Manipulations in the Peripheral Stalk of the Saccharomyces cerevisiae F1F0-ATP Synthase*

The Saccharomyces cerevisiae F1F0-ATP synthase peripheral stalk is composed of the OSCP, h, d, and b subunits. The b subunit has two membrane-spanning domains and a large hydrophilic domain that extends along one side of the enzyme to the top of F1. In contrast, the Escherichia coli peripheral stalk has two identical b subunits, and subunits with substantially altered lengths can be incorporated into a functional F1F0-ATP synthase. The differences in subunit structure between the eukaryotic and prokaryotic peripheral stalks raised a question about whether the two stalks have similar physical and functional properties. In the present work, the length of the S. cerevisiae b subunit has been manipulated to determine whether the F1F0-ATP synthase exhibited the same tolerances as in the bacterial enzyme. Plasmid shuffling was used for ectopic expression of altered b subunits in a strain carrying a chromosomal disruption of the ATP4 gene. Wild type growth phenotypes were observed for insertions of up to 11 and a deletion of four amino acids on a nonfermentable carbon source. In mitochondria-enriched fractions, abundant ATP hydrolysis activity was seen for the insertion mutants. ATPase activity was largely oligomycin-insensitive in these mitochondrial fractions. In addition, very poor complementation was seen in a mutant with an insertion of 14 amino acids. Lengthier deletions yielded a defective enzyme. The results suggest that although the eukaryotic peripheral stalk is near its minimum length, the b subunit can be extended a considerable distance.

F 1 F 0 -ATP synthase is the principal enzyme for ATP generation in most organisms. The overall structure and mechanism of the enzyme is highly conserved. The catalytic sites are located in the F 1 sector, and ATP synthesis is driven by proton translocation through the F 0 sector. Proton translocation results in movement of the rotor stalk that extends through the center of F 1 , and it is this rotation that results in the conformational changes in F 1 that account for the binding change mechanism (1)(2)(3). The function of the peripheral stalk is to hold F 1 against the movement of the rotor stalk (4).
Although the prokaryotic and eukaryotic peripheral stalks apparently serve the same core function, the subunit structures are very different. In bacteria, the peripheral stalk of the F 1 F 0 -ATP synthase is primarily composed of two identical b subunits that are thought to be in an extended ␣ helix. The b subunits each transit the membrane once as part of F 0 , extend alongside one noncatalytic ␣/␤ subunit interface to the top of F 1 where the carboxyl terminus of at least one b subunit interacts with the ␦ subunit (5,6). The hydrophilic portion of two subunits, b subunits are thought to be in a parallel coiled-coil conformation from the surface of the membrane to the top of F 1 (7). Support for this can be found in the right-handed coiled-coil structure of the peripheral stalks of the related enzyme, the A 1 A 0 -ATP synthase of Thermus thermophilus (8). In contrast, the eukaryotic peripheral stalk is composed of four different subunits named OSCP, h (or F 6 in Bos taurus), d, and b in Saccharomyces cerevisiae. The S. cerevisiae ATP4 gene encodes a b subunit that differs substantially in both structure and primary sequence from the bacterial b subunit. The b subunit traverses the membrane twice and extends in a largely ␣ helical conformation up one side of F 1 (9,10). In the B. taurus peripheral structure (11), the membrane-extrinsic portion of the b subunit participates in protein-protein interactions with the d and F 6 subunits. Finally, the carboxyl-terminal region of the b subunit associates with the OSCP located at the top of the F 1 sector (12). The major differences in subunit composition between the eukaryotic and prokaryotic peripheral stalks raised a question about whether the two stalks have fundamentally different physical and functional characteristics.
Previous work done in our laboratory showed that the E. coli enzyme displayed a high degree of tolerance for substantial insertions and deletions within the b subunit in the tether domain where the protein emerges from the surface of the membrane. At the extremes, F 1 F 0 -ATP synthase activity was detectable in Escherichia coli enzymes with b subunits containing deletions of 11 and insertions of up to 14 amino acids (13,14). An active F 1 F 0 -ATP synthase was also observed containing two b subunits with a size difference of up to 14 amino acids (15). These results were interpreted as evidence that for a considerable degree of flexibility within the E. coli peripheral stalk, not only were substantial changes in length tolerated, but chimeric proteins were constructed using sequences from b subunits of distantly related organisms within a functional F 1 F 0 -ATP synthase (16,17).
