The Saccharomyces cerevisiae nuclear gene RPM2 encodes a component of the mitochondrial tRNA-processing enzyme RNase P. Cells grown on fermentable carbon sources do not require mitochondrial tRNA processing activity, but still require RPM2, indicating an additional function for the Rpm2 protein. RPM2-null cells arrest after 25 generations on fermentable media. Spontaneous mutations that suppress arrest occur with a frequency of ~9 × 10−6. The resultant mutants do not grow on nonfermentable carbon sources. We identified two loci responsible for this suppression, which encode proteins that influence proteasome function or assembly. PRE4 is an essential gene encoding the β-7 subunit of the 20S proteasome core. A Val-to-Phe substitution within a highly conserved region of Pre4p that disrupts proteasome function suppresses the growth arrest of RPM2-null cells on fermentable media. The other locus, UMP1, encodes a chaperone involved in 20S proteasome assembly. A nonsense mutation in UMP1 also disrupts proteasome function and suppresses Δrpm2 growth arrest. In an RPM2 wild-type background, pre4-2 and ump1-2 strains fail to grow at restrictive temperatures on nonfermentable carbon sources. These data link proteasome activity with Rpm2p and mitochondrial function.
MITOCHONDRIA are vital organelles in eukaryotic cells. In addition to their role in respiration and oxidative phosphorylation, mitochondria are the site for such diverse cellular functions as heme biosynthesis, metabolite transport, amino acid biosynthesis, and lipid catabolism (Attardi and Schatz 1988; Pon and Schatz 1991). Mitochondrial biogenesis and function require input from both nuclear and mitochondrial genomes. In the yeast Saccharomyces cerevisiae, mitochondrial DNA codes for proteins required for oxidative phosphorylation, electron transport, mitochondrial protein synthesis, and mRNA processing. Ribosomal RNA, tRNA genes, and an RNase P RNA gene necessary for mitochondrial protein synthesis are also coded by the organelle genome. The maintenance of a wild-type mitochondrial genome is dependent on mitochondrial protein synthesis (Myerset al. 1985).
In facultative aerobes like S. cerevisiae, mutations in mitochondrial DNA are tolerated if cells are grown on fermentable carbon sources. Consequently, nuclear genes encoding respiratory components and factors required for mitochondrial gene expression affect respiratory growth but not growth on fermentable carbon sources. In contrast, nuclear mutations that interfere with organelle formation and/or maintenance disrupt growth on all carbon sources (De Pintoet al. 1999).
Yeast mitochondrial RNase P is a ribonucleoprotein complex required for removing 5′ leader sequences from mitochondrial tRNA precursors (Moraleset al. 1992; Dang and Martin 1993). The RNA component of mitochondrial RNase P (Rpm1r) is encoded by the mitochondrial genome (Shuet al. 1991). The protein component of mitochondrial RNase P (Rpm2p) is encoded in the nucleus. Biochemical and genetic analyses established Rpm2p as a subunit of mitochondrial RNase P (Moraleset al. 1992; Dang and Martin 1993). Rpm2p is the major protein found in highly purified preparations of mitochondrial RNase P (Moraleset al. 1992). Moreover, antibodies raised against Rpm2p immunoprecipitated both RNase P activity and Rpm1r from mitochondrial extracts (Dang and Martin 1993). Finally, strains with an insertional disruption midway through the RPM2 reading frame accumulated mitochondrial tRNA precursors with 5′ extensions, Rpm1r precursors, and were unable to grow on respiratory carbon sources (Moraleset al. 1992; Stribinskiset al. 1996). These data demonstrate that RPM1 and RPM2 encode subunits of RNase P and that the activity of this ribonucleoprotein complex is required for respiratory but not fermentative growth.
Unexpectedly, cells with a null allele of RPM2 could not grow on either respiratory or fermentable carbon sources, indicating that RPM2 had a function in addition to its role as a subunit of mitochondrial RNase P (Kassenbrocket al. 1995). This function can be provided by the N-terminal half of Rpm2p, explaining why strains with the insertional disruption grew on fermentable carbon sources. The nature of this second function is unclear, but it could be related to the isolation of RPM2 as one of two high-copy suppressors of a temperature-sensitive (ts) allele of TOM40/ISP42 (Kassenbrocket al. 1995). TOM40 encodes an essential membrane component of the mitochondrial import machinery (Vestweberet al. 1989; Moczkoet al. 1992; Rapaportet al. 1997; Hillet al. 1998). The N-terminal portion of RPM2 that is sufficient for growth on fermentable carbon sources also suppresses the TOM40 mitochondrial import defect (Kassenbrocket al. 1995). Rpm2p and Tom40p do not appear to interact when assayed with the two-hybrid system, suggesting that RPM2 may suppress the mutant allele without physically interacting with Tom40p (Groom 1995). Thus, the requirement for RPM2 for growth remains unclear.
RPM2-null mutants display a significant rate of phenotypic reversion on glucose-containing media. This rate is indicative of spontaneous second-site loss-of-function suppressor mutations in genes functionally linked to the essential growth function of RPM2. We labeled these second-site suppressors, suppressors of arrest (soa) mutants.
