the normal cytochrome c allele was iso-1-cytochrome c (cyc1-⌬ cyc7-H3-xx CYC3 ϩ ϫ CYC1 ϩ cyc7-⌬ CYC3 ϩ ) rather than diploid series in which the normal cytochrome c allele was iso-2-cytochrome c (cyc1-⌬ cyc7-H3-xx CYC3 ϩ ϫ cyc1-⌬ CYC7-H3 CYC3 ϩ ) because apo-iso-1-cytochrome c is much more labile than apo-iso-2-cytochrome c; therefore, the effi-ciency of production of holocytochrome c should be more sensitive to the inhibition caused by the presence of an altered apo-iso-2-cytochrome c.
The results presented below reveal that all diploid strains, except for the control strain, contained less than 50% of holocytochrome c, indicating that all of the altered apo-iso-2-cytochromes c partially displaced the normal forms. Furthermore, several of the diploid strains contained as little as 15% of the normal level, suggesting that these altered forms had a higher affinity than the normal protein for some components involved in import or maturation of cytochrome c.
The mutations causing less than 50% of holocytochrome c in the heterozygous strains are formally designated as dominantnegative or semidominant-negative mutations. Such dominant-negative mutations are generally encountered in the single copy heterozygous state when the mutant form of the protein has an abnormally high affinity for some other cellular component, thus displacing the normal form.
Classes of Mutants-Regions important for heme attachment and mitochondrial import were initially investigated with 10 deletions ( Fig. 2A), spanning the entire 112-amino acid sequence of iso-2-cytochrome c (Fig. 3). The mutants were analyzed for: (i) the levels and function of holocytochrome c in haploid series 2; (ii) the levels and subcellular localization of the altered apocytochromes c in haploid series 2 (Fig. 4); and (iii) the displacement of normal cytochrome c by the altered cytochromes c in the diploid series (Fig. 4).
The levels of holocytochromes c in haploid series 2 are presented in Fig. 2. Representative results of low temperature spectrophotometric recordings are shown in Fig. 5. Analysis of the subcellular distributions included: (i) determination of the amount of cytochrome c that was associated with mitochondria but not necessarily protected from proteinase K digestion; (ii) determination of the amount of cytochrome c inside mitochondria which was protected from digestion with externally added proteinase K; and (iii) determination of the susceptibility of each cytochrome c to proteinase K digestion in the presence of Triton X-100, which disrupts the mitochondrial membrane. Based on these measurements, each of the 25 mutationally altered iso-2-cytochromes c could be assigned to one of five classes (Class I-Class V, see Table IV). All but Class II and Class V altered cytochromes c were deficient in heme attachment.
Class I Mutants Lack Heme Attachment but Are Accumulated Inside Mitochondria-The design of this study was based on the properties of a previously examined mutant in which the two cysteine residues required for heme attachment were replaced by serine residues (Dumont et al., 1991). As confirmed in the present study, no holocytochrome c was formed from a mutant in which serine residues had been substituted for Cys 23 and Cys 26 . However, in strains that overexpressed the heme lyase, 40% of the cellular complement of this altered form of apocytochrome c accumulated in the mitochondria. The results presented in Fig. 2 show that expression of the Cys 23 3 Ser 23 Cys 26 3 Ser 26 cytochrome c also interfered with formation of the normal form of iso-1-cytochrome c when both forms were expressed in diploid cells. Such cells contained only 15% of the normal level of iso-1-cytochrome c compared with the 50% level that was seen in strains containing one copy of the CYC1 gene encoding iso-1-cytochrome c, and no other cytochrome c genes. The data in Fig. 2 also demonstrate similar effects when alanine, rather than serine, was substituted for the two cysteine residues. Substitution of alanine or serine for either one of the two cysteine residues resulted in a phenotype that was indistinguishable from substitutions at both positions (Fig. 2C).
These mutants having Cys 23 or Cys 26 replacements were assigned to Class I.