In the present work, the length of the S. cerevisiae peripheral stalk was manipulated to determine whether the F 1 F 0 -ATP synthase had the same tolerance to insertions and deletions as the prokaryotic enzyme. A synthetic ATP4 gene (ATP4 syn ) was designed to facilitate site-directed mutagenesis and expressed in S. cerevisiae with an ATP4 gene disruption. Characterization of the mutants suggested that the S. cerevisiae peripheral stalk possesses a limited tolerance for changes in its length.

EXPERIMENTAL PROCEDURES
Materials and Media-Molecular biology enzymes and oligonucleotides were obtained from New England Biolabs and Invitrogen, respectively. Reagents, chemicals, and media were acquired from Sigma-Aldrich, Fisher Scientific, and MP Biomedicals.
YPD medium was used as a rich medium for growth of S. cerevisiae. The medium consisted of 1% Bacto-yeast extract, 2% Bacto-peptone, and 2% glucose. The carbon source of this rich medium can be substituted with either glycerol, in the case of YPG, 2 or galactose, in the case of YPGal. Complete synthetic drop-out medium, denoted CSM nutrient, was used to select for auxotrophic markers with S. cerevisiae. The defined medium omitted the appropriate nutrient for the auxotrophic marker. The medium contained 0.67% yeast nitrogen base, 2% glucose, and 0.077% nutrient drop-out mix (MP Biomedicals). Sporulation medium contained a limited amount of glucose and nutrients in order to induce sporulation of a diploid S. cerevisiae strain. Sporulation medium had 1% potassium acetate, 0.1% Bacto-yeast extract, 0.05% glucose. Solid S. cerevisiae medium was made by addition of 1.6% agar to any of the previously described media.
Construction of ATP4 syn Expression Vectors-The ATP4 syn shuttle vectors were constructed by a multiple-step protocol. The ATP4 syn gene was synthesized by Genscript (Piscataway, NJ) and was originally received as an insert in the plasmid pUC57 (Fig. 1). Flanking restriction sites EcoRI and NotI were used to facilitate moving the A4TP4 syn gene into shuttle vector pRS313. PCR primer pairs, AW114/AW115 and AW116/ AW117, were used to amplify the 150 bp upstream and downstream regions of the ATP4 gene from S. cerevisiae genomic FIGURE 1. The ATP4 syn gene. A, restriction map of the ATP4 syn gene. The distance of 100 bp is marked by a bar. B, complete sequence of ATP4 syn gene. Nucleotides changed from the wild type ATP4 to insert enzyme restriction sites through silent mutations and to alter codon preference are highlighted. Restriction sites are underlined, and the name for the enzyme for each restriction site is indicated below each sequence. DNA (Table 1). Restriction sites included in both primer pairs facilitated movement of the DNA into positions flanking the ATP4 syn gene, resulting in expression shuttle vectors pAW33 (Ap r , ATP4 syn , and HIS3) and pAW36 (Ap r , ATP4 syn , and URA3). The plasmids were confirmed by sequencing.
Plasmid Complementation of a ⌬ATP4 Strain-Primers AW83 and AW84 were used to amplify the KanMx6 cassette and flanking DNA from a haploid ATP4 disruption strain obtained from ATCC (catalogue no. 4012750) ( Table 1). The linear PCR product was transformed directly into BY4743, a diploid strain homozygous for the wild type ATP4 gene. S. cerevisiae transformations were performed essentially as described by Geitz et al. (18). Incorporation of the KanMx6 cassette into one ATP4 allele of the cells was detected by screening for growth on solid YPD media supplemented with 250 g/ml (active) G418 (YPDϩG418) and YPG medium. Colony PCR was performed was used as an initial screening tool to determine genotypes of candidate strains (19). Isolated genomic DNA was used for sequencing to confirm genotypes (20).
