Bovine Coupling Factor 6, with Just 14.5% Shared Identity, Replaces Subunit h in the Yeast ATP Synthase*

The mammalian mitochondrial ATP synthase is composed of at least 16 polypeptides. With the exception of coupling factor F6, there are likely yeast homologs for each of these polypeptides. There are no obvious yeast homologs of F6, as predicted from primary sequence comparison of the putative peptides encoded by the open reading frames in the yeast genome. In this manuscript, we demonstrate that expression of bovine F6 complements a null mutant in ATP14 gene in yeast Saccharomyces cerevisiae. Subunit h of the yeast ATP synthase is encoded by ATP14 and is just 14.5% identical to bovine F6. Expression of bovine F6 in an atp14 null mutant strain recovers oxidative phosphorylation, and the ATP synthase is active, although functioning with a lower efficiency than the wild type enzyme. Like subunith, bovine F6 is shown to interact mainly with subunit 4 (subunit b), a component of the second stalk of the enzyme. These data indicated the subunith is the yeast homolog of mammalian coupling factor F6.

The F 0 F 1 -ATP synthase is the major enzyme responsible for the aerobic synthesis of ATP. It exhibits a tripartite structure consisting of a head piece (F 1 catalytic sector), base piece (F 0 , membrane sector), and two connecting stalks. F 1 is a watersoluble unit retaining the ability to hydrolyze ATP. F 0 is embedded in the membrane and is mainly composed of hydrophobic subunits forming a specific proton conducting pathway. The connecting stalks are composed of components from both F 1 and F 0 . When F 1 and F 0 are coupled, the enzyme functions as a reversible H ϩ -transporting ATPase or ATP synthase (1,2).
The establishment of the crystal structure of the major part of the bovine F 1 (3) allowed experiments that demonstrated that the enzyme is a molecular rotary motor, as shown by the ATP-dependent rotation of the ␥-subunit (4,5) and consistent with the binding site hypothesis of Boyer (6). In Escherichia coli, F 1 and F 0 are linked by two stalks, one of which is made of subunits ␥ and ⑀, and these also constitute a part of the rotor (7). Three other subunits, ␦ of F 1 and the two b-subunits of F 0 , are also involved in the binding of F 1 to F 0 and are thought to form the second stalk of the stator. The stator is thought to fix the ␣ 3 ␤ 3 oligomer to the a-subunit, thus allowing rotation of the c-subunit oligomer together with the ␥and ⑀-subunits. (8 -10) while holding the head piece in place.
The E. coli ATP synthase and the bovine enzyme are composed of 8 and 16 different subunits, respectively (11). In the case of Saccharomyces cerevisiae, the F 0 F 1 -ATP synthase is composed of at least 13 different subunits involved in the structure of the enzyme; the disruption of any of the corresponding structural genes leads to a lack of assembly of the holo-complex (12). Recently, the establishment of the structure of the yeast enzyme at 3.9 Å resolution revealed the structure of the F 1 and the subunit c oligomer of F 0 (13).
Among the supernumerary subunits of the yeast ATP synthase F 0 , subunit h has been described as an essential component because inactivation of the ATP14 gene led to a lack of oxidative phosphorylations (14). Recently cross-linking experiments (15) have positioned this hydrophilic and acidic component of 10,408 Da close to subunit 4 (subunit b), a component of the second stalk of the ATP synthase. In this paper we report the complementation of a yeast strain devoid of the yeast ATP synthase subunit h by a single copy vector bearing a DNA sequence encoding the bovine coupling factor 6. This is a rather remarkable result because subunit h and bovine F 6 share only 14.5% sequence identity.

EXPERIMENTAL PROCEDURES
Yeast Strains and Nucleic Acid Techniques-The S. cerevisiae strain D273-10B/A/H/U (MAT␣, met6, ura3, his3) (16) was the wild type strain. The strain with the null mutation in ATP14 (MAT␣, met6, ura3, his3, ATP14::URA3) has been described (14). The ⌬ATP14 strain containing the plasmid pbF 6 , was obtained by transformation of the null mutant in ATP14 gene by the nonintegrative single copy vector, pRS313, which contains the coding region of mature bF 6 , 1 the leader peptide of the ␤-subunit of the yeast ATP synthase, and the upstream and downstream transcriptional controlling elements of the ATP2 gene.