Six of the 10 deletions spanning the CYC7 gene gave rise to phenotypes typical of Class I mutations. The deletions in this class included ⌬(Val 37 -Arg 47 ) and the five deletions encompassing the region from Thr 58 to Lys 112 , comprising most of the carboxyl-terminal half of the molecule. Each of these mutant apo-iso-2-cytochromes c lacked heme attachment. Furthermore, each was imported into mitochondria and interfered with the formation of holo-iso-1-cytochrome c in diploid strains with approximately the same efficiency as the mutations involving substitutions for the cysteine residues involved in heme attachment. Thus it appeared that these deletions did not reside in sequences required for mitochondrial import or binding to CCHL.
The ⌬(Ile 104 -Lys 112 ) deletion has been included in this class of mutants although the percentage of this form of apocyto-

FIG. 2. Summary of the mutations in the iso-2-cytochrome c and their effects.
The various alterations of iso-2-cytochrome c are listed on the left. Normal denotes the normal iso-2-cytochrome c. No iso-2-cyt. c denotes the total deficiency of iso-2-cytochrome c. Deletions are denoted by ⌬, followed by the first and last residue, for example, ⌬(Ala1-Lys14) denotes the mutant in which the sequence from Ala 1 to Lys 14 is deleted. Amino acid replacements are indicated using arrows, for example, Cys23 3 Ala23 denotes the mutant in which the cysteine residue at position 23 is replaced by the alanine residue. Ept ϩ ⌬(Ala1-Lys14) denotes the mutant in which the amino-terminal peptide from Ala 1 to Lys 14 is replaced by the 12CA5 epitope. The iso-2-cytochrome c constructs are schematically shown in the center column of the figure.
For panels A and B, the solid box is the intact amino acid sequence of iso-2-cytochrome c, and the open box is the part of iso-2-cytochrome c sequence which has been deleted. The hatched box is the 12CA5 epitope. Amino acid residues Cys 23 and Cys 26 , the sites of the heme attachment, are shown above the protein constructs. For panel C, only part of the amino-terminal region is shown. Deletions are again indicated by the open boxes, and amino acid replacements are shown above the protein constructs. The results are listed on the right. Holo-Cyt. c refers to the total amount of holocytochrome c in intact cells. Internal Cyt. c refers to the amount of cytochrome c inside the mitochondria which is protected from digestion by externally added proteinase K (see the Mito.ϩPro.K lane of Fig. 4). Asso. Cyt. c refers to the amount of cytochrome c which is associated with the mitochondria after the mitochondria are isolated and washed (see the Mito. lane of Fig. 4). Holo. in 2N refers to the amount of holocytochrome c produced in the heterozygous diploid strains and reflects the extent of displacement of holocytochrome c by the altered apo-iso-2-cytochromes c. The cytochrome c level that is associated with the mitochondria and the level inside the mitochondria were calculated from the comparisons of the intensities of the cytochrome c bands of the cytoplasm and the relevant mitochondria fractions when various amounts of total protein of each fraction were loaded in the Western blot analyses. For purposes of this calculation, mitochondria were assumed to constitute 2% of total cellular protein. chrome c localized within mitochondria was slightly lower than that seen in the other members of this class. This less efficient accumulation suggests that sequences at the extreme carboxyl terminus of cytochrome c may play a minor role in some stages of targeting or binding to CCHL. In a diploid strain, expression of the ⌬(Ile 104 -Lys 112 ) deletion inhibited formation of normal holocytochrome c to the same extent as other Class I mutations.

