ASC1/RAS2 Suppresses the Growth Defect on Glycerol Caused by the atp1–2 Mutation in the Yeast Saccharomyces cerevisiae *

To better define the regulatory role of the F 1 -ATPase a -subunit in the catalytic cycle of the ATP synthase complex, we isolated suppressors of mutations occur-ring in ATP1 , the gene for the a -subunit in Saccharomyces cerevisiae . First, two atp1 mutations ( atp1–1 and atp1–2 ) were characterized that prevent the growth of yeast on non-fermentable carbon sources. Both mutants contained full-length F 1 a -subunit proteins in mitochon- dria, but in lower amounts than that in the parental strain. Both mutants exhibited barely measurable F 1 - ATPase activity. The primary mutations in atp1–1 and atp1–2 were identified as Thr 383 3 Ile and Gly 291 3 Asp, respectively. From recent structural data, position 383 lies within the catalytic site. Position 291 is located near the region affecting subunit-subunit interaction with the F 1 b -subunit. An unlinked suppressor gene, ASC1 ( a subunit complementing) of the atp1–2 mutation (Gly 291 3 Asp) restored the growth defect phenotype on glyc- erol, but did not suppress either atp1–1 or the deletion mutant D atp1 . Sequence analysis revealed that ASC1 was allelic with RAS2 , a G-protein growth regulator. The introduction of ASC1/RAS2 found misreading in the ATP1 sequence previously reported (18). We re-sequenced the wild-type ATP1 by using a genomic clone and used corrected sequence for comparison. The nucleotide se- quences described in this article are available from the DDBJ/EBI/ GenBank™ nucleotide sequence data bases under accession numbers D37948 ( ATP1 ), D37949 ( atp1–2 ), D88458 ( atp1–1 ), and D37950 ( ASC1 / RAS2 ). Computer Analysis— Homology searches were calculated by BLAST (basic local alignment search tool) (23) against all data compiled in the NCBI data base NCBI). Complementation Tests— Growth on a non-fermentable carbon source, glycerol, was examined by incubation on YPGE medium at 30 °C for 3–4 days. Preparation of Mitochondria— Cells were grown in 50 ml of YPDM medium. After 24 h of incubation at 30 °C, cells ( ; 2–4 3 10 7 cells/ml) were harvested and mitochondria were prepared according to method reported previously (24). heated at 100 °C for 5 min, and then subjected to SDS-polyacryl-amide gel electrophoresis (10% polyacrylamide gel, Tefco Corp., Tokyo, Japan). After electrophoretic transfer of gel-resolved F 1 a -subunit, bound anti-F 1 a -subunit antibody on the polyvinylidene difluoride mem- branes were assayed mainly by using a Western-Light rabbit kit (Tropix, Inc., Bedford, MA).

Mitochondrial ATP synthase functions as a key enzyme for ATP production in eukaryotic cells (1). The enzyme is controlled in response to the energy demands of the cell (2). Although considerable attention has been given to the central role of mitochondrial ATP production in the initiation of programmed cell death (apoptosis), little is known about the regulation of ATP synthase during its biogenesis and energy transduction or its links to growth regulatory pathways (3).
The enzyme complex is composed of the F 1 -ATPase (catalytic sector) and the transmembrane F 0 proton channel (embedded in inner membrane) (4 -6). Both F 1 and F 0 are necessary for ATP synthase activity, whereas F 1 alone retains the ability to hydrolyze ATP (F 1 -ATPase) (7). The F 1 -ATPase consists of five different subunits: ␣, ␤, ␥, ␦, and ⑀. The minimum unit for F 1 -ATPase activity resides on ␣-␤-subunit dimer (8). The catalytic center is considered to be in the ␤-subunit (9,10), and the ␣-subunit has been reported to play a role in the formation of the catalytic site with it (11). The ␣-subunit also assists the assembly of other subunits of the F 1 -ATPase (12) by acting as a chaperone to assist assembly (13). In yeast, all but three F 0 subunits of the enzyme are encoded on nuclear DNA. In order to examine the control mechanism(s) of the F 1 ␣-subunit in the complex assembly and function, we characterized several mutants (14) and isolated extragenic suppressors of mutations in the ATP1 gene. One extragenic suppressor for the point mutant, atp1-2, was RAS2, a well known regulator of cell proliferation and signal transduction (15)(16)(17). This work reveals for the first time that the cellular growth regulatory activity of RAS2 is linked in some manner with the biogenesis or function of mitochondrial enzyme complexes.
Gene Library-The Sau3AI pool of yeast genomic DNA from DC5 was cloned into the BamHI site of a vector YEp13 for the construction of yeast genomic library (19).