The expression plasmid for expression of bF 6 , pbF 6 , was made essentially as described (17). In this scheme, the coding region of mature bF 6 replaces the coding region of the gene encoding the ␤-subunit of the ATP synthase, ATP2. This allows the expression of bF 6 to be under the same controls as that of the ATP2 gene. The coding region of bF 6 is amplified by PCR, and this PCR fragment is used to directly replace the coding region of the ATP2 gene. The replacement occurs by site-specific homologous recombination effected in yeast. In addition to the bases required to direct the synthesis of the coding region of bF 6  primers also contain a 30-base sequence that targets the PCR product to the desired site of recombination on the ATP2 gene, in this case at the end of the region coding for the leader peptide and at the 3Ј-untranslated part of the gene.
Recombination is effected in vivo in yeast and occurs across a linearized plasmid (pRS314) (18) that contains the ATP2 gene that is cut in the center (gap repair) (for a diagram of this method, see Ref. 17). The recombination event occurs at the region encoding the leader peptide, at one end, and at the stop codon of the coding frame of bF 6 (a clone containing the cDNA for bovine F 6 was kindly provided by Dr. John E. Walker, Cambridge, UK). In this manner, the coding region of mature bF 6 precisely replaces the coding region of the ATP2 gene.
The DNA sequence of the chimeric gene was sequenced to ensure the correct recombination event and to ensure the absence of any mutations. DNA sequence analysis was performed at the Iowa State University sequencing facility (Ames, Iowa). The chimeric gene was removed from the plasmid by digestion with XbaI and SalI and subcloned into the low copy vector pRS313 at the same restriction sites (18).
The forward and reverse primers used in the PCR reaction were: bx_F 6 ⅐pri: CTT CTA TCC ACT TCG TGG AAA AGA TGT ATG GCC TCA aat aag gag ctt gat cct gtg and bx_F 6 ⅐rev⅐pri: CTT CCC TTG GTT TAA GCT TTA TTT CTT CTA gga ttg tgg ttt ctc gac, respectively. The lowercase letters correspond to the region that primes DNA synthesis in the PCR reaction using the cDNA of bF 6 as the template. The capital letters correspond to the target site for homologous recombination in the plasmid containing the gene encoding the ␤-subunit of the F 1 -ATPase (ATP2). The underlined sequence corresponds to the additional Ala-Ser codons. These were added because it is frequently the case that Ala-Ser are the first two amino acids after the processing site of the leader peptide. The Ala-Ser are thus positioned after and adjacent to the leader sequence of what corresponds to the leader peptide of the ␤-subunit of the ATPase.
Biochemical Procedures-Cells were grown aerobically at 28°C in a complete liquid medium containing 2% lactate as carbon source (19) and harvested in logarithmic growth phase. Mitochondria were prepared according to (20), frozen as droplets in liquid nitrogen, and stored at Ϫ70°C. Protein amounts were determined according to Lowry et al. (21) in the presence of 5% SDS using bovine serum albumin as a standard. Oxygen consumption rates were measured with NADH as substrate (22). Variations of transmembrane potential (⌬⌿) were evaluated by measurement of fluorescence quenching of rhodamine 123 with an SFM Kontron fluorescence spectrophotometer (23). The specific ATPase activity was measured at pH 8.4 according to Somlo (24) and modified as follows. Freshly prepared mitochondria were diluted with the same volume of either the isolation buffer (0.6 M mannitol, 2 mM EGTA, 10 mM Tris-maleate, pH 6.8) or 0.75% Triton X-100 (w/v). Aliquots were taken for protein concentration measurement, and determination of ATPase activity was done as follows: mitochondrial protein (50 g) was incubated for 2 min in reaction medium (0.9 ml, 0.2 M KCl, 3 mM MgCl 2 , 10 mM Tris-HCl, pH 8.4) in the presence or absence of F 0 inhibitors. The reaction was started with the addition of 5 mM ATP and stopped after 2 min by the addition of 0.3 M trichloroacetic acid.
Mitochondrial Triton X-100 extracts were sedimented on sucrose gradients as follows. Mitochondrial protein (3 mg at 10 mg/ml) was incubated with an equal volume of 0.75% (w/v) Triton X-100. The extract was spun at 100,000 ϫ g for 10 min at 4°C. The supernatant was loaded on the top of a 10 -30% linear sucrose gradient containing 0.1% Triton X-100, 1 mM ATP, 1 mM EDTA, and 10 mM Tris-HCl, pH 8.0. The gradient was centrifuged at 4°C for 17 h, 46 min in a SW41 rotor (Beckman) at 41,000 rpm, and fractions (1 ml) were collected and used for analysis.