Plasmid pAW36 (ATP4 syn , Ap r , and URA3) was transformed into strain AW2 (MATa/MAT␣ his Ϫ ura Ϫ atp4::KanMx6/ϩ) ( Table 2). Sporulation was induced by a shift to solid sporulation medium. The resulting tetrads were dissected using a micromanipulator and haploid colonies were screened for growth on three types of media: (i) solid CSM without uracil (CSM-URA) medium, (ii) solid YPDϩG418 medium, and (iii) solid YPG to test for functional ATP synthase activity. Colonies were genotyped for a KanMx6 cassette disruption of the chromosomal ATP4 gene and the presence of the plasmid-borne ATP4 syn gene. The strain was tested for mating type using Cold Springs Harbor Method, found to be Mata and was named AW3 (Mata his Ϫ ura Ϫ atp4::KanMx6, pAW36 (ATP4 syn , Ap r , and URA3)).
Construction of Mutant Plasmids-Mutagenic oligonucleotides were designed to either insert or delete 4, 7, 11, or 14 codons (Fig. 2). Each mutant ATP4 syn gene was constructed in essentially the same manner by ligation-mediated site-directed mutagenesis (21). Mutant ATP4 syn expression plasmids were transformed into the S. cerevisiae strain AW3 (his Ϫ ura Ϫ atp4::KanMx6, pAW36 (Ap r , URA3, and ATP4 syn )) to yield cells that contained both the pAW36 plasmid (Ap r , URA3, and ATP4 syn ) and a mutant plasmid (Ap r , HIS3, and mutant ATP4 syn ). The pAW36 plasmid was selected against by replica plating the S. cerevisiae cells on solid YPDϩ 0.1% 5-fluoroorotic acid medium. After selection on solid YPDϩ5-fluoroorotic acid media, surviving cells are patched on to solid CSM-His (CSM without histidine) medium. Genomic DNA from selected patches grown on solid CSM-His medium were isolated and sequenced for the presence of the mutant ATP4 syn gene.
Preparation of Mitochondria-For mitochondrial preparation, the strains were grown in YPGal as a carbon source to allow for growth of the Ϫ strains. Biological F 1 F 0 -ATP synthase activity was determined by plating serial dilutions on plates containing rich media with 2% glycerol and 0.1% glucose (YPGD) and incubating at 30°C for 3 days.
Preparation of mitochondria from each mutant strain was done by scaling down the method of Meisinger et al. (22). A 5-ml overnight culture grown at 30°C was used to innoculate 50 ml of YPGal and allowed to grow to stationary phase. 500 ml of YPGal was inoculated with between 5-20 ml of the stationary culture and grown for 20 h.
To make spheroplasts, cells were harvested, washed once with sterile double-distilled H 2 O, and weighed. The cells were suspended at 2 ml buffer/g wet cells in prewarmed DTT buffer (100 mM Tris-H 2 SO 4 , pH 9.4, 10 mM DTT). The suspension was incubated with gentle agitation (80 rpm) for 20 min at 30°C. The cells were recovered by centrifuged and washed with zymolyase buffer (ZB) (1.2 M sorbitol, 20 mM potassium phosphate, pH 7.4) at 7 ml/g wet cells. Cells were suspended in ZB containing zymolyase 20-T (Fisher Scientific) at 7 ml/g wet cells and incubated at 30°C with slow shaking for 30 -45 min. After zymolyase treatment, the cells were harvested and washed in ZB at 7 ml/g wet cells.