Plasmids-A plasmid pYCL12-5 (YCp type) was constructed as follows; approximately 2.9 kb of EcoRI-SphI fragment having ATP1 with its 5Ј-and 3Ј-flanking regions was cloned into a derivative of YCp50 in * This work was supported in part by National Institutes of Health Grant RO1GM36536 and American Heart Association Grant 92006620 (to M. G. D.). 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.
The nucleotide sequence reported in this paper has been submitted to the DDBJ/GenBank™/EBI Data Bank with accession numbers D37949 (atp1-2) and D88458 (atp1-1).
DNA Sequencing-Nucleotide sequencing was performed by dideoxy chain termination method (21), and autoradiograms were obtained according to the method previously reported (22). In the process of sequencing, we found misreading in the ATP1 sequence previously reported (18). We re-sequenced the wild-type ATP1 by using a genomic clone and used corrected sequence for comparison. The nucleotide sequences described in this article are available from the DDBJ/EBI/ GenBank™ nucleotide sequence data bases under accession numbers D37948 (ATP1), D37949 (atp1-2), D88458 (atp1-1), and D37950 (ASC1/RAS2).
Computer Analysis-Homology searches were calculated by BLAST (basic local alignment search tool) (23) against all data compiled in the NCBI data base (supplied October 10, 1997 by NCBI).
Complementation Tests-Growth on a non-fermentable carbon source, glycerol, was examined by incubation on YPGE medium at 30°C for 3-4 days.
Preparation of Mitochondria-Cells were grown in 50 ml of YPDM medium. After 24 h of incubation at 30°C, cells (ϳ2-4 ϫ 10 7 cells/ml) were harvested and mitochondria were prepared according to the method reported previously (24).
Determination of F 1 -ATPase Activity-F 1 -ATPase activity was measured by minor modification of the method of Pullman et al. (25) by following the decrease of the absorbency at 340 nm. The assay mixture consisted of 50 mol of Tris acetate (pH 7.4), 1 mol of MgCl 2 , 0.02 mol of NADH, 2 mol of phosphoenolpyruvate, 2 mol of ATP, 13 g of lactate dehydrogenase, and 16 g of pyruvate kinase in 1 ml.

RESULTS
Biochemical Properties of atp1-2 or atp1-1 Mutants-Yeast strains carrying atp1-2 (XJY11) and atp1-1 (XJY12) mutations fail to grow on the non-fermentable carbon source ( Fig. 1) and lack ATPase activity (Table I) (14,18). Both strains, however, were able to synthesize their full length ␣-subunits, although in reduced amounts (Fig. 2). The amount of mutant ␣-subunit was approximately 47% of the parental wild type strain DC5 (Table I) in atp1-2 and 52% in atp1-1 (14,18). The F 1 -ATPase activity in the atp1-2 mutant was barely measurable above that of ATP1 deletion, SKY4010. Thus, the atp1-2 mutation essentially led to complete loss of ATPase activity (Table I). The enzyme activity of ⌬atp1 strain, SKY4010, could be restored to 54% of wild-type enzyme activity by introduction of ATP1 on pYCL12-5. The recovery of enzyme activities in strains atp1-2 and atp1-1 transformed with the same plasmid was 43% and 37%, respectively (Table I). The lower ATPase specific activity of the point mutants transformed with the centromeric plasmid pYCL12-5 suggests that the mutant ␣-subunit protein probably competes with the plasmid-encoded wild type subunits for assembly into the functional enzyme complex.
Characterization of Mutations in atp1-1 and atp1-2-The nucleotide sequence of atp1-2 (Table II) revealed a guanine to adenine change at the second position in codon 291, is strictly conserved in ATP1 genes of other organisms (Table III), and results in an amino acid substitution from Gly 291 to Asp. Sequence analysis of atp1-1 also revealed a cytosine to thymine at the second position in codon 383 (Table II) generated an amino acid substitution from Thr 383 to Ile. The residue at Thr 383 is also a strictly conserved amino acid residue among all of the ATP synthase ␣-subunit genes reported so far. Therefore, these data indicate that both Gly 291 and Thr 383 of the F 1 ␣-subunit are each essential for function of the ATP synthase.
