Differential Ubiquitin-dependent Degradation of the Yeast Apo-cytochrome c Isozymes*

The yeast Saccharomyces cerevisiaecontains two forms of cytochrome c, iso-1- and iso-2-cytochrome c, which are encoded by the nuclear genesCYC1 and CYC7, respectively. The cytochromesc are synthesized in the cytosol, imported into mitochondria, and subsequently modified by the covalent attachment of heme through the action of cytochrome c heme lyase, which is encoded by CYC3. Apo-iso-2-cytochrome c but not apo-iso-1-cytochrome c was observed incyc3 − mutants. Furthermore, pulse-chase experiments previously demonstrated that the lack of apo-iso-1-cytochrome c was due to its rapid degradation. We report herein that this degradation of apo-iso-1-cytochromec is dependent on ubiquitination and on the action of the proteasome. Diminished degradation of apo-iso-1-cytochromec was observed in pre2-2 and pre1-1mutants having altered proteasome subunits; in ubc1,ubc4, and ubc5 strains lacking one or more of the ubiquitin-conjugating enzymes; and in strains blocked in multi-ubiquitination by overproduction of the abnormal ubiquitin-K48R ubiquitin. In addition, we have used epitope-tagged ubiquitin to demonstrate that apo-iso-1-cytochrome c but not apo-iso-2-cytochrome c is ubiquitinated. Furthermore, the degradation of apo-iso-1-cytochrome c was diminished when the N-terminal region was replaced with the N-terminal region of apo-iso-2-cytochrome c, indicating that this region may be the target for degradation. We suggest that ubiquitin-dependent degradation of apo-iso-1-cytochromec is part of the regulatory process controlling the preferential expression of the iso-cytochromes c.

The yeast Saccharomyces cerevisiae contains two forms of cytochrome c, iso-1-cytochrome c and iso-2-cytochrome c, which are encoded by the nuclear genes CYC1 and CYC7, respectively. Iso-1-cytochrome c and iso-2-cytochrome c, which are 80% identical, normally comprise 95 and 5% of total cytochrome c, respectively, in aerobically grown, derepressed cells (1). Both iso-cytochromes c are synthesized in the cytosol, as apo-cytochrome c, and subsequently translocated into the mitochondria. Heme is covalently attached to the apo-cytochromes c by cytochrome c heme lyase (CCHL), 1 which is en-coded by the gene CYC3, resulting in the formation of the mature holo-cytochromes c (2). Import of the apo-cytochromes c is dependent on the action of CCHL, and cyc3-⌬ mutants, lacking CCHL, accumulate apo-cytochromes c in the cytosol (3). However, apo-1 (apo-iso-1-cytochrome c) is not detected in cyc3-⌬ mutants, whereas apo-2 (apo-iso-2-cytochrome c) is present at the corresponding level of holo-iso-2-cytochrome c in related CYC3 ϩ strains. Dumont et al. (4) demonstrated with pulse-chase experiments that unimported apo-1 is rapidly degraded.
In this study, we have demonstrated that apo-1 degradation requiresfunctionalproteosomesandismediatedbytheubiquitindependent pathway. The ubiquitin system is the major pathway for targeting and selective degradation of proteins in the cytosol and nucleus of eukaryotes (5)(6)(7)(8)(9)(10)(11). Ubiquitin is a 76residue protein that is joined reversibly to protein substrates by linkage between the ␣-carboxyl group of ubiquitin and ⑀amino groups of certain lysine residues of the acceptor proteins through an isopeptide bond in a multistep reaction. Ubiquitin is first activated through the formation of of a thiol ester bond of C-terminal glycine by a ubiquitin-activating enzyme. Ubiquitin then forms a thiol ester bond with a second protein, one of the ubiquitin-conjugating enzymes, E2. The transfer of ubiquitin to protein substrates by an E2 is done with or without the participation of a ubiquitin-protein ligase, E3. Additional ubiquitin molecules are usually added to the substrate by the same enzyme cascade, in which ubiquitin moieties are linked to the lysine 48 of each previously conjugated ubiquitin molecule until multi-ubiquitin chains are formed, which facilitates a more rapid degradation by the action of the 26 S proteosome or sometimes by the action of vacuoles.