The Class II Mutant Lacking the Amino-terminal Region Is Partially Defective for Production of Holocytochrome c and for
Mitochondrial Accumulation of Cytochrome c-The haploid series 2 strain with the ⌬(Ala 1 -Lys 14 ) contained 10% of the normal amount of holocytochrome c. The corresponding haploid series 1 strain expressing the amino-terminal deletion in the presence of a high level of CCHL accumulated approximately 10% of the total cellular complement of immunoreactive cytochrome c inside mitochondria ( Fig. 2A). Since most of the ⌬(Ala 1 -Lys 14 ) that is imported appears to be converted to holocytochrome c, the amino-terminal sequence did not appear to be required for attachment of heme but might be required for the initial association with mitochondria or binding to CCHL. The effects of the amino-terminal deletion were not due to a requirement for a particular sequence in this region, because replacement of the deleted 14 amino acids with 9 amino acids of unrelated 12CA5 epitope led to production of holocytochrome c inside mitochondria at a level that was indistinguishable from the normal level.
Class III Mutants Affecting the Region from Gly 15 to Gln 24 Lack Holocytochrome c and Exhibit Greatly Diminished Mitochondrial Accumulation of Apocytochrome c-The ⌬(Gly 15 -Gln 24 ) deletion completely blocked conversion to holocytochrome c and prevented accumulation of any detectable apocytochrome c inside mitochondria. The sequences responsible for this phenotype were further defined by examining the deletions ⌬(Gly 15 -Arg 22 ), ⌬(Gly 15 -Leu 18 ), and ⌬(Phe 19 -Arg 22 ).
The ⌬(Gly 15 -Arg 22 ) deletion was similar to the ⌬(Gly 15 -Gln 24 ) deletion. Thus, loss of Gln 24 and Cys 23 , which is one of the sites of covalent linkage of heme, does not explain the Class III phenotype. The lack of a role for Cys 23 was also supported by analysis of the Cys 23 3 Ser 23 and Cys 23 3 Ala 23 mutations, which did not prevent mitochondrial accumulation of apocytochrome c.
The ⌬(Gly 15 -Leu 18 ) and ⌬(Phe 19 -Arg 22 ) deletions each exhibited only partial defects in the accumulation of mitochondrial apocytochrome c compared with the entire ⌬(Gly 15 -Arg 22 ) deletion. This indicated that sequences from each of these two smaller regions might be important for this process. The phenotypes of the ⌬(Phe 19 -Arg 22 ) deletion and the single substitution Phe 19 3 Ala 19 were indistinguishable, indicating that Phe 19 might be the critical residue in the Phe 19 -Arg 22 region. Thus, Phe 19 and one or more residues in the Gly 15 -Leu 18 region appeared to be required for efficient mitochondrial accumulation of apocytochrome c.
Even though the ⌬(Gly 15 -Gln 24 ) apocytochrome c did not accumulate inside mitochondria to any appreciable extent, expression of this altered apocytochrome c in the heterozygous diploid strain displaced the formation of the normal holocytochrome c to the same extent as the Class I apocytochromes c, which were accumulated efficiently inside mitochondria (see Fig. 2).
Class IV Mutations Affecting His 27 Prevent Translocation across the Mitochondrial Outer Membrane-The ⌬(Gln 25 -Lys 36 ) deletion resulted in normal association of apocytochrome c with mitochondria. However, in contrast to Class I mutants, which were accumulated efficiently inside mitochondria, most of apocytochrome c containing the ⌬(Gln 25 -Lys 36 ) deletion remained protease-accessible on the mitochondrial surface. Expression of this mutant in diploid cells led to some inhibition of holo-iso-1-cytochrome c formation, although not to the same extent seen for the Class I mutants.
Since even normal apocytochrome c did not associate with mitochondria that lacked CCHL (Dumont et al., 1991), it seemed likely that binding the ⌬(Gln 25 -Lys 36 ) apocytochrome c on the mitochondrial surface would depend on CCHL. This was tested by measuring mitochondrial association of this mutant in a strain expressing a single copy of CYC3 (encoding CCHL), instead of the overexpressing strains used in the other subcellular localization study. The lower level of CCHL reduced the amount of mitochondrially associated cytochrome c from 40 to 3% of the total cellular level. Thus, interaction with CCHL was required to maintain even surface-exposed precursor in association with mitochondria.