To lyse the cells, the pellet was carefully suspended at 6.5 ml/g wet cells precooled homogenization buffer (0.6 M sorbitol, 10 mM Tris-HCl, pH 7.4, 1 mM EDTA, and 1 mM PMSF). All steps were performed at 4°C. The spheroplasts were lysed with 15 strokes of a glass Teflon homogenizer. Afterward the homogenate was diluted 2-fold with homogenization buffer and cen- The restriction sites for SalI, NcoI, NotI, and SacI are underlined, boldface, underlined and italicized, and boldface and italicized, respectively. trifuged at 1500 ϫ g for 5 min to pellet cell debris and nuclei. The supernatant was transferred to a clean centrifuge tube and centrifuged at 4000 ϫ g for 5 min. To recover the mitochondrial fraction, the supernatant was transferred to a second clean centrifuge tube and centrifuged for 15 min at 12,000 ϫ g. The pellet was gently suspended in 250 l of 4°C SEM buffer (250 mM sucrose, 1 mM EDTA, 10 mM MOPS-KOH, pH 7.4). The mitochondria were stored at 4°C until ATPase assays or a Western blot could be performed. Remaining mitochondria were flash frozen and stored at Ϫ80°C. Protein concentration was determined using the bicinchoninic acid method (23). Assays of F 1 F 0 -ATP Synthase-Assays of ATP hydrolysis were performed essentially as described by Tzagoloff and Ackerman (24). Mitochondrial protein samples (30 g protein/ well) were loaded onto a 12% Tris-glycine SDS Bio-Rad TGX Gel. Following electrophoresis, the mitochondrial proteins were transferred to a PVDF membrane (Bio-Rad) by electroblotting (110 V for 20 min at 4°C). Detection was performed using the Rodeo ECL Western kit (Affymetrix), including the washing of the membrane with Rodeo Blocker. The rabbit antiserum against the S. cerevisiae b subunit was a generous gift from Dr. Rosemary Stuart (Marquette University). The rabbit antiserum against the b subunit was used at a concentration of 1:1000. The secondary antibody, anti-rabbit IgG (Amersham Biosciences) was used at a 1:30,000 concentration. The antibody was detected using chemiluminescence (Affymetrix). The signals were visualized using high performance chemiluminescence film (GE Healthcare) and a Kodak X-Omat. Signal strength was assessed using Un-Scan-It gel digitizing software (Silk Scientific, Inc.).

RESULTS
Plasmid Complementation with ATP4 syn -To construct a strain lacking a functional ATP4 gene, it was disrupted by a linear piece of DNA containing the KanMx6 cassette by homologous recombination. However, the deletion of the ATP4 gene from the genome was known to cause S. cerevisiae cells to go either Ϫ or 0 (25). To overcome this problem, a scheme was devised to maintain a copy of the native ATP4 gene or the ATP4 syn gene in the cells during all steps involved in chromosomal ATP4 gene deletion strain construction.
Wild type diploid strain, BY4743 (ATCC), was transformed with a linear PCR product containing the KanMx6 cassette flanked by an ATP4 sequence. The ATP4 sequence provided a means for homologous recombination with a single ATP4 allele. Transformants were selected for growth on YPDϩG418 medium, genomic DNA was prepared and sequenced. The resulting heterozygous strain was named AW2 (MATa/MAT␣ his Ϫ ura Ϫ atp4::KanMx6/ϩ). Plasmid pAW36 (ATP4 syn , Ap r , and URA3) was transformed into strain AW2. Sporulation was induced resulting in tetrads. Haploid colonies were genotyped for the KanMx6 cassette disruption of the chromosomal ATP4 gene and the presence of the plasmid-borne ATP4 syn gene. The strain was subsequently tested for mating type, found to be Mata, and named AW3 (Mata his Ϫ ura Ϫ atp4::KanMx6, pAW36 (ATP4 syn , Ap r , URA3)). Plasmid complementation of the chromosomal disruption was tested by patching AW3 (ATP4 syn ) on YPG medium. The strain grew robustly at a level comparable with the wild type parental strain BY4743. This showed that the synthetic ATP4 syn gene allowed phenotypically normal F 1 F 0 -ATP synthase activity in vivo. A second strain AW4 (Mata his Ϫ ura Ϫ atp4::KanMx6, pAW36 (ATP4 syn , Ap r , and HIS3)) was constructed by plasmid shuffling to serve as a positive control for subsequent experiments.