The conclusions on the degradation of apo-1 and the lack of degradation of apo-2 were based on the use of defined conditional mutants of the proteosome, on mutants lacking one of more of the E2-conjugating enzymes, on the inhibition of multiubiquitination by overexpression of the altered ubiquitin Ub-K48R, and on the identification of ubiquitinated forms of apo-1 but not apo-2. Because multi-ubiquitinated proteins also can be degraded by the action of vacuoles (or lysosomes), we have also examined pep4-⌬ mutants, which are defective in a subset of vacuolar proteinases and which are reduced in the degradation of numerous proteins.

MATERIALS AND METHODS
Yeast Strains and Plasmids-The yeast strains used in this study are listed in Table I. Strain WCG4-11/22a, obtained from Dr. Dieter H. Wolf, along with our laboratory strains, were used to construct pre2-2 cyc7-⌬ and pre1-1 cyc7-⌬ strains. CYC3 ϩ genes were disrupted by transforming ura3-52 strains with a 3-kilobase pair EcoRI-XhoI linear fragment, which was derived from the plasmid pAB609 and which corresponds to a segment encompassing the CYC3 gene in which the coding region was replaced by the URA3 gene. Strains B-10105 through B-10108, B-10109 through B-10112, and B-101155 through B-10158 represent three tetrads resulting from a cross of B-7553 and B-10104. Strains B-10120 through B-10124 were prepared by disrupting one or more of the UBC1, UBC4, and UBC5 genes in strain B-8107 with, respectively, the following plasmids treated with the following endonucleases: pS-UBC1, HindIII; pUBC4HIS3, SphI-EcoRI; and pUB-CLEU2, SacI-PstI (see Refs. 12 and 13). YEp96 (also denoted p[P CUP1 -Ub]), YEp110 (also denoted p[P CUP1 -Ub-K48R), and YEp112 (also denoted p[P CUP1 -HA-Ub]) are 2-m plasmids that encode normal (Ub), Ub-K48R, and HA-Ub, respectively, under the control of the copperinducible CUP1 promoter, as described by Hochstrasser et al. (14). Ub denotes normal ubiquitin; Ub-K48R denotes Ub with a K48R replacement; and HA-Ub denotes ubiquitin with a YPYDVPDYA epitope from hemagglutinin (HA) of influenza virus attached to the N terminus of Ub. Disruptions and plasmids were introduced by conventional transformation procedures.
The cyc1-851, cyc1-864, cyc1-867, and cyc1-885 alleles were generated by transforming strain   (15) directly with synthetic oligonucleotides, as described by Yamamoto et al. (16). Subsequently, the CYC3 gene was disrupted in each of the strains as described above, resulting in strains B-10627 to B-10630 (Table I). Strain B-10631 is isogenic to this series.
PCR and Sequencing-The deletions or disruptions cyc7-67, cyc7-⌬::CYH2, cyc3-⌬::URA3, ubc1-⌬::LEU2, ubc4-⌬::HIS3,. and ubc5-⌬::LEU2 were identified or confirmed from the size of the pertinent PCR fragment, as compared with size of the fragment obtained with the wild type allele, as described previously (17). The pre2-2 and pre1-1 mutant alleles were identified by sequencing the appropriate PCR product encompassing nucleotide positions 974 and 595, respectively, using the dideoxy chain termination method (18)  Media and Growth Conditions-Pulse-chase and Western blot analyses of apo-1 from strains involving pre2-2 or various ubc-⌬ strains were carried out with yeast grown to stationary phase in 15 ml of YPD (1% Bacto-yeast extract, 2% Bacto-peptone, and 2% dextrose). Pulse-chase and Western blot analyses of apo-1 from strains containing p[P CUP1 -Ub], p[P CUP1 -Ub-K48R), and p[P CUP1 -HA-Ub] plasmids were carried out with yeast grown to stationary phase in 15 ml of SD (0.67% Difco yeast nitrogen base and 2% dextrose) containing the required supplements. Cells were collected by centrifugation at 5000 ϫ g for 1 min, washed once in sterile distilled water, and resuspended in the same volume of fresh SD medium containing the required supplements and 100 M CuSO 4 ; the culture was incubated for an additional 5-8 h. All yeast strains were grown at 30°C except for pre2-2 strains and related control strains, which were grown at 23°C.