His 27 appears to be the key residue in this region responsible for the accumulation of apocytochrome c in a protease-inaccessible compartment of mitochondria. Apocytochrome c containing a ⌬(Thr 28 -Lys 36 ) deletion accumulated in a protease-protected state with the same efficiency seen for Class I mutants. Furthermore, the single substitution His 27 3 Ala 27 exhibited a phenotype indistinguishable from deletion of the entire region from Gln 25 to Lys 36 .
Class V Deletion of His 48 -Tyr 57 Allows Holocytochrome c Formation-The ⌬(His 48 -Tyr 57 ) mutant produced holocytochrome c at 40% of the normal level. The deleted region corresponds to a loop in the three-dimensional structure. Thus, the structure appeared to be able to accommodate a rearrangement in which two residues that were in close proximity (Arg 47 and Thr 58 ) were directly joined instead of being connected by a loop. Two additional deletions in the corresponding region of iso-1cytochrome c were previously shown to allow formation of holocytochrome c. These two iso-1-cytochrome c deletions lacked the following residues (iso-2-cytochrome c numbering system): FIG. 5. Low temperature (؊196°C) spectrophotometric recordings of a series of diploid strains. The following representative strains contained the indicated levels of normal cytochrome c. A, residues 49 -62 (cyc1-453) (Sherman et al., 1975;Hampsey et al., 1988) and residues 52-59 (cyc1-817) (Fetrow et al., 1989). In addition, this region is absent in the S-type bacterial cytochromes c (Matsuura et al., 1982), including cytochrome c 551 , which lacks residues corresponding to positions 46 -68 of iso-2-cytochrome c.
Not all of the apocytochrome c containing the ⌬(His 48 -Tyr 57 ) deletion was converted into holocytochrome c. Approximately 30% of the immunologically detectable protein was found in the cytoplasm. Since strains expressing this mutant contained a normal overall cellular level of immunologically detectable cytochrome c, 70% of which was mitochondrial, there appeared to be additional apocytochrome c inside mitochondria. The existence of this population of mitochondrial apocytochrome c, which was presumably bound to CCHL, was consistent with the observed partial displacement of normal holo-iso-1-cytochrome c by the ⌬(His 48 -Tyr 57 ) mutant in diploid strains.

DISCUSSION
The sequences in cytochrome c which are responsible for directing the protein from its site of synthesis in the cytoplasm to its functional site on the outer surface of the mitochondrial inner membrane have not been well characterized previously. A major difficulty in identifying such sequences has been the requirement for heme to be covalently attached to apocytochrome c to observe efficient import. This makes it difficult to distinguish sequence determinants that are important for heme attachment and concomitant folding of the protein from sequence determinants that are actually involved in subcellular targeting.
To circumvent these difficulties, we have made use of a previous observation that altered forms of apocytochrome c which are incapable of having heme covalently attached can be imported efficiently into mitochondria that have an abnormally high level of CCHL (Dumont et al., 1991). Determination of the subcellular distribution of apocytochrome c in such CCHLoverproducing strains has the following advantages for the identification of sequences involved in targeting. (i) The assay is performed in vivo with a system that is known to be capable of efficiently importing and maturing normal cytochrome c. (ii) Comparisons can be made among different forms of cytochrome c which are incapable of undergoing covalent heme attachment. (iii) The assay does not involve the creation of fusion proteins that could alter targeting. (iv) Apo-iso-2-cytochrome c is stable in the cytoplasm and mitochondria, allowing immunologic assay of the subcellular distribution of the protein.

