Degradation of cytochrome oxidase subunits in mutants of yeast lacking cytochrome c and suppression of the degradation by mutation of yme1.

We have confirmed by spectral analysis that cytochrome oxidase is not present in strains of the yeast Saccharomyces cerevisiae having a primary deficiency in cytochrome c, and we have demonstrated by immunological procedures that such strains lack the mitochondrially encoded subunits I, II, and III of cytochrome oxidase. Furthermore, pulse-chase experiments demonstrated that subunit II is rapidly degraded in vivo. This degradation can be at least partially suppressed by disruption of the nuclear gene YME1, which encodes a putative ATP-Zn-dependent protease. We suggest that the cytochrome oxidase subunits are not properly assembled in the absence of cytochrome c, and that Yme1 and possibly other proteases degrade the unassembled mitochondrial-encoded subunits of cytochrome oxidase.

S. cerevisiae contains two isoforms of cytochrome c, iso-1-and iso-2-cytochrome c, which are encoded by the nuclear genes, CYC1 and CYC7, respectively (12,13). Strains completely deficient in cytochrome c can be produced either by cyc1 Ϫ cyc7 Ϫ double mutations (13) or by mutation of the CYC3 gene that encodes heme lyase, which catalyzes the covalent attachment of the heme group (14,15). Spectral examination of intact cells revealed that such mutants lacking cytochrome c are also deficient in cytochrome oxidase, presumably as a secondary effect of the cytochrome c deficiency (13,15,16). Also, cytochrome oxidase is exceedingly sensitive to glucose repression in mutants having trace amounts of cytochrome c, as found with certain "leaky" cyc3 mutations (14), and in mutants having low levels of function, as found with certain cyc1 Ϫ missense mutations. The lack or diminished levels of cytochrome oxidase was also observed in mutants of Neurospora crassa deficient in cytochrome c (17)(18)(19).
In this study, we have used immunological procedures to demonstrate that cytochrome c-deficient cyc1 Ϫ cyc7 Ϫ mutants lack the mitochondrial-encoded cytochrome oxidase subunits I, II, and III and have reduced amounts of the nuclear-encoded subunits IV, V, VI, and VIa. In addition, pulse-chase experiments demonstrated that subunit II is rapidly degraded in vivo.
Furthermore, a number of genes encoding, or presumably encoding, mitochondrial proteases were disrupted, and the levels of the cytochrome oxidase subunits were examined. If these proteases are responsible for the degradation of the cytochrome oxidase subunits, the disruptions should act as suppressors in cytochrome c-deficient strains.
The nuclear gene PIM1 encodes an ATP-dependent protease located in the mitochondrial matrix (20,21). Because Pim1 is in the mitochondrial matrix and cytochrome oxidase is in the intermembrane space, and because pim1-⌬ mutants become Ϫ , Pim1 was not expected to be the protease acting on the cytochrome oxidase subunits. On the other hand, YTA10 (22), also denoted AFG3 (23), encodes a protease that acts on incompletely synthesized polypeptides in the mitochondrial inner membrane (24). YME1 (25) and RCA1 (26), also denoted YTA11 and YTA12, respectively (22), encode putative proteases related to Afg3 and represent members of a family of ATPases similar to proposed proteolytic complexes in Escherichia coli and yeast. However, afg3-⌬, and rca1-⌬ disruptions are deficient in the cytochrome oxidase subunits I, II, and III.
Only the yme1-⌬ disruption, but not pim1-⌬, afg3-⌬, or rca1-⌬, increased the level of the cytochrome oxidase subunits II and III in cytochrome c-deficient strains. These results suggest that the cytochrome oxidase subunits are not properly assembled in the absence of cytochrome c, and that Yme1 and possible other proteases degrade the unassembled subunits I, II, and III.
This degradation of cytochrome oxidase subunits in strains lacking cytochrome c at least superficially resembles the degradation of labile forms of cytochrome c in strains lacking either cytochrome c 1 or cytochrome oxidase, its physiological partners (27).