Antibodies-Anti-cytochrome c antibody was prepared essentially as  Matner and Sherman (19). Antibodies to the HA epitope (12CA5) were obtained from Berkeley Antibody Co. Inc. Pulse-Chase Labeling and Immunoprecipitation of Cytochrome c-Yeast strains grown as described above were collected by centrifugation, resuspended in 12 ml of semisynthetic sulfate-free medium (20), lacking yeast extract and containing 2% raffinose, 0.1% glucose, and, if required, 100 M CuSO 4 . After 1 h of incubation, [ 35 S]methionine (1300 Ci/mmol, Amersham Corp.) was added to a concentration of 0.125 Ci/ml, and the cells were incubated for an additional 15 min, followed by a chase in 30 mM (final concentration) methionine. Identical volumes of cells (1.5 ml) were taken at the times indicated, collected by centrifugation at 5000 ϫ g for 1 min, washed once in sterile distilled water, and resuspended in 0.4 ml of sterile distilled water. Cells were lysed in an equal volume of 0.4 M NaOH containing 1.7% 2-mercaptoethanol. After incubating the cell suspension on ice for 10 min, 0.2 ml of 100% trichloroacetic acid was added to precipitate the cell and protein mixture. After 10 min on ice, each sample was centrifuged at 15,000 ϫ g for 10 min, the supernatant was removed, and the pellet was washed in 1 ml of acetone and recentrifuged at 15000 ϫ g for 10 min. The pellet was washed once more in acetone, allowed to dry, and subsequently solubilized by boiling for 5 min in 2 ϫ loading buffer (4% SDS, 0.125 M Tris-HCl, pH 6.8, 1 mM EDTA, 20% glycerol, 10% 2-mercaptoethanol, 2 mM phenylmethylsulfonyl fluoride, and 0.002% bromphenol blue). The solubilized samples were centrifuged at 15,000 ϫ g for 5 min, and the pellets discarded. Apo-1 was immunoprecipitated as described previously for holo-1-cytochrome c (17). Proteins were separated on a 10% SDS-polyacrylamide gel (21); the gels were dried; apo-1 was visualized by autoradiography using Kodak BIOMAX MR imaging film, and the intensity of the bands was quantified by densitometry.
Western Blot Analysis-Yeast were grown in the appropriate conditions, and proteins were either solubilized or immunoprecipitated and separated on a 10% SDS-polyacrylamide gel, as described above. Proteins were transferred to nitrocellulose by standard procedures (22). Nonspecific protein binding was blocked by 2 h of incubation in 10% nonfat dried milk or 5% fetal calf serum (Life Technologies, Inc). The filters were incubated for 3 h with either anti-HA antibody or anticytochrome c antibody at concentrations of 1:5000 or 1:1000, respectively. The filters were visualized using either anti-mouse or anti-goat horseradish peroxidase or alkaline phosphatase development reagents (Bio-Rad).