Protein Folding and Heme Attachment
Deletions covering 75% of the sequence of iso-2-cytochrome c (Thr 28 -Lys 112 ) accumulated in mitochondria to the same level as cytochrome c containing substitutions only for the two cysteine residues, Cys 23 and Cys 26 , which are the sites of covalent heme attachment. However, with the exception of ⌬(His 48 -Tyr 57 ), where the deletion apparently happened to coincide with a dispensable loop in the three-dimensional structure of the protein, none of the deletion-containing apocytochromes c that were efficiently imported could be converted into holocytochrome c. This suggests that mutations that interfere with the native structure of the mature protein also render cytochrome c incapable of having heme covalently attached. Since apocytochrome c appears to be an unfolded protein (Babul and Stellwagen, 1972;Fisher et al., 1973), either CCHL recognizes native-like structures in the precursor, or protein folding is coupled to heme binding so that the formation of the thioether linkages cannot proceed unless the protein-heme product can adopt a stable native conformation.
In a previous analysis of single amino acid replacements at 16 different sites throughout iso-1-cytochrome c which caused deficiency of function, only replacements at Cys 23 , Cys 26 , and His 27 (iso-2-cytochrome c numbering system) completely prevented the formation of holocytochrome c (Hampsey et al., 1988). Thus, except for these three amino acid residues that comprise and adjoin the actual site of covalent heme attachment, it seems likely that no short motifs are required for the covalent attachment of heme. On the other hand, the failure of the numerous deletion-containing cytochromes c to be converted into holocytochrome c indicates that successful catalysis of heme attachment by CCHL requires the existence of an overall structure. Since most of the deletion-containing apocytochromes c accumulate efficiently inside mitochondria, structural distortions that prevent heme attachment do not necessarily prevent binding to CCHL, and furthermore, the actual site of heme attachment in cytochrome c may not be the same as the sequence determinants involved in the binding to heme attaching enzyme.
The failure of most of the deletion-containing alleles of iso-2cytochrome c to undergo covalent heme attachment is in sharp contrast to the report of Veloso et al. (1984) that partially purified CCHL is capable of attaching heme to a peptide corresponding to residues 1-25 of horse apocytochrome c. Either the efficiency of the reaction in vitro is very low, or the sequence requirement for CCHL action is much more stringent in vivo.

Sequences Involved in Import
Altered cytochromes c that do not accumulate in mitochondria in the presence of excess CCHL could be defective in any of the following aspects: (i) they could interact improperly with a hypothetical cytoplasmic machinery, failing to reach the mitochondrial surface; (ii) they could fail to undergo the correct initial interaction with a hypothetical mitochondrial receptor or import apparatus; (iii) they could be incapable of being translocated across the mitochondrial outer membrane; or (iv) they could fail to bind to CCHL. By combining subcellular fractionation with competition studies in diploid strains, we have distinguished the following four different types of sequence alterations with different effects on import.
1. The Amino Terminus-Deletion of the first 14 amino acids of the protein reduces the efficiency of association of precursor with mitochondria or CCHL but still allows inefficient conver- sion into holocytochrome c. The effect of the deletion is not sequence-dependent, since replacement of the missing residues with nine unrelated amino acids restores normal competence for import and conversion to holocytochrome c. The nonessential nature of amino acids in this region was indicated earlier by the recovery of more than 50 amino acid replacements and deletions in the amino-terminal region of iso-1-cytochrome c which did not affect function or production of holocytochrome c significantly. Only those shortened by six or more amino acid residues had diminished levels (Baim et al., 1985;Hampsey et al., 1988). Since normal iso-1-cytochrome c is four amino acids shorter at its amino terminus than iso-2 cytochrome c and since the two isozymes apparently share the same import pathway, a six-amino acid deletion in iso-1-cytochrome c is equivalent to a 10-amino acid deletion in iso-2-cytochrome c. Apparently, at least four amino acids of any type are required preceding Gly 15 for near normal production of holocytochrome c.