Here, we describe five independent soa suppressor strains that fall into three different complementation groups, all of which suppress lack of growth of Δrpm2 strains on glucose media. We demonstrate that two of these mutant alleles affect 26S proteasome function. PRE4 was identified as the wild-type allele of soa1 and encodes the β-7 subunit of the 20S proteasome core (Grollet al. 1997). UMP1 is the wild-type allele of soa2 and encodes a chaperone necessary during the assembly of the 20S proteasome core complex (Ramoset al. 1998). Thus, defects in proteasome activity compensate for the lack of RPM2.
MATERIALS AND METHODS
Strains, plasmids, and media: Standard yeast manipulations were used (Sherman 1991; Sherman and Hicks 1991; Kaiseret al. 1994). Yeasts were transformed with plasmid DNA using a lithium acetate method (Chenet al. 1992). The yeast strains used in this study are listed in Table 1. We used the centromeric pRS300 series of plasmids for subcloning (Sikorski and Hieter 1989).
Rich glucose media, YPD, included 1% Bacto-yeast extract, 2% Bacto-peptone, and 2% glucose. Rich glycerol-ethanol media, YPGE, contained 1% Bacto-yeast extract, 2% Bacto-peptone, 3% (v/v) glycerol, and 2% (v/v) ethanol. Synthetic complete (SC), synthetic complete lacking specific amino acids (SC trp−), and synthetic dextrose minimal media (SD) contain 0.67% Bacto-yeast nitrogen base without amino acids, 2% glucose, and appropriate supplements (Kaiseret al. 1994). Solid media for plates included 2% Bacto-agar. Culture media reagents were Fisher Scientific (Pittsburgh) or Difco (Detroit) brand. G418, canavanine, and CdCl2 aqueous solutions were sterilized by passage through a 0.2-μm filter (Micron Separations Inc.) and added to autoclaved plate media at the concentrations reported in results.
Construction of null allele of RPM2: YMW1 and YMW2 are haploid derivatives of W303 cells and were a generous gift from Dr. Michael Walberg (Zieleret al. 1995). Deletion of RPM2 was performed in a YMW1/YMW2 diploid, YML9. The complete RPM2 open reading frame (ORF) was replaced with the heterologous dominant G418 resistance gene kanMX using a SFH-PCR-based gene disruption technique (Wachet al. 1994; Wach 1996). RPM2/kanMX primers were 5′ CAACAAACAGAAAAGAAAAAACAAGCATACACAAACGAAAATGCGTACGCTGCAGGTCGAC 3′ and 5′ GTTATAACAGTGAAATAAATAAAATATCAAATTGTAAAGGTCAATCGATGAATTCGAGCTCG 3′. Underlined sequences in oligonucleotides mark those parts of the sequence that are complementary to pFA6-MCS in plasmid pFA6-kanMX6 (Wach 1996). Southern blot analysis and PCR confirmed the proper insertion of kanMX at an RPM2 locus, specifically starting one nucleotide (nt) 3′ of the first ATG of the RPM2 ORF through to the penultimate nt 5′ to the STOP codon.
Suppressors of the RPM2-null mutation: Three different diploid RPM2/Δrpm2::kanMX strains, YML34, YML37, and YML41, were sporulated and haploid spores were separated by tetrad analysis on YPD plates. The sporulation pattern observed in all cases was a 2:2 ratio of big/small colonies on YPD at 30° after 3 days. The small colonies do not increase in size after an additional week at 30°. A total of 14 small Δrpm2 colonies were individually suspended in 200, 300, or 400 μl of YPD. A small aliquot of cells was counted using a hemocytometer. The remaining cells were plated to YPD at 30°. Suppression of Δrpm2-arrested growth results in phenotypic revertant colonies, most likely occurring through spontaneous second-site suppressors we called suppression of arrest alleles (soa alleles). A maximum of four different colony sizes was seen within revertant populations. Individual colonies were picked and replated to score for homogeneous growth. Only those Δrpm2 soa colonies that grew homogeneously were studied further. To determine if the soa mutations were recessive or dominant to Δrpm2, Δrpm2 soa cells were first mated to ρ+ Δrpm2 SOA strains containing a wild-type copy of RPM2 on a YEp352 covering plasmid. A 5-FOA shuffle was performed to lose the YEp352 RPM2 plasmid (Sikorski and Boeke 1991).
Identification of wild-type SOA alleles: Seventeen independent Δrpm2 soa phenotypic revertants were backcrossed with isogenic wild-type strains, YMW1 or YMW2, to isolate the soa mutations in an otherwise wild-type background and determine if suppression originated from single-site mutations. Diploids were sporulated and haploid progeny were separated by tetrad analysis. The growth of haploid progeny was evaluated on rich and minimal glucose and glycerol media at 16° and 37°. Tight conditional growth phenotypes were observed in five independent backcrosses, permitting us to pursue these soa strains further. Tetrad analysis patterns of parental, nonparental, and tetratype indicated that the mutations originated as single, nuclear loci. The ability of RPM2 SOA/RPM2 soa diploids to grow under restrictive conditions demonstrated that soa1, soa2, soa3, soa4, and soa5 were recessive.