RESULTS
Determination of the steady-state level and turnover of apo-1. As described above, apo-1 is rapidly degraded in the cytoplasm when import into the mitochondria and maturation is blocked by the deletion of CYC3, which encodes CCHL. In contrast, apo-2 is not degraded under these conditions (4). We have investigated the pathway of degradation by examining whether or not the level of apo-1 is diminished in strains either having alterations or deficiencies in the following components: iso-1-cytochrome c having the stabilizing global suppressor N52I; defective proteasomal subunits encoded by pre2-2 and pre1-1; protease-defective vacuole due to the pep4-⌬ mutation; the ubiquitin Ub-K48R, which prevents multi-ubiquitination; and the lack of one or more of the E2-conjugating enzymes Ubc1p, Ubc4p, and Ubc5p. These studies were carried out with strains lacking CCHL due to the cyc3-⌬::URA3 deletion, and lacking apo-2 due to either the cyc7-67 or cyc7-⌬::CYH2 deletion (denoted cyc7-⌬ below). The steady-state levels of apo-1 in these CYC1 ϩ cyc3-⌬ cyc7-⌬ derivatives were determined by Western blot analysis, and the turnover of apo-1 was determined by pulse-chase experiments. Finally, we investigated why apo-1 is targeted for degradation through the ubiquitindependent pathway, while apo-2 is not, by examining apo-1 with various altered N-terminal sequences, including the Nterminal region corresponding to apo-2.
Apo-1 Degradation Is Not Diminished by the N52I Global Suppressor-Previous work has established that the N52I replacement in iso-1-cytochrome c acts as a "global" suppressor, counteracting defects located at at least three different sites in the molecule (23, 24). G6S, G27S, and H33P are three replacements that result in nonfunctional iso-1-cytochrome c, whose function is at least partially restored by the single amino acid replacement N52I in vivo. Furthermore, the N52I iso-1-cytochrome c has an enhanced stability in vivo (25). In addition, the N52I replacement causes an unprecedented increase in the thermodynamic stability of the protein in vitro (26). Examination of various replacements of Asn-52 suggested that hydrophobic interactions are the main factor involved in stabilizing the N52I iso-1-cytochrome c (26,27).
We have examined N52I apo-1 to determine if the degradation may be dependent on the overall thermostability of the protein. The result of the steady-state levels indicated that the normal apo-1 and the N52I apo-1 in strains B-8107 and B-8108, respectively, were equally degraded (data not presented), indicating that the degradation may not be dependent on the overall stability. Thus, the difference in apo-1 and apo-2 degradation cannot be simply explained by the difference in stability. However, we have not directly demonstrated the N52I replacement stabilizes apo-cytochrome c, as it does for holo-cytochrome c.
Functional Proteasomes Are Required for Degradation of Apo-1-We have established that functional proteasomes are required for degradation of apo-1 by examining the levels of apo-1 in pre2-2 and pre1-1 mutants having defective proteasomal subunits (28,29). The pre2-2 mutation corresponds to a T 3 C nucleotide substitution at nucleotide position 974, causing an Ala 3 Val replacement in the Pre2p subunit, whereas the pre1-1 mutation corresponds to a C 3 T nucleotide substitution at nucleotide position 595, causing a Ser 3 Phe replacement in the Pre1p subunit (28,29).
Inhibition of Ubiquitination Decreases Degradation of Apo-1-We have established that ubiquitination is required for degradation of apo-1 by examining the levels of apo-1 in a series of isogenic strains lacking one or more of the E2-conjugating enzymes Ubc1p, Ubc4p, and Ubc5p, which are generally required for the bulk turnover of short-lived proteins and which functionally overlap (12,13). The UBC1 ϩ , UBC4 ϩ , and UBC5 ϩ genes were disrupted in strain B-8107 (CYC1 ϩ cyc7-⌬ cyc3-⌬), either singly or in pairs. Pulse-chase experiments with cells that were approaching the stationary phase of growth, presented in Fig. 3, indicated that the degradation of apo-1 was diminished to a small degree in each of the ubc4-⌬ and ubc5-⌬ single mutants and that this diminution was more pronounced in the ubc1-⌬, and the ubc1-⌬ ubc4-⌬ and ubc4-⌬ ubc5-⌬ double mutants.