Wild type eukaryotic cytochromes c from numerous species vary considerably in their amino-terminal regions. Many of these proteins are nine amino acid residues shorter than iso-2-cytochrome c, and some are two amino acid residues longer than iso-2-cytochrome c (Hampsey et al., 1988;Moore and Pettigrew, 1990). In spite of this variation, the import and heme attachment processes of cytochrome c appear to be fundamentally the same in all eukaryotes. For example, horse cytochrome c, expressed under control of the normal yeast iso-1-cytochrome c promoter, was produced in yeast at a level corresponding to 66% of the normal abundance of iso-1-cytochrome c, even though the amino terminus of horse cytochrome c corresponds to Gly 10 of iso-2-cytochrome c (Hickey et al., 1991).
A minimal requirement for the length of the amino-terminal region of a cleavable mitochondrial matrix targeting signal fused to dihydrofolate reductase has been observed by Ungermann et al. (1994). In that case, the minimal length appeared to be required for the ATP-dependent interaction of the precursor with mitochondrial Hsp70. In the case of cytochrome c, the length could be required for stable interaction with the mitochondrial outer membrane, with CCHL, or with some yet-to-be identified factors. However, it is surprising that the length requirement for cytochrome c resides at the extreme amino terminus of the protein, since cytochrome c contains no cleavable targeting signal, and the other targeting signals that we have uncovered are located at least 15 residues from this end. This suggests either that there is nonspecific anchoring of the end of the chain in order to expose some more specific signals farther from the end or that the extreme amino terminus is involved in stabilizing a translocation-competent conformation. The amino-terminal region of apocytochrome c could also enhance membrane-active properties that would facilitate penetration across the mitochondrial outer membrane (Jordi et al., 1989b).
2. Region Encompassing Gly 15 -Arg 22 -The ⌬(Gly 15 -Arg 22 ) apocytochrome c did not associate with mitochondria, although expression of this mutant cytochrome c in a heterozygous diploid strain led to inhibition of formation of coexpressed normal holocytochrome c. Two possible explanations for this observation are: (i) that mutations in the Gly 15 -Gln 24 region lead to formation of a stalled complex between apocytochrome c and a cytoplasmic factor involved in import, thereby titrating the amount of factor available for import of normal cytochrome c; or (ii) that mutations in this region allow a loose association with mitochondrial components that can dissociate during subcellular fractionation but is still tight enough to allow competition in vivo for maturation of normal cytochrome c.
There is one previous report of a cytoplasmic factor involved in cytochrome c import in an in vitro system . Furthermore, a cytoplasmic factor called mitochondrial import stimulation factor from rat liver has been found to promote mitochondrial import of proteins other than cytochrome c. The ATPase activity of mitochondrial import stimulation factor was stimulated by the presence of unfolded mitochondrial precursors, including apocytochrome c but not holocytochrome c. This stimulation of ATPase was reversed by the presence of mitochondrial outer membranes, apparently because of binding of the mitochondrial import stimulation factor-precursor complex to sites on the membranes (Hachiya et al., 1994). Import of cytochrome c into mitochondria does not appear to depend on cytoplasmic heat shock-related proteins to the same extent as import of other mitochondrial proteins, since, in contrast to other mitochondrial precursors, apocytochrome c can be imported efficiently into isolated mitochondria from wheat germ translation systems, which are at least partially deficient in such heat shock-related proteins (Dumont et al., 1988;Murakami et al., 1988). There appear to be at least two sequence determinants in the region Gly 15 -Arg 22 , which affect mitochondrial accumulation, including Phe 19 and at least one residue in the region Gly 15 -Leu 18 . The equivalent of Phe 19 is found in the sequence of every known eukaryotic cytochrome c (see Hampsey et al., 1988), in agreement with the possible importance of this residue. The residues found at the equivalent of positions Ala 16 , Thr 17 , and Leu 18 vary considerably among eukaryotic species and have been mutated with little effect on cytochrome c function. Gly 15 is invariant among eukaryotic cytochromes c, suggesting that it could play an important role. However, replacements of Gly 15 by alanine, serine, cysteine, or aspartic acid residues allowed the formation of substantial levels of holocytochromes c which were at least partially functional (Hampsey et al., 1988;Auld and Pielak, 1991). Thus, deletion of the region Gly 15 -Leu 18 might lead to the loss of functionally redundant sequence elements or disrupt a local conformation that was not affected by replacement of any single amino acid residue. Sprinkle et al. (1990) found that