MATERIALS AND METHODS
Genetic Nomenclature and Yeast Strains-The symbols CYC1 ϩ and CYC7 ϩ denote the wild-type nuclear genes encoding, respectively, iso-1-cytochrome c and iso-2-cytochrome c. The cyc1-1011 symbol corresponds to a frameshift-TAA mutation at codon 85, resulting in a complete deficiency of iso-1-cytochrome c function; whereas cyc7-67 denotes a 400-base pair deletion, resulting in a complete deficiency of iso-2cytochrome c function. The cyc1-1011 and cyc7-67 symbols are abbreviated in this paper as cyc1-⌬ and cyc7-⌬, respectively. Thus, cyc1-⌬ cyc7-⌬ strains are completely deficient in cytochrome c. YME1 ϩ and yme1-⌬ denote, respectively, the wild-type and disrupted gene encoding a putative mitochondrial Zn 2ϩ -ATP-dependent protease. Similar genetic symbols are used to denote the other alleles of the putative proteases. ϩ denotes the normal mitochondrial genome, whereas Ϫ denotes mutants completely or partially deficient in mitochondrial DNA and thus are unable to carry out mitochondrial protein synthesis, resulting in deficiencies in cytochrome oxidase subunits I, II, and III, as well as cytochrome b and other mitochondrial-encoded proteins.
Low Temperature Spectrophotometric Recordings of Intact Cells-Total amounts of cytochromes a⅐a 3 , b, c, and c 1 were quantitatively estimated by absorbance recordings of intact cells at Ϫ196°C, with an Aviv Model 14 spectrophotometer as described previously (32).
Quantitation of Cytochrome Oxidase Subunits-Yeast were grown to stationary phase in 15 ml of YPD (1% yeast extract, 2% Bacto-peptone, 2% dextrose); the cells were collected by centrifugation at 5000 ϫ g for 1 min and resuspended in 4 ml of cold sterile distilled water. Cells were lysed essentially by the procedure of Yaffe and Schatz (34). The cells were lysed in an equal volume of 0.4 M NaOH containing 1.7% 2-mercaptoethanol. After incubating the cell suspension for 10 min on ice, 200 l 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 15,000 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 Tris⅐HCl, pH 6.8, 1 mM EDTA, 20% glycerol, 10% 2-mercaptoethanol, 2 mM phenylmethylsulfonyl fluoride, 0.002% bromphenol blue). For comparison, mitochondrial extracts were also prepared as described (35) and solubilized in the same manner. These solubilized samples were centrifuged at 15,000 ϫ g for 5 min, and the pellets were discarded. The levels of each of the subunits were estimated by loading various stepwise dilutions onto a 10% SDS-polyacrylamide gel (36) and comparing the intensities of the bands to those from the normal control. Protein was transferred to nitrocellulose by standard procedures (37). Nonspecific protein binding was blocked with a 2-h incubation with 5% fetal calf serum (Life Technologies, Inc.). The filters were incubated for 3 h with primary antibody at the following dilutions: CoxI (20 ng/ml), CoxII (1:2000), CoxIII (0.5 g/ml), CoxIV (1:500), CoxV (1:400), CoxVI (1:200), and CoxVIa (1:100). The filters were visualized using either anti-mouse anti-goat horseradish peroxidase or alkaline phosphatase development reagents (Bio-Rad).
Pulse-Chase Labeling of Cells and Immunoprecipitation of CoxII-Yeast strains were grown as described above and resuspended in 12 ml of semisynthetic sulfate-free medium (38) lacking yeast extract and containing 2% raffinose, 0.1% glucose, and 1 mg/ml cyclohexamide. After the cells were incubated for 15 min at 30°C, [ 35 S]methionine (1300 Ci/mmol, Amersham) 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 2 ml of cells were taken at the indicated times, and the cells were lysed and immunoprecipitated by the same method previously described for cytochrome c (27), except that 1 ml of polyclonal antiserum to CoxII was added to each sample. Proteins were separated as described above, the gels were dried, and CoxII was visualized by autoradiography.