In addition, multi-ubiquitination was inhibited by overexpressing Ub-K48R, an altered ubiquitin in which the normal lysine 48 was replaced by arginine. Ub-K48R ubiquitin can still conjugate to proteins but is unable to function as an acceptor for multi-ubiquitination (37,38). B-8107 (CYC1 ϩ cyc7-⌬ cyc3-⌬) was transformed with the plasmids p[P CUP1 -Ub] and p[P CUP1 -Ub-K48R], which have the normal Ub and Ub-K48R ubiquitin, respectively, under control of the copper-inducible CUP1 promoter, as described by Hochstrasser et al. (14). Pulse-chase analysis, presented in Fig. 4, demonstrated that degradation is diminished in the presence of Ub-K48R. The slight difference in apo-1 degradation rates, presented in Figs. 3 and 4, presumably reflects the difference in the media and the fact that ubiquitin was overproduced in only experiments described in Fig. 3.
These results with the ubc1-⌬, ubc4-⌬ and ubc5-⌬ mutants and the overexpressed Ub-K48R ubiquitin established that ubiquitination is involved in the degradation of apo-1.
Identification of Ubiquitinated Forms of Apo-1-Ubiquitination of proteins has been detected by overexpressing either ubiquitin or epitope-tagged ubiquitin (14,39). We have demonstrated that apo-1 is ubiquitinated by overexpressing both normal ubiquitin and a HA epitope-tagged ubiquitin under control of the copper-inducible CUP1 promoter, as described by Hochstrasser et al. (14).
single or multi-copies of ubiquitin were expressed. This was anticipated because of the lack of degradation of apo-2 as described previously (4). Thus, apo-2 is not a substrate for ubiquitination but does serve as a specific control for such immunoprecipitation experiments.
In a similar experiment, the isogenic strains B-6748 (cyc1-⌬ cyc7-⌬) and B-8107 (CYC1 ϩ cyc7-⌬ cyc3-⌬) were transformed with the p[P CUP1 -HA-Ub] plasmid encoding ubiquitin with an HA epitope. The proteins were immunoprecipitated with anticytochrome c antibody, separated by electrophoresis, and were probed with anti-HA antibody, or anti-cytochrome c antibody, as shown in Fig. 6. The position of unconjugated apo-1 was revealed on one side of the gel (lanes [5][6][7][8], by probing with anti-cytochrome c, whereas the position of the ubiquitinated apo-1 was revealed on the other side (lanes 1-4), by probing with anti-HA antibody, therefore identical samples are in lanes   1 and 5, lanes 2 and 6, lanes 3 and 7, and lanes 4 and 8. The HA-tagged ubiquitinated species of apo-1 is clearly seen with proteins from strain B-10128 (CYC1 ϩ cyc7-⌬ cyc3-⌬ p[P CUP1 -HA-Ub]) (lane 2) but not with proteins from the control strains (lanes 1, 3, and 4). The size of the ubiquitinated apo-1 was approximately 8 kDa larger than the size of unconjugated apo-1, suggesting that apo-1 was mono-ubiquitinated. However, determination of the size and number of ubiquitin polypeptides by electrophoretic positions is known to be inaccurate (40). Identical patterns were observed with similar experiments repeated on numerous occasions, with various dilutions of the anti-HA antibody, and with pre2-2 strains that produce higher levels of apo-1 (results not presented).
banding pattern indicative of multi-ubiquitinated forms was observed.