Isolation of an atp1-2 Suppressor Gene-Suppressor genes to atp1-2 (Gly 291 3 Asp) that restored growth of the mutants on a non-fermentable carbon source were selected as described (see "Experimental Procedures"). From 10,000 Leu ϩ transformants, five were able to grow on glycerol and were further characterized. Two of the five plasmids exhibited restriction maps different to that of ATP1 but shared a common region. We denoted the gene contained in these fragments as ASC (for ␣-subunit complementing gene). The ASC1 suppressor of atp1-2 did not restore the growth defect on glycerol of atp1-1 (Fig. 1). Thus, these results support that mutations in atp1-2 and atp1-1 may affect different functions or cellular activities of the F 1 -ATPase ␣-subunit. ASC1 failed to restore the growth on glycerol of other ATPase subunit mutants including the ATP1 deletion mutant SKY4010 (data not shown). When the ASC1 gene was placed on a centromeric plasmic, the atp1-2 transformant grew on glycerol but at one third the rate of the wild type (180 min doubling, Fig. 3). Thus, low copy gene dosage of ASC1 suppressed the atp1-2 mutation. When the atp1-1 (Thr 383 3 Ile) was transformed with the same genomic library, only ATP1 genes and no other were isolated on different plasmids using this selection.
ASC1 Is Identical to RAS2-To identify ASC1, the common region of the plasmid, inserts were sequenced (Fig. 4). This indicated that ASC1 was identical to a previously sequenced gene RAS2. This gene is a homologue to human oncogene involved in signal transduction (30,31). Further, Southern hybridization of chromosomes separated by contour-clamped homogenous electric field gel electrophoresis confirmed that digoxigenin-labeled ASC1 hybridized with chromosome XIV on which RAS2 is localized (data not shown) (32). Primary clone analysis also showed that ASC1 hybridized with ATCC clone 70762 alone, which contains RAS2 (data not shown). Finally, the yeast RAS2 gene obtained from another source suppressed atp1-2 in the same manner as ASC1 (data not shown).
Biochemical Properties of the atp1-2 Mutant Transformed with ASC1/RAS2-The presence of the plasmid encoded ASC1/RAS2 in the atp1-2 mutant caused no change in mitochondrial ATPase activity as long as the cells were grown on glucose (Table I). On glucose in this medium, there is no selective pressure for the mutant to retain the plasmid encoded ASC1/RAS2. However, when the transformant was grown on     the non-fermentable substrate like glycerol, the F 1 -ATPase activities increased in atp1-2 almost 6-fold. The level of enzymatic activity in atp1-2 transformed with ASC1 reached 30% that of the parental strain grown under the same conditions (Table IV). These results suggest that ASC1/RAS2 functions as part of the regulatory circuit linked to the control of F 1 -ATPase subunit synthesis on a non-fermentable carbon source.
The amount of F 1 -ATPase ␣-subunit protein was measured under different growth conditions to determine if increases in activity reflected enzyme content. The amount of Atp1-2p in XJY11 grown on glycerol only increased 1.3-fold following the introduction of ASC1/RAS2 (Table IV) on a centromeric plasmid. The amount of ␣-subunit increased about 2-fold in both the XJY11 and XJY12 mutants transformed with ASC1/RAS2 on glucose (Table I). This increase in the F 1 -ATPase ␣-subunit protein occurred when additional copies of ASC1/RAS2 were present. To further confirm the positive regulatory effect of RAS2 on Atp1p, we determined the level of F 1 ␣-subunit in a RAS2 deletion mutant. Disruption of ASC1/RAS2 in DC5 decreased the level of Atp1p compared with that of the parental strain (data not shown).
The combination of protein and enzymatic analysis (Tables I  and IV) was consistent with a model that additional copies of RAS2 increased the amount of the F 1 -ATPase subunit protein in some fashion. In the case of the mutant containing a partially functional atp1-2 subunit, the increase in protein likely provides a threshold level of mutant subunit and activity sufficient to support growth on glycerol. If additional RAS2 gene product suppressed atp1-2 by causing increased levels of the partially active Atp1-2p, then we should restore growth on glycerol of the atp1-2 mutant by introducing and expressing additional copies of the mutant gene. To test this, additional copies of atp1-2 were placed into the XJY11 mutant on the multicopy plasmid (pYEatp1-2). Following transformation, the level of Atp1-2p and F 1 -ATPase activity increased and growth was restored on glycerol (Table IV). These data support the model that ASC1/RAS2 controls the amount of ␣-subunit and ATP synthase activity in mitochondria in response to a nonfermentable carbon source. DISCUSSION In this paper we show that the specific mutant of ATP1, atp1-2, but not atp1-1, can be partially suppressed by RAS2, a mediator of signal transduction. Additionally, this is the first report of mutation sites in ATP1 in S. cerevisiae and will allow further examination of the role of this subunit in cell growth and its control. These studies identify an essential residue in ␣-subunit function at Thr 383 in the active site and another at Gly 291 , which retains partial enzymatic function and can be suppressed by RAS2 in multiple copies (Fig. 5). There are three copies of ATP1 in yeast (39,40); however, the results of sequence analysis of an extensive number of cloned PCR products indicate that only one mutation site was present in each mutant described here. In the mitochondria from atp1-2 as well as atp1-1, the F 1 -ATPase activity was barely measurable, although Western blotting revealed that there were 47% or 52% level of ␣-subunits, respectively, compared with the level in the wild-type strain. Expression of ATP1 in the same cell with either Atp1-2p or Atp1-1p yielded a lower specific activity than that for ATP1 expression in the absence of the mutants alleles. This indicated competition for assembly between the defective ␣-subunits, Atp1-2p or Atp1-1p, and the wild type ␣-subunit expressed from ATP1.