substitutions of glutamine, glutamic acid, and asparagine for lysine residues at the equivalent of positions 14, 16, and 17 (iso-2-cytochrome c numbering system) in D. melanogaster cytochrome c led to loss of import. The lysine residues at positions 16 and 17 are not conserved in iso-2-cytochrome c. The third of these lysine residues, at position 14, is not indispensable based on the results of aminoterminal deletions described above. Nonetheless, our identification of a range of mutations over the amino-terminal 25% of the sequence of cytochrome c which lead to defects in mitochondrial accumulation targeting agrees with the identification by Sprinkle et al. (1990) of lesions in this region which are capable of preventing targeting. In contrast to these results, Nye and Scarpulla (1990b), assaying for targeting of an iso-1-cytochrome c-chloramphenicol acetyltransferase fusion protein, found that they could delete as much as the first 67 amino acids of the cytochrome c sequence without affecting targeting to mitochondria. The results reported in the present manuscript do not support the proposal by Nye and Scarpulla (1990b) that there is functional redundancy between amino and carboxyl regions of the protein.
Because single localized deletions were examined in our study, we have not ruled out the possibility that, in addition to the sequences we uncovered, additional alterations at two or more nonadjacent sites might lead to impaired mitochondrial accumulation.
3. Role of His 27 -His 27 serves as a ligand for the heme iron in native holocytochrome c and is conserved in all eukaryotic cytochromes c. Previously, alteration of this residue was shown to prevent mitochondrial accumulation of holocytochrome c, but the effects on import could not be distinguished from those on heme attachment (Dumont et al., 1988). In the present experiments, where import was examined in the absence of heme attachment, replacement or deletion of His 27 led to association of the altered apocytochrome c with mitochondria, but most of the mitochondrially associated precursor remained protease-accessible on the mitochondrial outer surface. The bound precursor appears to represent a form of the protein which is stalled in the import process, either because His 27 is required for interaction with translocation machinery of the mitochondrial outer membrane or because the histidine is involved in recognition of the precursor by CCHL. Since no binding to mitochondria is observed in the absence of CCHL (Dumont et al., 1991), simple failure to bind would not result in accumulation of the partially translocated form. Instead, the altered precursor must be entering either directly or indirectly into some aberrant interaction with CCHL.
The observation that apocytochrome c lacking His 27 fails to associate with mitochondria unless they contain high levels of CCHL confirms the role of CCHL in maintaining the stalled form of the precursor. The simplest explanation of this result is that the altered apocytochrome c adopts a membrane-spanning topology, rendering it accessible to external proteases at the same time as it is bound to CCHL inside the mitochondrial outer membrane. However, CCHL has been reported to be associated predominantly with the mitochondrial inner membrane, not the outer membrane (Dumont et al., 1991). There are a number of ways of reconciling these findings. (i) CCHL may not interact directly with the stalled precursor but, rather, act in an indirect way to maintain the association of the altered apocytochrome c with mitochondria. (ii) CCHL may transiently become associated with the outer membrane in vivo, or it may be localized to sites where the two membranes are in close contact. (iii) The apparent localization of CCHL to the inner membrane in isolated mitochondria may result from redistribution during fractionation.
4. The Carboxyl Terminus-We detect only a slight diminution in mitochondrial accumulation resulting from deletion of the nine amino acid residues at the carboxyl-terminal region of iso-2-cytochrome c. However, like most of the deletions studied, removal of the carboxyl terminus effectively blocks conversion into holocytochrome c. These results contradict earlier claims that critical targeting sequences reside at the carboxyl terminus of cytochrome c (Matsuura et al., 1981;Stuart et al., 1987). Such claims were based either on competition with a carboxylterminal peptide of mitochondrial import in an in vitro system, which could be subject to nonspecific interference at high peptide concentrations, or on studies with a mutant cytochrome c containing 27 unrelated amino acid residues fused to the carboxyl terminus, which could lead to spurious mistargeting. Furthermore, neither of these previous studies clearly distinguished effects on heme attachment from effects on import. On the other hand, our finding that the carboxyl terminus is not critical for mitochondrial accumulation of cytochrome c is in agreement with the two previous systematic studies of cytochrome c sequences involved in targeting (Nye and Scarpulla, 1990b;Sprinkle et al., 1990).