The N-terminal Sequence May Target Apo-1 for Ubiquitination and Degradation-As shown in Fig. 7, Western blot analysis was used to determine the steady-state levels of the following apo-cytochromes c: TEFKAGSAKK-, CYC1 ϩ ; AKESTGFKPGSAKK-, CYC7 ϩ ; AKESTGFKPGSAKK-, cyc-  These results revealed that the degradation of apo-1 was diminished when the the N-terminal region was replaced with the N-terminal region of apo-2 (cyc1-885). The cyc1-885 apo-1 was approximately at the same level as CYC1 ϩ apo-1 in pre2-2 (Fig.  1). Furthermore, the cyc1-864 and cyc1-867 apo-1-cytochromes c, but not the cyc1-851 apo-1, also showed slight increases, suggesting that the normal the N-terminal region of apo-1 is at least in part the target for ubiquitin-dependent degradation. DISCUSSION We have established that turnover of apo-1 requires functional proteasomes and that this degradation is mediated by the ubiquitin pathway. Diminished degradation of apo-1 occurred in pre2-2 and pre1-1 strains having altered proteasome subunits; in ubc1, ubc4, and ubc5 strains lacking one or more of the ubiquitin conjugating enzymes; and in strains blocked in ubiquitination by overproduction of the abnormal Ub-K48R ubiquitin. In addition, overproduction of normal Ub and epitope-tagged HA-Ub ubiquitin was used to demonstrate that apo-1 is ubiquitinated. Furthermore, the inhibition of degradation by Ub-K48R ubiquitin and the banding pattern of B-10130 (Fig. 5), representing multple conjugates of apo-1, indicates that multi-ubiquitination may be the active form.
Because the pep4-⌬ mutation had no effect on apo-1 degradation, the vacuole does not appear to play a role in the degradation. Also, the N52I global suppressor did not protect apo-1 from degradation. In contrast, the N52I suppressor protects certain labile forms of holo-iso-1-cytochrome c from degradation that occurs within mitochondria. 2 The ability to rapidly degrade certain proteins has been shown to be important in metabolic regulation (41). Studies of the protein degradation by ubiquitination and subsequently by the proteosome have suggested that important regulatory proteins, as well as abnormal proteins, are targeted for degradation (12,14,42,43). This study on apo-1 provides the first example of a nuclear-encoded mitochondrial protein that is degraded by the ubiquitin-dependent pathway. Furthermore, we suggest that the differential degradation of apo-1 and apo-2 is part of the regulatory process controlling the preferential expression of the iso-cytochromes c.
Under partial anaerobiosis or glucose repression, the absolute amount of iso-1-cytochrome c diminishes but the absolute amount of iso-2-cytochrome c increases, and the relative proportion of iso-2-cytochrome c can be greater than iso-1-cytochrome c (44). Even though transcription is by far the major means for the differential expression of the iso-cytochromes c, we suggest that the differential degradation of apo-1 and apo-2 also may play a role. During partial anaerobiosis and repression, the levels of heme and possibly CCHL are low, leading to the accumulation of apo-cytochrome c. Because iso-2-cytochrome c is apparently required under these physiological conditions, it is to the advantage of the cells to preferentially degrade apo-1, resulting in a still higher proportion of apo-2  4 and 8). The samples in lanes 1-4 were probed with the anti-HA antibody (12CA4), whereas the samples in lanes 5-8 were probed with the anti-cytochrome c antibody to confirm the presence and position of apo-1. Therefore, identical samples are in the following: lanes 1 and 5, lanes 2 and 6, lanes 3 and 7, and lanes 4 and 8. Cells were grown to stationary phase at 30°C in SD medium, transferred to fresh SD medium containing 100 M CuSO 4 , and incubated for 5-8 h. Subsequently, the proteins were solubilized, separated by SDS-PAGE in a 10% gel, transferred to nitrocellulose, and probed by the pertinent antibody. The arrows indicate the positions of ubiquitinated apo-1 (lanes 1-4), and apo-1 (lanes [5][6][7][8]. The position of the molecular weight markers are indicated in the middle of the figure. CYC1 ϩ and CYC7 ϩ encode normal iso-1 and iso-2, respectively. The abnormal residues in apo- 1 (cyc1-885, cyc1-851, cyc1-867, and cyc1-864) are underlined. The increased level of the cyc1-885 apo-1 containing the N-terminal sequence of apo-2 suggests that the N-terminal region of apo-1 may be, at least in part, the target for ubiquitin-dependent degradation. which, after import, results in a higher proportion of iso-2cytochrome c.