RPM2 soa1 yeast (37B1.14OCD3) were transformed with a pRS200 centromeric genomic library, YPH1, (Sikorski and Hieter 1989; American Type Culture Collection (ATCC) no. 77165) and screened for suppression of the cold-sensitive (cs) phenotype at 16° on YPGE. A plasmid containing an 8.8-kb insert suppressed this conditional growth phenotype. Subcloning identified a 2.2-kb fragment containing a single ORF, PRE4, that supported growth of RPM2 soa1 cells at 16° on YPGE. Linkage of the mutant phenotype with the PRE4 locus was confirmed by integrating a PRE4 URA3 linearized plasmid at the PRE4 locus in RPM2 soa1 cells (Rothstein 1991). These cells were mated to isogenic YMW2, and haploid progeny were scored for uracil auxotrophy and for growth under restrictive conditions.
RPM2 soa2 yeast (34D1.51OCC1) were similarly transformed with a YCp50 centromeric genomic library, CEN BANK A (Roseet al. 1987; ATCC no. 37415), and screened at 37° on YPGE. Plasmid-based suppression was seen with an 11-kb insert. Subcloning identified UMP1 as the wild-type allele suppressing the soa2 ts phenotype. Linkage of the mutant phenotype with the UMP1 locus was confirmed by integrating a UMP1 URA3 linearized plasmid at the UMP1 locus in RPM2 soa2 cells (Rothstein 1991). These cells were mated to isogenic YMW1, and haploid progeny were scored for uracil auxotrophy and for growth under restrictive conditions.
Nucleic acid techniques and DNA sequencing: Standard DNA manipulations and methodologies were used (Sambrooket al. 1989). Restriction enzymes and T4 DNA ligase were from New England Biolabs (Beverly, MA) or MBI Fermentas. PCR DNA polymerase enzymes used include TAQ (Fisher), Klentaq-LA polymerase mix (Sigma, St. Louis), and Vent (New England Biolabs). The primers for PCR recovery of sequence at the PRE4 locus were upstream primer 5′ CATTCCGTGTAGTGCCAATA 3′ and downstream primer 5′ GGTTACCCGTCAATCCTTTT 3′. The equivalent PCR primers for the UMP1 locus were upstream primer 5′ CGGCACTAATCTAATCTTTCAA 3′ and downstream primer 5′ TGGCAATTTTATCCTACCTGAG 3′.
DNA sequencing was performed by the University of Wisconsin Biotechnology Center, DNA Synthesis and Sequencing Facility (Dr. Charles Nicolet). PCR products were sequenced directly and after cloning into a PCR-TA cloning vector (Invitrogen). Both strands of each PCR product were completely sequenced to assure accuracy. Plasmid DNA was isolated using either the Bio-Rad (Richmond, CA) or Qiagen miniprep kits. BLAST searches compared and aligned nucleotide sequences (Altschulet al. 1990), while the SIM alignment portion of the ExPASy home page was used for amino acid comparisons. Amino acid numbering of the mature Pre4p is based on information from the Protein Data Bank (Brookhaven, NY) under accession no. 1RYP (Grollet al. 1997).
Protein analysis: Protein was isolated from log-phase cultures. Cells were grown at 30° to an OD640 of 0.5 and then transferred to 16° for 3.5 hr. Cells were collected by centrifugation and stored at −70°. Frozen pellets were resuspended in 200 μl NET-NP (150 mm NaCl, 5 mm EDTA, 50 mm Tris-HCl, pH 7.5, 0.5% NP-40) plus protease inhibitors. Protease inhibitors were diluted 1:10 (Complete mini cocktail tablets; Boehringer Mannheim, Indianapolis). An equal volume of sterile glass beads was added, and cells were disrupted for 3 min in a mini-bead beater (Biospec Products) followed by 15 min at −70° and 3 more min in the bead beater. Samples were then boiled and cleared of particulate matter by centrifugation at 12,000 × g in a microfuge. The concentration of proteins in the supernatant was determined (Bio-Rad protein assays). A total of 1 μg of protein was loaded onto a 10% SDS/polyacrylamide gel, run, and blotted to Immobilon-P (Millipore, Bedford, MA) membranes. Probes included a rabbit polyclonal antiubiquitin antibody (Haas and Bright 1985) generously provided by Dr. A. Haas, and a mouse monoclonal antiactin antibody (Boehringer Mannheim). Enhanced chemiluminescence (Amersham Life Sciences, Arlington Heights, IL) was used for signal detection.
Microscopy: Nuclear and mitochondrial DNA were stained with the vital dye DAPI (4′,6-diamidino-2-phenylindole) in at least four independent experiments for each strain (Kaiseret al. 1994). Approximately 107 cells were pelleted in a microfuge with a 5-sec pulse and resuspended in 70% ethanol on ice. Cells were fixed for 10–30 min on ice and washed twice with H2O. Cells were then resuspended in 1 μg/ml DAPI and mounted using either aqueous Light antifade or Prolong antifade (Molecular Probes, Eugene, OR). Observations were made on a Nikon Optiphot light microscope with a ×60 Plan-Apo 1.4 NA objective.