Cross-linking Experiments and Analyses-Frozen mitochondria isolated from wild type and mutant cells were thawed, washed twice with 0.6 M mannitol, 2 mM EGTA, 50 mM triethanolamine-HCl, pH 8.0, and suspended in the same buffer at a protein concentration of 10 mg/ml. Samples were incubated for 30 min at 30°C in the presence or absence of DSP dissolved in dimethylformamide, and the reaction was quenched by the addition of 10 mM Tris, pH 8.0. The ATP synthase was extracted with 0.375% (w/v) Triton X-100, and after incubation for 15 min at 4°C, the extract was spun at 100,000 ϫ g for 15 min at 4°C, and the supernatant was used for analysis.
Bovine F6 was released from the ATP synthase (free bF 6 ) by sonication as follows. Washed mitochondria from yeast, ⌬ATP14 ϩ bF 6,, were suspended in 100 mM NaCl, 22 mM triethanolamine, pH 8.0, at a final concentration of 10 mg protein/ml and sonicated four times for 30 s at 120 V (Annemasse sonicator). The sample was centrifuged at 435,000 ϫ g for 1 h at 4°C. Aliquots of the supernatant containing free bF 6 were incubated with DSP and analyzed either by Western blot or loaded on a 10 -30% linear sucrose gradient, as described above.
SDS-polyacrylamide gel electrophoresis was according to Schä gger and Von Jagow (25). Nitrocellulose membranes (Membrane Protan BA83, 0.2 m from Schleicher & Schuell) were used for Western blot analyses. The polyclonal antibodies, anti-bF 6 , were kindly provided by Dr. Youssef Hatefi (Scripps Research Institute, La Jolla, CA) and used with a dilution of 1:7,500. Antibodies against subunit 4 and the ␣-subunit were used with dilutions of 1:10,000 and 1:100,000, respectively. Membranes were incubated with peroxidase-labeled antibodies and visualized with the ECL reagent of Amersham Pharmacia Biotech. Molecular mass markers (Benchmark Prestained Protein Ladder) were from Life Technologies, Inc. Sequence alignments were performed as in (26). Secondary structure predictions were made with the program at the PredictProtein Server.

RESULTS
The Coding Sequence of the Bovine Coupling Factor 6 Complements a Null Mutant Strain Devoid of Subunit h-For expression in yeast, the cDNA encoding mature bF 6 was fused to the leader sequence of the yeast ␤-subunit. In addition, the codons for Ala-Ser were added to the front end of the codons encoding the mature bF 6 , as described (17). The expectation is that leader sequence will be cleaved in front of the Ala-Ser residue after import into the mitochondrion providing an Nterminal sequence of: ASNKELD . . .
The strain containing the null mutation in the ATP14 gene (14), which encodes subunit h, was transformed by a single copy vector (pRS313) bearing the selection marker HIS3 and the DNA encoding the bovine F 6 , pbF 6 . The transformants were tested for growth on a complete medium containing glycerol as the sole carbon source, YPG (3% glycerol, 2% peptone, 1% yeast extract). Growth on medium containing glycerol or lactate indicates that the cells are able to make ATP via oxidative phosphorylation and thus have a functional ATP synthase. All of the transformants were able to grow on YPG at 28°C but not at 37°C (Fig. 1). One of the transformants was selected, named ⌬ATP14 ϩ bF 6 , and this strain had a generation time of 300 min as compared with 164 min for the wild type strain, at 28°C in liquid medium containing lactate as the carbon source. Loss of the plasmid, by growing the cells on a complete medium with glucose as carbon source, resulted in a concomitant loss in the ability to grow on glucose minimal medium devoid of histidine and on YPG medium (not shown). Thus, these results indicate a functional homology between yeast subunit h and bF 6 .