RESULTS AND DISCUSSION
Cytochrome Levels in Vivo-The levels of the cytochromes a⅐a 3 , b, c 1 , and c in the isogenic series of strains were determined by low temperature spectrophotometric recordings of intact cells (Fig. 1). Strain B-8514 (curve A) (CYC1 ϩ cyc7-⌬ ϩ ) shows a normal complement of cytochromes a⅐a 3 , b, c 1 , and c. (The cyc7-⌬ defect is normally not manifested in the presence of CYC1 ϩ because iso-2-cytochrome c contributes to only approximately 5% of the total amount of cytochrome c (16).) Strain B-8123 (curve C) (cyc1-⌬ cyc7-⌬ ϩ ), which contains a primary deficiency in cytochrome c, had no detectable amount of cytochrome oxidase as indicated by the absence of a cytochrome a⅐a 3 peak at 602.5 nm. The cytochrome content in strain B-9614 (curve B) (CYC1 ϩ cyc7-⌬ yme1-⌬ ϩ ), having the yme1-⌬ disruption, was nearly the same as the normal strain B-8514 (curve A), except that all of the cytochromes are slightly diminished, similar to slightly repressed cells. Similarly, strain B-9621 (curve D) (cyc1-⌬ cyc7-⌬ yme1-⌬ ϩ ) was nearly the same as strain B-8123 (curve C), demonstrating that yme1-⌬ does not suppress the cytochrome a⅐a 3 deficiency caused by the absence of cytochrome c. The deficiencies of cytochromes a⅐a 3 and b, due to the lack of translation of the mitochondrial genes COX1, COX2, COX3, and COB1, can be seen in strains B-8516 (curve E) (CYC1 ϩ cyc7-⌬ Ϫ ) and B-9622 (curve F) (cyc1-⌬ cyc7-⌬ Ϫ ).
Analysis of Cytochrome Oxidase Subunits-Western blot analysis of the isogenic strains with specific antibodies was used to estimate, relative to the normal value, the levels of the 1. Low temperature (؊196°C) spectrophotometric recordings of a series of isogenic yeast strains. The strains were grown on 1% sucrose at 30°C for 3 days, and the absorption spectra were recorded as described previously (32). 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.

FIG. 2. Western blot analysis using monoclonal antibodies for CoxI, CoxII, and CoxIII, as described under "Materials and Methods.''
. The peaks at approximately 577 nm, seen in curves B-F, is due to zinc protoporphyrin. mitochondrial-encoded subunits, I, II, and III, as illustrated in Fig. 2, and the nuclear-encoded subunits IV, V, VI, and VIa. The levels of the subunits were determined quantitatively by comparing the intensities of the bands derived from the various strains to diluted samples prepared from the normal strain B-8514, similar to the procedure used by Calavetta and Capaldi. 1 The results, presented in Table I, indicated that strain B-9614 (CYC1 ϩ cyc7-⌬ yme1-⌬ ϩ ) had slightly diminished levels of all of the examined subunits, although the spectral peaks of cytochrome a⅐a 3 at 602.5 nm were nearly identical (Fig. 1). Strain B-8123 (cyc1-⌬ cyc7-⌬ ϩ ), which contains a primary deficiency in cytochrome c, lacked the mitochondrialencoded CoxI, CoxII, and CoxIII and had considerably decreased levels of CoxIV, CoxV, CoxVI, and CoxVIa, thus explaining the lack of cytochrome oxidase as indicated by spectral analysis (Fig. 1). Most importantly, the levels of CoxII and CoxIII, but not CoxI, were at least partially restored in strain B-9621 (cyc1-⌬ cyc7-⌬ yme1-⌬ ϩ ), thus suggesting that yme1-⌬ suppresses the degradation of some of the subunits. As expected, the Ϫ strains B-8516 (CYC1 ϩ cyc7-⌬ Ϫ ) and B-9622 (cyc1-⌬ cyc7-⌬ Ϫ ) lacked the mitochondrial-encoded subunits I, II, and III and contained diminished levels of the nuclearencoded subunits IV, V, VI, and VIa.