Identification and isolation of spontaneous second-site suppressors of Δrpm2: Sporulation of RPM2/Δrpm2::kanMX diploids and subsequent growth of haploid progeny on rich glucose yielded a 2:2 ratio of small/large colonies. More than 90% of the cells from the small colonies with the Δrpm2 allele lacked buds, indicating that they had ceased growth.
As a first step toward understanding this phenotype, we determined the number of cells in these small colonies. An average of ~4.4 × 107 cells per small colony was obtained by direct counting of cells in 14 independent colonies. This number is consistent with ~25 rounds of cell division from the initial spore. The mean spontaneous phenotypic reversion rate from 14 independent Δrpm2::kanMX colonies, plated on rich glucose at 30°, was between 1.5 × 10−7 and 1.4 × 10−6 revertants/original spore. Phenotypic revertants were designated suppressor of arrest (soa). At 30° on rich glucose media revertant colonies grew at different rates from each other and from isogenic wild-type strains. Seventeen Δrpm2 soa double mutants that demonstrated growth defects on YPD, at either 16° or 37°, were selected for further study.
Of 17 soa mutants, 5 demonstrated tight temperature-sensitive (Ts−) or cold-sensitive (Cs−) phenotypes on rich YPGE media linked to the soa phenotype when separate from the Δrpm2 mutation. Subsequent crosses demonstrated that each of the soa mutations were recessive (materials and methods). The conditional phenotypes were used in screening yeast genomic libraries for wild-type alleles of the soa loci.
Pairwise crosses of each of these five soa haploids were made, and diploids were plated under restrictive growth conditions. These crosses indicated that there were three soa complementation groups (Table 2). Analysis of Δrpm2 soa/Δrpm2 SOA diploid strains showed that mutations in complementation groups soa1 and soa3 were recessive and unable to suppress the Δrpm2 phenotype in the presence of their wild-type alleles. The soa2 allele, on the other hand, was semidominant and partially able to suppress Δrpm2 in the presence of wild-type SOA2.
The SOA1 locus is PRE4, which encodes a β subunit of the 20S proteasome: RPM2 soa1 cells were transformed with a low-copy centromeric genomic library (YPH1) and screened for complementation of their Cs− respiratory growth phenotype (materials and methods). A single plasmid containing an 8.8-kb insert complemented restrictive growth at 16° on YPGE. Subcloning the six ORFs in this insert revealed that PRE4 was responsible for the complementation. PRE4 encodes the β-7 subunit of the 20S proteasome core complex and is essential for cell viability (Hiltet al. 1993; Grollet al. 1997). Figure 1A shows PRE4 suppression of Cs− and Ts− respiratory growth phenotypes of RPM2 soa1 cells. To determine if soa1 is an allele of PRE4, a plasmid containing wild-type PRE4 and URA3 was integrated at the PRE4 locus in soa1 cells. These cells were crossed to wild type, sporulated, and tetrads were dissected. The haploid progeny exhibited a 4:0 ratio of wild-type/mutant segregation pattern of YPGE growth at 16°, demonstrating that soa1 is a mutant allele of PRE4. The soa1 allele is the second allele of PRE4 identified and was termed pre4-2 (Hiltet al. 1993). The previously identified pre4-1 allele produces a truncated protein with decreased peptidylglutamyl peptidase hydrolyzing (PGPH) activity that does not affect cell growth (Hiltet al. 1993).
The SOA2 locus is UMP1, which encodes a 20S proteasome maturation factor: Two overlapping plasmids complementing the soa2 defect were isolated (materials and methods). Subcloning the six possible ORFs revealed that UMP1 was the complementing gene. Figure 1B shows UMP1 suppression of the Ts− respiratory growth phenotype of RPM2 soa2 cells. UMP1 also suppresses the Ts− respiratory phenotype of the other member of the soa2 complementation group. UMP1, like PRE4, is necessary for the function of the 20S core proteasomal complex. UMP1 encodes a chaperone involved in 20S proteasome assembly (Ramoset al. 1998). To determine if soa2 mutations were alleles of UMP1, a plasmid containing wild-type UMP1 was integrated at the UMP1 locus in soa2 cells. These cells were crossed to wild type, sporulated, and tetrads were dissected. The haploid progeny revealed a 4:0 ratio of wild-type/mutant segregation pattern of growth on YPGE growth at 37°, demonstrating that the soa2 locus is UMP1. We named the soa2 allele ump1-2.