Phenotypic Analyses of the Complemented Yeast Strain-Mitochondria were prepared from ⌬ATP14 ϩ bF 6 strain to examine the effectiveness of bF 6 in replacing subunit h in the structure and function of the ATP synthase. Respiration rates were measured with NADH as substrate (Table I). In the presence of CCCP, the uncoupled respiration rates of mitochondria isolated from the wild type and ⌬ATP14 ϩ bF 6 strains were similar. The main difference between the two preparations is the respiration rate associated with the phosphorylation of ADP (State 3) where the respiration rate of ⌬ATP14 ϩ bF 6 mitochondria is only 51% of the uncoupled respiration rate, as compared with 63% for the wild type. As a result, the respiratory control and the ATP/O ratio of ⌬ATP14 ϩ bF 6 mitochondria are lower than those of wild type mitochondria, thus reflecting a lower ATP synthase activity and a lower efficiency of oxidative phosphorylation.
At pH 8.4, the ATPase activity of ⌬ATP14 ϩ bF 6 mitochondria is partially inhibited by DCCD but not by oligomycin, both of which are inhibitors of F 0 (Table II). Addition of Triton X-100 (0.375%) solubilizes the ATP synthase and increases the ATPase activity of wild type mitochondria by removing IF 1 , the F 1 inhibitor protein. This stimulation also occurred for the ⌬ATP14 ϩ bF 6 mitochondrial extract, but the sensitivity to DCCD was nearly abolished. These data suggest that the ATP synthase isolated from ⌬ATP14 ϩ bF 6 mitochondria was highly unstable, consistent with the temperature-sensitive phenotype of the cells on YPG medium (Fig. 1).
Potential (⌬⌿) measurements were performed in the respiration medium (Fig. 2). With ethanol as a substrate, the addition of ADP promoted a transient decrease in the fluorescent quenching of rhodamine 123 because of proton uptake through F 0 during ATP synthesis. This effect was less pronounced with mitochondria isolated from ⌬ATP14 ϩ bF 6 consistent with the low rate of state 3 respiration. The reversibility of the ATP synthase was also examined. Ethanol addition promoted a strong fluorescence quenching of rhodamine 123 because of the respiratory chain activity, and this was reversed by potassium cyanide. Finally, addition of ATP caused a fluorescent quenching because of its hydrolysis and the coupled pumping of protons out of the mitochondrion. Clearly, the ⌬⌿ generated by ATP hydrolysis in mitochondria isolated from ⌬ATP14 ϩ bF 6 is sensitive to either oligomycin or DCCD at concentrations used to inhibit the wild type mitochondria, but the intensity of the fluorescent quenching was not as large, and it was not as stable as compared with the wild type sample. These results suggest that the ATP synthase from ⌬ATP14 ϩ bF 6 is uncoupling during the course of the assay. However, the addition of oligomycin or DCCD did not affect the transmembrane potential generated by the respiratory chain, suggesting that passive proton conduction through F 0 is not occurring in the absence of either ADP or ATP.
Bovine F 6 Associates with the Yeast ATP Synthase-The prior genetic and biochemical studies indicate that bF 6 can correct for the loss of subunit h, albeit not to wild type levels. If bF 6 is acting directly by substituting for subunit h in the ATP synthase, bF 6 should be associated with the ATP synthase devoid of subunit h. A number of biochemical studies were performed to test this hypothesis. First Western blot analysis using anti-bF 6 antiserum detected a band with a relative mo-lecular mass of 10.6 kDa that was present in the SDS-dissociated ⌬ATP14 ϩ bF 6 , but not wild type, mitochondrial sample (Fig. 3A). The mature bF 6 has a calculated molecular mass of 9,116 Da assuming that processing occurs just prior to the Ala-Ser residues. If the protein is not processed, other than removing the initiating Met, then it would have a mass of 12,653 Da. These results suggest the 10.6-kDa peptide represents bF 6 and is likely processed at the predicted point. However, N-terminal sequence analysis has not been done on the yeast-expressed bF 6 to verify the site of processing.