The unusual observation that RAS2 could specifically suppress the atp1-2 mutation indicates a relationship between a Arrows indicate open reading frames. The fragment (NheI-SphI, 3.7 kb) from pMTY1 was digested using enzymes and generated DNA fragments shown were cloned into YEp13, resulting in pMTYBH20, pMTYBH10, YEpNB20, YEpBS10, and YEpHH5, respectively. pSIY101 contains the disrupted open reading frame 2 by the insertion of LEU2. Yeast strain XJY11 was transformed with modified plasmids to Leu ϩ , and their glycerol phenotypes were subsequently analyzed as described under "Experimental Procedures." growth regulatory pathways involving RAS and mitochondrial energy transduction. Analysis of the atp1-2 mutation site revealed that it is not located in the catalytic domain like atp1-1 but at the boundary where subunit-subunit interaction is noted (Fig. 5). All of the information presented in this study is consistent with the ␣-subunit of atp1-2 exhibiting residual activity that in the presence of multiple copies of RAS2 reaches a threshold sufficient for growth on a mitochondrial-dependent substrate. On the other hand, the presence of additional copies of RAS2 in the atp1-1 mutant, XJY12, does not partially restore F 1 -ATPase activity or growth on glycerol. This is because the atp1-1 mutation is in the catalytic domain (Fig. 5).
Previous studies have shown that threshold levels of the energy transducing ATPase complex containing as little as 15% of the oligomycin-sensitive ATPase activity of wild type strains is sufficient to support growth on glycerol (41). This is consistent with the model that increasing the level of Atp1-2p and/or enhancing its assembly will yield sufficient energy transducing complex for growth on glycerol. This model for suppression of atp1-2 was confirmed in the experiment in which the gene dosage of atp1-2 was increased using a multicopy plasmid containing atp1-2. The resulting transformants exhibited an increase in F 1 -ATPase activity and were able to grow on glycerol (Table IV). Thus, increased Atp1-2p due to gene dosage was a necessary condition to support growth on a non-fermentable carbon source. In the XJY12 transformant, the amount of Atp1-1p also increased (Table I) due to ASC1/RAS2. Thus, ASC1/RAS2 appears to be involved in the regulation of ␣-subunit content in mitochondria. On glucose, the lack of ASC1/ RAS2 stimulation of ATPase protein and activity likely reflects the loss of ASC1/RAS2 plasmid due to non-selective growth conditions.
RAS2 has been previously characterized in yeast in earlier studies (16). One phenotype of different alleles of ras2 is the failure to grow on a non-fermentable carbon source, although a firm characterization of the mechanism of this remains open. RAS2 activates adenylyl cyclase, followed by activation of cAMPdependent protein kinases (42)(43)(44). The presence of a cAMPdependent protein kinase on the inner membrane of mitochondria has been reported (45,46). More recent work indicates a role for protein kinase activity in the regulation of mitochondrial transcription (47,48). Dupont et al. also reported that ccs1/ira2, an attenuator gene of RAS1 and RAS2, confers the resistance to inhibitors of the F 0 part of ATP synthase and increased in the ATP synthesis rate in mitochondria (49,50). This could suggest a role for RAS in the assembly of mitochondrial and nuclear encoded subunits of the complex.
The present study directly links a growth mediator to mitochondrial energy coupling. We demonstrate here that the ASC1/RAS2 stimulates the F 1 -ATPase in the atp1-2 mutant. This activation allowed growth on glycerol. Studies of the relationship between mitochondrial function and growth regulatory pathways have yet to be convincingly established in detail. However, considerable activity has recently focused on the pathways that link the function of mitochondria and mitochondrial components to the growth regulatory activities with which they are now firmly established. Most recently, Akt, a protein kinase B, which has been shown to act as an anti apoptotic regulator at many points in the pathway, is itself regulated by phosphorylation (3). So far, RAS2 functions have been investigated with glucose as a carbon source. It is well established that the RAS2-cAMP pathway is activated by glucose and that RAS2 is necessary to grow on a non-fermentable carbon source. Studies are currently in progress to better understand if these regulatory pathways share common control points and why gene dosage of RAS2 can influence the amount and function of ATPase subunits in mitochondria.