We have not resolved the question of whether the same regions of cytochrome c which we have determined to be necessary for mitochondrial accumulation are also sufficient for accumulation. Expression of amino-terminal regions of apocytochrome c as short fragments in vivo does not result in accumulation of sufficient quantities of protein for detection using an epitope tag, presumably because of rapid degradation (data not presented). On the other hand, fusion of apocytochrome c fragments to a larger carrier protein can lead to mistargeting (see Nye and Scarpulla, 1990a).

Pathway of Cytochrome c Import into Mitochondria
Based on the results presented above and on previous studies, we present in Fig. 6 a tentative model for cytochrome c targeting, emphasizing several of the following unresolved aspects of the process.
1. After synthesis in the cytoplasm, apocytochrome c may associate with a cytoplasmic factor. The existence of such a factor is suggested by our observation that expression of altered forms of apocytochrome c which do not bind or enter mitochondria inhibits the formation of the coexpressed normal cytochrome c.
2. Apocytochrome c, either alone or in conjunction with a cytoplasmic factor, associates with mitochondria and adopts a membrane-spanning topology. Upon becoming exposed to the intermembrane space, sequences near the amino terminus of apocytochrome c bind either directly to CCHL or to a CCHLassociated factor. Such a factor would normally be present in excess relative to CCHL, or its level would increase upon overexpression of CCHL. Both the amino and carboxyl termini exhibit membrane-active properties, although only amino-terminal fragments are able to translocate across the lipid bilayer (Zhou et al., 1988;Jordi et al., 1989aJordi et al., , 1989b. However, it is not currently known whether the pathway of import of cytochrome c involves a direct interaction with the lipid phase.
3. His 27 of apo-iso-2-cytochrome c interacts with CCHL or some other CCHL-associated component of the translocation machinery in order to allow the carboxyl-terminal portion of the protein to be translocated into the mitochondrial intermembrane space.
4. If CCHL has remained on the inner membrane, it associates with the precursor following diffusion across the intermembrane space.
5. Altered apocytochromes c that are unable to have heme covalently attached remain trapped in the intermembrane space in a dead end complex with CCHL. Normal precursors FIG. 6. Model for mitochondrial import of cytochrome c. Cytochrome c is synthesized in the cytoplasm as apocytochrome c (Apo), lacking heme. Apocytochrome c binds to a putative factor in the cytoplasm (Cytoplasmic factor). Class III mutants (Table IV) appear to interact with the hypothetical cytoplasmic factor but are unable to translocate across the mitochondrial outer membrane. Apocytochrome c diffuses reversibly across the mitochondrial outer membrane and binds to cytochrome c heme lyase (CCHL), which is located on the mitochondrial inner membrane facing the intermembrane space. Class IV mutants involving His 27 (Table IV) appear to bind to CCHL while remaining exposed to the outside surface of the mitochondria. Apocytochrome c can bind to CCHL even in the absence of heme attachment, as indicated by the Class I (Table IV) mutants. After heme attachment, catalyzed by CCHL, holocytochrome c (Holo) folds into its native globular state, is released from CCHL, and remains trapped in the intermembrane space.
are covalently linked to heme in a reaction catalyzed by CCHL. Folding to the native globular structure occurs concomitant with this reaction.
6. Finally, matured normal precursors are released into the intermembrane space. In their folded form, they cannot translocate back across the outer membrane (see Fig. 6).