Mutant Strain Construction-Insertion and deletion mutagenesis of the E. coli peripheral stalk suggested a considerable degree of plasticity, but it was unknown whether this property was shared with the eukaryotic enzyme (13,14). To test the eukaryotic stalk, a comparable set of mutations were designed to lengthen and shorten the b subunit of S. cerevisiae. The ATP4 syn genes were constructed with 4 -14 codons either deleted or inserted. In the case of the insertions, the coding sequence was duplicated with alternative codons for the amino acids wherever possible. The number of codons inserted or deleted was based on the periodicity of an ␣ helix. The site for the manipulations was selected by inspection of the high-resolution structure of the B. taurus peripheral stalk (11) and the cryoelectron microscopy image of the S. cerevisiae F 1 F 0 com-plex (26). In both structures, the peripheral stalk narrowed to the width of a single ␣ helix where the b subunit is likely to emerge from the membrane. This narrowing of the peripheral stalk roughly coincides with the region in the E. coli enzyme where much of the insertion and deletion work was done. The S. cerevisiae and B. taurus b subunits share 44% sequence similarity. The target within the peripheral stalk was chosen by aligning the deduced amino acid sequence of the B. taurus b subunit with that of the S. cerevisiae b subunit (Fig. 3). This alignment revealed that the narrow region in the B. taurus peripheral stalk structure was homologous to amino acids b K78 -A94 in S. cerevisiae. However, two short segments of highly of conserved amino acids appear at b D79 -A83 and b V100 -D108 , so mutations were designed to minimize modification of these sites. Therefore, the segment of the ATP4 syn gene encoding amino acids b A81-A94 was selected as the target for mutagenesis.
Growth Properties of Mutant Strains-Recombinant ATP4 syn genes were tested for the ability to complement the disruption of the endogenous ATP4 gene in vivo by growth on YPG and YPGD. Use of glycerol as a carbon source requires a functional, intact F 1 F 0 -ATP synthase, and inclusion of a limited amount of dextrose in YPGD media allowed for slow growth of petite mutants.
The insertion and deletion strains showed dramatically different growth phenotypes. Deletions of seven or more amino acids showed no growth on glycerol and little growth on YPGD medium, suggesting the petite phenotype (Fig. 4A). In contrast, the ATP4 syn ⌬4 strain grew well on YPG and YPGD and was comparable with that of the control AW4 (ATP4 syn ) strain. Insertions of up to eleven amino acids also showed robust growth on both types of media (Fig. 4B). Significantly, less growth was observed for the ATP4 syn ϩ14 strain, indicating the enzyme was just barely capable of sustaining biologically significant function.
Expression of Recombinant b Subunits-It was likely that levels of expression and/or import of the recombinant b subunits might be affected due to the introduction of mutations. Western blots were performed on mitochondria isolated from the mutant strains using anti-b subunit antiserum to determine steady-state levels of the recombinant b subunits (Fig. 5). The positive control from strain AW4 (ATP4 syn ) showed a strong signal indicating that the synthetic gene product was expressed and imported into the mitochondria (Fig. 5, A and B, lane 2). Levels of the ATP4 syn ⌬4 subunit were comparable with ATP4 syn . The ATP4 syn ⌬7 and ATP4 syn ⌬11 subunits were detected at much lower levels  Amino acid sequence alignment of B. taurus and S. cerevisiae b subunits. The protein sequences were obtained from the Pubmed protein data base and were aligned using protein BLAST. The method used was the compositional adjustment method. The two sequences had 24% sequence identity and 44% sequence similarity. MARCH 25, 2011 • VOLUME 286 • NUMBER 12 JOURNAL OF BIOLOGICAL CHEMISTRY 10159 (Fig. 5B). A band from ATP4 syn ⌬14 subunit was only observable by prolonged exposure of the film.

Yeast F 1 F 0 -ATPase Peripheral Stalk
In contrast, the ATP4 syn ϩ4 subunit was present at levels comparable with the positive control. As the b subunit was further lengthened, expression levels decreased. Although both of the strains expressing ATP4 syn ϩ7 and ATP4 syn ϩ11 subunits had a normal biological phenotype, the level of expression was only 60% of that of the positive control. Extension to 14 amino acids showed a further decrease of expression levels to about 50% of the positive control. This 10% decrease in expression resulted in a disproportionate change in growth phenotype. Although successfully imported into the mitochondria, it seems that the ATP4 syn ϩ14 subunit may not be efficiently incorporated into an F 1 F 0 complex. Therefore, expression level of the b subunit did not necessarily directly correlate to growth phenotypes of the mutant strains. F 1 F 0 -ATPase Activity-In mitochondria-enriched samples, the greatest single source of ATP hydrolysis activity is attributed the F 1 F 0 -ATP synthase. Traditionally, the amount of P i release that is specifically assigned to the enzyme is the amount of activity lost by addition of the F 1 F 0 -ATP synthase inhibitor, oligomycin. Oligomycin inhibits the enzyme by interacting with the F 0 proton channel at a site thought to be located at the interface between the a subunit and the ring of c subunits (27).