Sensitivity of soa mutants to canavanine and CdCl2: To determine if proteasome function was compromised in pre4-2, ump1-2, and soa3 strains, cultures were incubated under a variety of stressful conditions that increase cellular dependence on proteasomal degradation. Exposure of cells to increasing concentrations of the arginine analogue canavanine causes a general increase of mutant proteins that are normally degraded by the proteasome. Relative to wild-type cells, both pre4-2 and ump1-2 cells showed a hypersensitivity to canavanine, a phenotypic characteristic of cells with defective proteasome activity (Figure 2; Chen and Hochstrasser 1996; Ramoset al. 1998). RPM2 pre4-2 cells have decreased colony size relative to wild-type strains at 5 μg/ml canavanine and show no growth between 30 and 35 μg/ml canavanine (Figure 2A). Similar results are seen with the RPM2 ump1-2 mutants that do not grow on media containing 40 μg/ml canavanine. In contrast, the RPM2 soa3 cells grow to near-wild-type levels when spotted on the SD-canavanine plates. The combined effect of a null mutation of RPM2 with any of the soa mutations led to poor growth on SD media alone and no growth in the presence of low concentrations of canavanine (5–10 μg/ml; data not shown). This implies that cells lacking RPM2 have an increased sensitivity to stressful conditions.
A second test of proteasome function is the sensitivity of cells to oxidative damage caused by CdCl2. RPM2 pre4-2 and RPM2 ump1-2 cells grow poorly in the presence of 30 μm CdCl2 (Figure 2B). The double Δrpm2 soa mutants failed to grow under the same conditions (data not shown). Similar to results obtained with canavanine, RPM2 soa3 cells were not sensitive to CdCl2. Both the canavanine and CdCl2 sensitivities were overcome if RPM2 pre4-2 and RPM2 ump1-2 cells were transformed with PRE4 and UMP1, respectively, on low-copy vectors (Figure 2, A and B). The soa3 complementation group does not show evidence of defective 26S proteasome activity. Our investigations into the relationship of soa3 as a suppressor of Δrpm2 yeast are ongoing.
Degradation of ubiquitinated proteins is inhibited in pre4-2 yeast: As a further test of impaired proteasome activity in pre4-2 cells, we examined the level of ubiquitinated proteins at permissive and restrictive growth conditions (Haas 1988; chen and Hochstrasser 1995; Rinaldiet al. 1998). Wild-type and mutant cells were grown at 30° in YPGE media. Aliquots from log-phase cultures were either transferred to 16° for 3 hr or maintained at 30°. Total protein was isolated, fractionated by SDS-PAGE, transferred to polyvinyldifluoride membrane, and probed with antiubiquitin antibody. The blot showed an increased signal in RPM2 pre4-2 extracts vs. wild-type cells at 16° (Figure 3; Haas and Bright 1985). The diffuse signal in the molecular weight range 35–250 kD, which reflects a general increase in ubiquitinated proteins, is typical for cells with impaired proteasome activity.
The pre4-2 allele contains a missense mutation: The pre4-2 mutation is a guanine-to-thymidine base change at nucleotide 157 in the PRE4 ORF, which results in a valine-to-phenylalanine substitution at amino acid (aa) 12 in the mature protein. Sequence comparison of pre4-2p with the three most closely related 20S proteasome β-7 subunits and with the archtype β subunit of Thermoplasma acidiphilium showed that this amino acid substitution resides within a highly conserved region. This change introduces a bulky Phe at the start of β-strand 2 within the β-sandwich tertiary structure described in Groll et al. (1997).
Sequence of ump1-2 and ump1-3 reveals both contain nonsense mutations: Cells lacking UMP1 are viable but grow slowly (Ramoset al. 1998). Ump1p is postulated to interact with proPre2p and block protease maturation until proteasome assembly is complete. Both ump1-2 and ump1-3 contained nonsense mutations resulting in premature stop codons at codons 10 or 36, respectively.
Morphology of Δrpm2 soa and RPM2 soa mutants: Cells lacking RPM2 fail to produce mature mitochondrial tRNA, leading to an absence of mitochondrial protein synthesis and petite S. cerevisiae strains. The state of the mitochondrial genome was analyzed using a combination of DAPI staining and genetic crosses. DAPI staining revealed mitochondrial DNA in Δrpm2 pre4-2 cells, while Δrpm2 ump1-2 and Δrpm2 soa3 showed no extranuclear fluorescent staining (Figure 4). Mitochondrial DNA was distributed throughout the elongated, interconnected Δrpm2 pre4-2 cells even though nuclear DNA did not appear to be segregating properly. To determine if Δrpm2 pre4-2 maintained a wild-type mitochondrial genome, the cells were mated to YMW2 ρo, an isogenic strain containing a wild-type nuclear genome but lacking mitochondrial DNA. The resultant diploids were unable to grow on YPGE, demonstrating that Δrpm2 pre4-2 cells are ρ−.
Proteasome degradation of cyclins involved in cell cycle regulation has been well documented (Chunet al. 1996; Kinget al. 1996; Hershko 1997). The Δrpm2 pre4-2 and Δrpm2 ump1-2 mutants displayed a range of morphologies indicative of cells defective in cell cycle progression (Figure 4). Approximately 5–10% of Δrpm2 pre4-2 cells grown at 30° on YPD were multiply budded or showed interconnected, elongated morphologies (Figure 4B). After an overnight shift to the restrictive temperature of 16°, the entire culture displayed aberrant morphologies with many dead, serially connected ghosts observed (data not shown). Δrpm2 ump1-2 cultures of interconnected cells continue to divide at 37° but do so quite slowly with decreased viability (data not shown). At 30° an interconnected, elongated morphology was most evident when the Δrpm2 ump1-2 cultures reached stationary phase (Figure 4C). Δrpm2 soa3 cultures appeared as wild-type cells at both permissive and restrictive temperatures (Figure 4D).