This view of preferential degradation of apo-1 is consistent with observing higher proportions of iso-2-cytochrome c in mutants which are partially deficient in mitochondrial import, including cyc2 mutants and certain cyc3 mutants having partial deficiencies of CCHL (1,4,45).
One major question concerns the identification of the signals for degradation that is present in apo-1 but absent in iso-2. In general, the recognition of protein substrates for ubiquitin-dependent degradation is complex and has become increasingly diverse, as a multitude of different proteins targeted for degradation have been identified. Degradation by the ubiquitin system apparently requires at least two sites, a recognition site for binding a E2-conjugating or a E3 ligase protein and an acceptor site encompassing a lysine residue for ubiquitination. The two sites can be in close proximity or separated and can even reside in different subunits of an oligomeric protein (46). The diversity of the signals may be due in part to the numerous E2, Ubc enzymes which are encoded by a multigene family containing at least 13 members in yeast (10).
Several types of degradation signals have been uncovered in screens with artificial sequences that destabilize reporter proteins (47,48). One extensively studied degradation signal involves the "N-end rule pathway" (48), also denoted class I signals (47), in which Ubc2p and Ubr1p cooperate in the degradation of proteins, based on the recognition of signals at the N termini of polypeptide chains. Class II signals, also uncovered in a screen of artificial sequences, contain amphipathic ␣-helix structures and require Ubc6p, Ubc7p, and either Ubc4p or Ubc5p (47). Class III signals are short linear sequences of aliphatic residues, such as isoleucine and leucine, and require Ubc4p and Ubc5p.
Critical elements for degradation of naturally occurring proteins have been identified and characterized by site-directed alterations, including the following: c-Mos involving Pro-2 and dephosphorylation of Ser-3 (49); cyclins involving a degenerate sequence, RXALGXIXN, termed the "destruction box" (11, 50 -52); c-Jun involving residues 31-57 and 224 -331 of the protein (53); and Mat␣2p repressor, which contains two degradation signals, one between 53 and 67 and a second between 136 and 140 (43). In addition, Rogers et al. (54) postulated that short lived proteins contained a region enriched in residues of proline, glutamic acid, serine, and threonine (PEST region).
An examination of the apo-1 and apo-2 sequences, which are approximately 80% identical, does not reveal any obvious differences that can explain the differential ubiquitin-dependent degradation. All cytochromes c are lysine-rich, including apo-1 and apo-2, which contain 16 and 17 lysine residues, respectively. Apo-1 contains lysine residues at positions 27, 59, and 104, which are absent in apo-2, whereas apo-2 contains lysine residues at positions 49, 63, and 108, which are absent in apo-1. One major difference between the iso-cytochromes c is that apo-2 contains four additional amino acid residues at the N terminus, Ala-Lys-Glu-Ser. We have examined the possible role of the N-terminal region by examining the levels of apo-1 of four altered forms presented in in Fig. 7. The increased levels in the three of the four altered forms, cyc1-885, cyc1-864, and cyc1-867 apo-1-cytochromes c, but not the cyc1-851 apo-1, indicates that N-terminal region of apo-1 is at least in part involved in the ubiquitin-dependent degradation. The cyc1-885 apo-1, containing an altered N-terminal region corresponding to the normal apo-2 N-terminal region, exhibited the greatest reduction in degradation (Fig. 7). Nevertheless, it should be noted that the level of cyc1-885 apo-1 is approximately equivalent to the level of apo-2 in the CYC7 ϩ strain, that holo-iso-2-cytochrome c comprises only approximately 5% of the holoiso-1-cytochrome c level, and that the level of apo-2 in cyc3-⌬ strains is approximately equivalent to the level of holo-iso-2cytochrome c in CYC3 ϩ strains. Thus, replacing the N-terminal region of apo-1 with the N-terminal region of apo-2 resulted in only a marginal protection against degradation. The sequences responsible for the differential ubiquitin-dependent degradation, in conjunction to the N-terminal region, remains to be determined.