In the absence of oligomycin, strain AW4 (ATP4 syn ) mitochondrial-fractions had significantly greater ATP hydrolysis activity than the negative control strain (Table 3). Mitochondria prepared from the ATP4 syn disruption strains displayed ATPase activity comparable with the negative control. Although the ATP4 syn ⌬4 strain grew well on YPG media, mitochondrial fractions prepared from the strain exhibited little or no F 1 F 0 -ATPase activity. The evidence suggested that the ATP4 syn ⌬4 enzyme was clearly intact and active in vivo, but the complex was apparently unstable during preparation of the mitochondrial sample. In the presence of oligomycin, the negative control had negligible oligomycin sensitivity (2%). Most of the deletion strains showed no appreciable oligomycin sensitivity (Table 3).
In stark contrast to the deletion strains, the insertion strain mitochondrial samples possessed abundant ATPase activity ( Table 4). Strains that contained b subunits with insertions up to eleven amino acids possessed ATPase activity that was indistinguishable from the positive control. Even the ATP4 syn ϩ14 sample displayed activity substantially greater than the negative control value. In the presence of oligomycin, the strain AW4 (ATP4 syn ) had a 57% decrease in ATP hydrolysis activity, a value comparable with that reported in the literature (28,29). Surprisingly, the ATP4 syn ϩ4, ATP4 syn ϩ11, and ATP4 syn ϩ14 strains showed little oligomycin sensitivity, suggesting that oligomycin inhibition was lost due to the insertion mutations.
Interestingly, both the ATP4 syn ⌬7 and ATP4 syn ϩ7 mitochondria reproducibly showed considerable oligomycin sensitivity. As noted before the ATP4 syn ⌬7 has the petite phenotype, whereas the ATP4 syn ϩ7 possesses a wild type growth pheno-    type. We suspect that these particular changes have a minimal effect on the orientation of the b subunit relative to the a subunit within the F 0 sector. This may account for the oligomycin sensitivity unique to those two subunits.

DISCUSSION
Here, we report the construction, expression, and characterization of a series of mutations in the b subunit of S. cerevisiae F 1 F 0 -ATP synthase. Plasmid shuffling was used to achieve the ectopic expression of altered b subunits in a S. cerevisiae strain carrying a chromosomal disruption of the ATP4 gene. Insertions of up to 11 amino acids yielded a wild type phenotype for growth on YPG medium, and mitochondrial-enriched fractions had readily detectable levels of the recombinant subunits with abundant ATP hydrolysis activity. This activity showed surprisingly little oligomycin sensitivity when considered in the context of growth phenotype. Additionally, expression of the ATP4 syn ϩ14 b subunit provided limited complementation to a strain carrying a disrupted ATP4 gene. In contrast, most of the deletion mutations resulted in a loss of F 1 F 0 -ATP synthase function. Only the smallest deletion subunit, ATP syn ⌬4, yielded a positive growth phenotype on YPG medium. The F 1 F 0 complex with this shortened subunit had no apparent activity in vitro, suggesting enzyme instability during preparative procedures.
Previous work done in our laboratory showed that the E. coli F 1 F 0 -ATP synthase could accept large changes in length of the b subunits (13,14). Those experiments were interpreted as evidence that considerable plasticity was an inherent property of the bacterial peripheral stalk. The experiments presented in this paper suggest that the eukaryotic peripheral stalk has similar plasticity, at least with respect to a short deletion and lengthy insertions. Changes in length ranging from the deletion of four to the insertion of 11 amino acids in the b subunit yielded an active enzyme. Assuming the predicted ␣ helical conformation remains intact, this represents a change in length of ϳ18 Å (Fig. 6A). Although this may not seem like a large change in length, the measured distance from the top of the membrane to the bottom of F 1 is ϳ40 Å. Therefore, these mutations result in a change of 45% in the length of the peripheral stalk segment between the F 1 and F 0 sectors in an intact F 1 F 0 complex. The additional three amino acids in the ATP4 syn ϩ14 subunit resulted in substantially less enzyme.