Mutant morphologies of cells at permissive and restrictive conditions were not as prominent within the soa1 and soa2 strains with wild-type RPM2. All the single soa mutants were respiratory competent with ρ+ mitochondrial genomes. We used DAPI staining to observe the nuclear and mitochondrial DNA patterns in these soa mutants. During log-phase growth at 30°, pre4-2 and ump1-2 cultures are noticeably different from wild-type cells in that they contain multiply budded and elongated cells (Figure 5, A, C, and E). However, this mutant phenotype is not displayed to the same degree as it was in Δrpm2 pre4-2 and Δrpm2 ump1-2 cultures. As in Figure 4, not all cell bodies contain nuclear DNA, but mitochondrial DNA is found throughout the interconnected cells. Under restrictive conditions (3–16 hr at 16° or 37°, respectively) both RPM2 pre4-2 and RPM2 ump1-2 cultures contained elongated, multiply budded cells similar to their Δrpm2 counterparts. Consistent with the results from the double mutant, RPM2 soa3 retains a wild-type morphology under permissive and restrictive conditions, supporting the conclusion that soa3 is not a proteasome-functional mutant (Figure 5, G and H).
The results presented here demonstrate that growth on fermentable carbon sources can be restored in null mutants of RPM2 by mutations in two different genes that disrupt 20S proteasome function. One encodes the proteasome β-7 subunit; the other encodes a chaperone necessary for assembly of an active 20S proteasome complex (Ramoset al. 1998).
Proteasome function is essential in all eukaryotic cells for the regulated degradation of targeted proteins. The ATP- and ubiquitin-dependent proteasome pathway plays a key role in the degradation of misfolded/mutant proteins, as well as in controlling cell growth and division through the degradation of regulatory proteins (Chunet al. 1996; Hilt and Wolf 1996; Kinget al. 1996; Hershko 1997). The 26S proteasome is the proteolytic component of the ubiquitin-mediated protein degradation pathway. It is a dumbbell-shaped symmetrical structure composed of a barrel-shaped 20S proteolytic core capped by 19S regulatory particles. The yeast 20S proteasome has been crystalized recently and contains 14 distinct but structurally related subunits: 7 of type α and 7 of type β organized into four stacked heptomeric rings in the configuration αββα (Grollet al. 1997). The 20S core complex is assembled from proprotein precursors including inactive precursors of the proteasome proteases (Nandiet al. 1997). All β subunits are encoded by essential genes.
Subunits within the 20S core complex can have both structural and catalytic roles in proteasome function. Three different β subunits each provide distinct N-terminal threonine proteases to the proteasome catalytic core. The PUP1, PRE2, and PRE3 genes encode β-subunit proteases with trypsin-like, chymotrypsin-like, and PGPH specificities, respectively (Arendt and Hochstrasser 1997; Heinemeyeret al. 1997; Ciechanover and Schwartz 1998). Pre4p is not catalytically active as a protease but is necessary for PGPH activity (Hiltet al. 1993; Heinemeyeret al. 1997). Mutations in proteasome β subunits, such as Pre4p, are thought to influence the conformation of catalytic subunits or affect the binding and orientation of substrate peptides. Consistent with a critical role for noncatalytic subunits in proteasome structure is the observation that a null mutant of PRE4 is lethal. The requirement for PRE4 is not, however, solely related to its role in PGPH activity, since the truncated allele pre4-1 lacks PGPH activity but displays no detectable effects on cell growth or on the degradation of a reporter protein in vivo (Hiltet al. 1993).
Our work suggests that the Val-to-Phe substitution in β-7 has global effects on proteasome activity. Cells with a pre4-2 allele exhibit biochemical and morphological stress-related phenotypes characteristic of cells with reduced proteasome activity. Moreover, pre4-2 cells accumulated higher levels of ubiquitin-protein conjugates relative to wild type. We interpret these data to indicate that the pre4-2 mutation affects more than just the PGPH activity of the proteasome. Biochemical, immunological, and crystallographic studies have shown that Pre4p is in close proximity to Pre3p and Pup1p (Grollet al. 1997; Heinemeyeret al. 1997; Koppet al. 1997). Thus, the pre4-2 mutation could potentially disrupt multiple proteolytic activities within the proteasome as well as overall structural integrity. Similar “poisoning” of the Drosophila melanogaster 20S proteasome by mutant β-2 or β-6 subunits has been reported (Smyth and Belote 1999). However, unlike the D. melanogaster mutants, pre4-2 is recessive to PRE4 in the presence or absence of RPM2, indicating that the wild-type protein is preferentially incorporated into the 20S proteasome during assembly.