In Vivo Degradation of Subunit CoxII-Pulse-chase labeling experiments were used to establish that the diminished levels of CoxII were due to degradation. The results with the B-8514 (CYC1 ϩ cyc7-⌬ ϩ ), B-8123 (cyc1-⌬ cyc7-⌬ ϩ ), and B-9621 (cyc1-⌬ cyc7-⌬ yme1-⌬ ϩ ) are presented in Fig. 3. There was little or no apparent turnover of CoxII in the normal strain B-8514 after 2 h. In contrast, CoxII was rapidly degraded in the cytochrome c-deficient strain B-8123, resulting in complete or almost complete degradation by 2 h. CoxII was also degraded in the related yme1-⌬ strain B-9621, but at a lower rate. These pulse-chase experiments are in complete agreement with the results of the steady-state levels ( Table I), indicating that the reduced levels are due to degradation.
Degradation of Unassembled Subunits and Protection by Protein-Protein Interactions-Studies by other workers indicated that deletion of any of the genes encoding components of cytochrome oxidase leads to diminution of other subunits, suggesting that the unassembled subunits are prone to degradation. Dowhan et al. (10) demonstrated that disrupting COX4, and thereby eliminating subunit IV, resulted in the loss of the cytochrome a⅐a 3 absorption peak and in the diminution of subunits I, II, III, and V, but not VI and VII. Deletion of subunits V, VI (40), and VII (41) or VIIa (42) also abolished the spectrum for cytochrome a⅐a 3 , although levels of other subunits were not reported. More recently, Calavetta and Capaldi 1 systematically investigated the steady state levels of subunits I, II, III, IV, V, VI, and VIa, in yeast strains specifically deleted for one or another of the mitochondrial genes COX1, COX2, COX3 or the nuclear gene COX10, which is involved in heme a synthesis. The study revealed that subunits I, II, III, and VIa were greatly diminished if any of these COX1, COX2, COX3, or COX10 genes were deleted. However, deletion of COX1 or COX2 but not COX3 decreased the levels of subunits IV and V, whereas only deletion of COX2 decreased the level of subunit VI. Deletion of COX10 also diminished the levels of subunit V, but not subunits IV and VI. Overall, these findings suggest that unassembled subunits are degraded, with the degree of degradation of each subunit probably reflecting their intrinsic stability and their association with other subunits of the complex. Thus, deletion of a single subunit could lead to a cascade of complex degradations, which is dependent on the protective interactions.
The results presented in this paper can be explained by the protection of one or more of the cytochrome oxidase subunits by cytochrome c. Cytochrome c exhibits high affinity binding to subunit II and low affinity binding to another subunit (43). The absence of cytochrome c may destabilize CoxII, and possibly CoxI and Cox III, thereby making these, as well as other subunits, susceptible to degradation.
The results shown in Fig. 2 and summarized in Table I  demonstrated that degradation of subunits II and III, but not subunit I, in a cytochrome c-deficient strain can be partially suppressed by the yme1-⌬ mutation, which prevents synthesis of the Yme1 presumptive protease. The simplest explanation of these results is that Yme1 is a protease that, by itself or in a complex with other proteins, is involved in turnover of unassembled subunits of cytochrome oxidase, confirming a previous report that turnover of unassembled subunits of cytochrome oxidase requires a metal/ion-dependent factor (11). In addition, Weber et al. 2 observed that CoxII was degraded in a cox4deficient strain, and that this degradation was suppressed by yme1-⌬. The lack of suppression by yme1-⌬ of the CoxI complete deficiency (Table I), and of the possible partial deficiencies of other subunits, suggests that still other proteases may be involved in the degradation process. In fact, the turnover of CoxV was not suppressed by yme1-⌬, pim1-⌬, afg3-⌬, or rca1-⌬ (results not presented).
In summary, these results indicate that cytochrome c protects cytochrome oxidase subunits from degradation, and our previous results (27) demonstrated that cytochrome oxidase or cytochrome c 1 protects certain labile forms of cytochrome c from degradation, a phenomenon that was previously believed to occur only with strongly interacting components of protein complexes.