The evidence from experiments involving deletions is less clear. In E. coli, deletion of seven amino acids was essentially wild type, and longer deletions resulted in the assembly of fewer F 1 F 0 -ATP synthases (13). In S. cerevisiae, deletions of seven or more amino acids resulted in an apparent total loss of enzyme function. Although there was abundant ATP synthesis activity apparent in vivo, even the ATP4 syn ⌬4 b subunit F 1 F 0 complex proved unstable during preparation of the mitochondria. For the longer deletions, it is unclear whether the enzyme failed to assemble with the deletion subunits or whether the F 1 F 0 complex was unstable. In any case, once the b subunit is too short its interactions within the F 1 F 0 complex are irrevocably compromised. The lack of F 1 F 0 -ATP synthase is associated with deformation or loss of the mitochondria leading to an overall deficiency in energy metabolism (30). The evidence would seem to suggest that the wild type S. cerevisiae b subunit is very near its minimum functional length. Considering the sequence similarity between S. cerevisiae and B. taurus b subunits, especially within the segment crystallized by Dickson et al. (11), it appears likely that eukaryotic F 1 F 0 -ATP synthases evolved to favor the shortest practicable peripheral stalks.
One of the most surprising findings was the lack of oligomycin sensitivity associated with the majority of the mutant ATP4 syn F 1 F 0 -ATP synthases. With the exception of the ATP4 syn ϩ7 enzyme, all of the mutant strains that contained functional F 1 F 0 -ATP synthases were essentially insensitive to oligomycin. The result suggests that a subtle conformational change occurred within the F 1 F 0 complex due to incorporation of the ATP4 syn ϩ4, ATP4 syn ϩ11, and ATP4 syn ϩ14 subunits. The shift appeared to be propagated within the F 0 sector to the oligomycin binding site at the proton channel. Importantly, the shift did not cause any apparent loss in channel function, and only inhibitor binding seemed to be affected. There are two important implications of this observation. First, this suggests that oligomycin sensitivity is not a valid measure of F 1 F 0 -AT-Pase activity in enzymes with b subunit mutations and by extension alterations in other peripheral stalk subunits. Further support for the conclusion is found in a study involving the successful functional replacement of the h subunit by the B. taurus F 6 subunit. This replacement resulted in an oligomycin insensitive enzyme (31). Second, propagation of a conformation shift in the b subunit implies that changes in peripheral stalk length may not be fully accounted for by structural changes within the stalk itself.
Recent physical biochemistry experiments have indicated that within the context of the entire enzyme, the E. coli peripheral stalk is much less compliant than the central stalk (32,33). Insertion of 11 amino acids within the b subunit yielded in a more flexible peripheral stalk, but not to a level that it reaches the pliability of the central stalk (35). A recent high resolution structure of the B. taurus F 1 -peripheral stalk complex showed the possibility of two hinges B. taurus peripheral stalk (Fig. 6B) (12). One is located in the OSCP subunit on top of the F 1 complex. The second appears to be located in the region of b 146 in the B. taurus enzyme. Additionally, the cryoelectron microscopy structure shows the peripheral stalk narrowing as it emerges from the membrane to a diameter consistent with a single ␣ helix provided by the b subunit (34). This is the region where the insertions and deletions within the b subunit were constructed. The narrowing of the peripheral stalk, along with the current experiments that show considerable plasticity is consistent with a third hinge in the area where the b subunit emerges from the membrane. This suggests that localized regions of higher plasticity are necessary for the peripheral stalk to be incorporated into the complex. Once the enzyme is assembled, the peripheral stalk then assumes a stiff conformation relative to the central stalk.
Two hinges within the b subunit, one located near the surface in the membrane and the other alongside F 1 at b 146 , might act as pivot points for the b subunit during F 1 F 0 assembly. This suggests a hinged stiff model that accommodates the present understanding of the peripheral stalk. Modest changes in the angle of these hinges would alter the pitch of the peripheral stalk, allowing incorporation of a shortened or a lengthened b subunit into the F 1 F 0 complex (Fig. 6B). Alterations in pitch may contribute to the loss oligomycin sensitivity observed with peripheral stalk recombinant subunits.