Proper proteasome assembly is critical to proteasome function. Unlike Pre4p, Ump1p is not an integral component of the 20S core complex. However, since Ump1p is involved in the assembly and maturation of the 20S core complex, mutations that affect its expression or function have global effects on proteasome function (Ramoset al. 1998). Thus, ump1-2, like pre4-2, appears to suppress the lack of growth on fermentative carbon sources in Δrpm2 cells by disrupting proteasome function. Yeast carrying null alleles of UMP1 accumulated 15S precursors to the 20S core complex and proPre2p (Ramoset al. 1998). They grow at much slower rates and are sensitive to stressful conditions. In agreement with these results, the ump1-2 mutant reported here had sensitivities to canavanine and CdCl2 characteristic of proteasome mutants, presumably because the cells had fewer fully assembled, proteolytically active 20S proteasome complexes. The semidominant nature of the ump1-2 allele in suppressing RPM2-null mutations suggests that a single wild-type UMP1 gene in a Δrpm2/Δrpm2 diploid does not support the assembly of wild-type levels of 20S proteasome complexes.
Rpm2p is a multifunctional protein that is required for mitochondrial RNase P activity and for an essential function, which, like mitochondria, is required during fermentative growth. The amino-terminal 714-aa portion of RPM2, which supports growth on glucose, is sufficient to serve as a high-copy suppressor of a mitochondrial protein import defect resulting from a mutant allele of TOM40 (Kassenbrocket al. 1995). The pre4-2 and ump1-2 suppression of Δrpm2 suggests a model relating proteasomal activity to an essential feature of mitochondrial biogenesis.
Two other proteasome mutants that suppress mitochondrially related phenotypes have been reported (Campbellet al. 1994; rinaldiet al. 1998). A mutant allele of RPT3 (ynt1-1) was identified as a suppressor of the enhanced rate of escape of mitochondrial DNA to the nucleus in strains with mutant alleles of YME1 (Thorsnesset al. 1993; Campbellet al. 1994). The RPT3 mutant also suppressed abnormal mitochondrial morphologies and growth defects associated with yme1 mutations. Yme1p is an ATP- and zinc-dependent protease component of the i-AAA proteolytic complex, which is localized to the inner mitochondrial membrane (Thorsnesset al. 1993; Leonhardet al. 1996; Weberet al. 1996). Unassembled subunits of ATP synthase and respiratory chain components are degraded, in part, through a pathway involving Yme1p (Pearce and Sherman 1995; Weberet al. 1996). RPT3, on the other hand, encodes one of six ATPase activities found in the base subcomplex of the proteasome 19S regulatory particle (Glickmanet al. 1998). Like cells with the pre4-2 allele, the mutant allele of RPT3 leads to cold-sensitive growth on nonfermentable carbon sources (Campbellet al. 1994). It seems unlikely that the RPT3 mutant suppresses YME1 mutations by replacing the mitochondrial protein as a functional component of the i-AAA complex. Instead, RPT3 may function to suppress YME1 mutants by disrupting the pathway used to degrade abnormal mitochondria (Campbell and Thorsness 1998). Recent studies have indicated that a vacuolar protease, Pep4p, plays an important role in the escape of mitochondrial DNA to the nucleus in YME1 mutants. Mutants in PEP4 suppress DNA transfer from mitochondria to nucleus in YME1 mutants without suppressing growth defects and abnormal mitochondrial morphologies. This suggests that the vacuole may function in the degradation of abnormal mitochondria in YME1 strains (Campbellet al. 1994; Campbell and Thorsness 1998). In contrast, RPT3 mutants suppress all phenotypes associated with the YME1 mutation, indicating that proteasome mutants help promote mitochondrial integrity.
A proteasome mutant has also been shown to suppress a mutation in the mitochondrial genome (Rinaldi et al. 1994, 1995). A search for MnCl2-generated suppressors of a mitochondrial tRNAASP point mutant led to the isolation of a yeast proteasomal gene RPN11/MPR1 (Zennaroet al. 1989; Rinaldi et al. 1994, 1998; Glickmanet al. 1998). RPN11/MPR1 is a non-ATPase subunit of the lid portion of the 19S proteasomal regulatory particle (Glickmanet al. 1998). Similar to pre4-2, mpr1-1 cultures grown to stationary phase accumulate ubiquitinated proteins and growth phenotypes were more prominent on glycerol medium. Cells with the mpr1-1 allele contain an increased quantity of total mitochondrial DNA per cell, suggesting that a decrease of proteasome activity can elevate the copy number of mitochondrial DNA (Rinaldiet al. 1998).
The proteasome mutants isolated here are unique with respect to the work described above because they specifically affect the 20S proteolytic core vs. the 19S regulatory cap subunits of the 26S proteasome. A suppression mechanism that stabilizes mitochondrial DNA through the loss of proteasome function, as observed in Rinaldi et al. (1998), cannot account for suppression in Δrpm2 yeast since the Δrpm2 ump1-2 and Δrpm2 soa3-1 cells are ρo. Alternatively, if the ynt1-1 proteasome mutant suppresses yme1 by supporting global mitochondrial integrity, thereby bypassing the vacuolar degradation pathway, a similar mechanism stabilizing mitochondria in the presence of decreased proteasomal degradation could account for phenotypic suppression in Δrpm2 pre4-2 and Δrpm2 ump1-2 strains.
We suggest that the mechanism by which proteasome mutants suppress the RPM2-null allele is indirect, affecting the steady-state level of key protein(s) that compensate for the loss of RPM2. Unlike mammalian cells, lower eukaryotes, including S. cerevisiae, do not degrade long-lived proteins through the ubiquitin-proteasome pathway (reviewed in Goldberget al. 1997). Thus, pre4-2 and ump1-2 strains should accumulate short-lived proteins, including regulatory proteins known to have profound effects on cell growth. Consistent with this interpretation is the observation that the null allele of RPM2 can also be suppressed by multiple copies of YBL066, which encodes a putative transcription factor, and PAB1, which encodes the poly (A)-binding protein (Groomet al. 1998; H. C. Heyman and N. C. Martin, personal communication). Both of these genes have the potential to increase the steady-state level of certain proteins, which, in turn, could suppress RPM2 mutants.
Proteasome degradation of outer mitochondrial membrane proteins could act as part of a general system of mitochondrial maintenance similar to that described for the ER membrane proteins, insuring that improperly folded proteins or misassembled complexes are degraded rapidly (Kopito 1997; Plemper and Wolf 1999). Essential functions of mitochondria that depend on outer mitochondrial membrane proteins include import, fusion, and distribution. Decreased proteasome activity would then impact mitochondrial function by increasing the longevity of mitochondrially associated proteins, potentially suppressing lethal phenotypes in strains containing mutations adverse for proper function or maintenance of mitochondria. This model accounts for our results while also encompassing the results obtained with both the YME1 and tRNAASP mutants (Zennaroet al. 1989; Campbellet al. 1994; Campbell and Thorsness 1998).
Nuclear gene products comprise the majority of mitochondrial proteins needed for respiratory and nonrespiratory growth. A mitochondrial import system is integral to the transport of these nuclear-encoded proteins into the different mitochondrial compartments. Increased stabilization of mitochondrial outer membrane proteins or increased rates of mitochondrial import could affect the growth characteristics of yeast under both respiratory and fermentative conditions. An RPM2 essential growth function directed at this level could be compensated for by increasing the half-life of specific proteasome targets. This hypothesis is supported by previous results that RPM2 was isolated as a high-copy suppressor of a defective allele of the essential mitochondrial import channel protein Tom40p (Kassenbrocket al. 1995).
Proteasomes could regulate the levels of proteins targeted to mitochondrial compartments. At least one case of control of mitochondrial protein levels by cystosolic proteasome activity has been established. Regulation to achieve proper levels of two isoforms of cytochrome c, encoded by two different nuclear genes and localized to the intermembrane space, occurs through a ubiquitin-proteasome degradation mechanism (Pearce and Sherman 1997). It is therefore likely that other proteins destined for mitochondria can be regulated by proteasome activity before mitochondrial import.
The direct turnover of mitochondria utilizing a pathway balancing proteasomal and vacuolar degradation activities would provide cells with a mechanism for monitoring and removing unneeded or improperly functioning mitochondria (Campbell and Thorsness 1998; Sutovskyet al. 1998). A hypothesis that mitochondrial turnover involves ubiquitin-tagged proteins is supported by studies of ubiquitin localization. Ubiquitin conjugates and ubiquitin-conjugating enzymes have been localized to mitochondrial membrane fractions and the mitochondrial outer membrane (Zhaung and McCauley 1989; Magnaniet al. 1991). Fisk and Yaffe (1999) demonstrated that the loss of an E3 ubiquitin ligase affects mitochondrial inheritance. A mechanism for protein ubiquitination on the cytosolic side of mitochondria would be useful in regulated degradation of mitochondrial proteins by the proteasome or for targeted turnover of mitochondria by the vacuole.
Cell fractionation experiments have previously shown that detectable Rpm2p is localized to the mitochondrial fraction (Kassenbrocket al. 1995). Rpm2p, as a component of mitochondrial RNase P activity, is localized to the matrix. However, the presence of low levels of Rpm2p in other mitochondrial fractions or in the cytosol cannot be ruled out. The limited growth of the small Δrpm2 colonies suggests that a proteasome target affecting mitochondrial function, or mitochondria themselves, is diluted out after 25 generations. The specific role of Rpm2p during cell growth remains unknown but we clearly demonstrated that decreased proteasome function can provide the resources necessary in the absence of RPM2. Our results suggest that Rpm2p has a role in regulating the synthesis or turnover of proteins important for growth. This role can be bypassed in the two novel proteasome mutants characterized here.
We thank Ms. Clarissa Moxely and Mr. Michael Mindrum for technical assistance with the complementation assays and library screens. We thank Dr. A. Haas for the generous donation of affinity purified polyclonal antiubiquitin antibody and Dr. M. Walberg for the generous donation of YMW1 and YMW2 yeast strains. This work was supported by National Institutes of Health grant 992332 to N.C.M.
Communicating editor: M. Johnston
- Received August 25, 1999.
- Accepted November 2, 1999.
- Copyright © 2000 by the Genetics Society of America