Association of bF 6 to the yeast ATP synthase missing subunit h was determined by immunoprecipitation of the detergent (Triton X-100)-solubilized ATP synthase complex using antibodies directed against the ␣-subunit. However, unlike ATP synthase from wild type yeast, the antibodies precipitated only the F 1 sector (not shown). This suggested that the association of F 1 and F 0 sectors was not stable during immunoprecipitation. This is consistent with the prior studies that indicated that the ATP synthase with bF 6 was not as stable as the wild type enzyme. To capture the interactions between bF 6 and components of the ATP synthase, the mitochondrial proteins were treated with the thiol-cleavable homobifunctional crosslinking reagent, DSP. Fig. 3B shows that the concentration of bF 6 was greatly decreased upon incubation of mitochondria with 0.2 and 0.5 mM DSP, concurrent with the presence of bands of high molecular masses, as revealed by Western blot analysis. The most intense band displayed a relative molecular mass of 36 kDa, suggesting that bF 6 (9.1 kDa) was associated with a peptide of about 27 kDa. Incubation of DSP-treated ⌬ATP14 ϩ bF 6 mitochondria with 2-mercaptoethanol to reverse the cross-linking, eliminated the major 36-kDa product, and provided a relative increase in the amount of uncrosslinked bF 6 . In the presence of 2-mercaptoethanol, nonspecific bands are observed in the 45-kDa region. These bands probably originate from aggregates that did not enter the gel in the absence of the reducing agent but did enter the gel in the presence of the reducing agent. Free bF 6 was also prepared from mitochondria isolated from yeast ⌬ATP14 ϩ bF 6 to exam-

TABLE II
ATPase activities of isolated mitochondria D273-10B/A/H/U (wild type) and ⌬ATP14 ϩ bF 6 mitochondria were prepared from yeast cells grown with 2% lactate as carbon source. Two different mitochondrial preparations were made for each strain, and measurements were performed in triplicate with 50 g of mitochondrial protein. Samples treated in the presence or absence of 0.375% Triton X-100 were prepared as described under "Experimental Procedures." Oligomycin (6 g/ml) and DCCD (6 g/ml) were added where indicated.  ine the behavior of bF 6 upon incubation with DSP in the absence of ATP synthase (Fig. 3C). The amount of free bF 6 decreased upon incubation with DSP, and this was reversed with 2-mercaptoethanol. Importantly, DSP cross-linking of free bF 6 did not provide any specific bands of higher molecular masses. This suggests that nonspecific cross-linking of free bF 6 occurred with other peptides, and these were reversed by reduction with 2-mercaptoethanol.
To determine whether bF 6 was cross-linked to the ATP synthase, mitochondria treated with DSP were extracted with Triton X-100, and the solubilized proteins were separated by sucrose-gradient centrifugation. Eleven fractions were collected and analyzed by Western blot. The samples were reduced before SDS-polyacrylamide gel electrophoresis to allow the clear identification of bF 6 . The blots were probed with antibodies raised against the ␣-subunit (an F 1 subunit), subunit 4 (an F 0 subunit), and bF 6 . The wild type ATP synthase sediments to fractions 4 -8 as shown by the cosedimentation of the ␣-subunit and subunit 4 in the sucrose gradient (Fig. 4A). For the ATP synthase from mitochondria isolated from yeast ⌬ATP14 ϩ bF 6, the ␣-subunit and subunit 4 were distributed in a much broader range of fractions, suggesting both a more complex mixture of species and a less stable assembly of the ATP synthase. This occurred in the absence of reaction with DSP (not shown) or in the presence of 0.2 mM DSP (Fig. 4B). However, when the concentration of DSP was increased to 0.5 mM, then a significant fraction of bF 6 was seen to sediment with the other subunits of the ATP synthase (Fig. 4C). This new location of bF 6 in the sucrose gradient was not the result of nonspecific cross-linkings of bF 6 to other proteins, because reaction of DSP with bF 6 , which has been freed and separated from F 1 F 0 stayed at the top of the gradient, the same position of bF 6 without reaction with DSP (Fig. 4, D and E). In conclusion, the results support the hypothesis that bF 6 associates to the yeast ATP synthase.
DSP Primarily Cross-links Bovine F 6 to Subunit 4 -If bF 6 was acting directly by replacing subunit h, then not only should it associate with the ATP synthase, but it should also interact with the same peptides of the ATP synthase as subunit h. To test this hypothesis, the protein partners of bF 6 were determined by Western blot analysis after cross-linking the peptides with DSP. The results from Fig. 3 indicate that the major cross-linked species with bF 6 had a molecular mass of about 36 kDa, suggesting that bF 6 was cross-linked to a peptide 20 -30 kDa in size. As a consequence, antibodies raised against subunits of the ATP synthase whose molecular mass were within this range, OSCP and subunits 6, d, and 4, were used in the analysis. In addition, long polyacrylamide slab gels were used to clearly identify the cross-linked products (Fig. 5).
In a wild type context, numerous cross-linked products involving subunit 4 have been reported, and the most intense band that was found in the 36-kDa region has been identified as 4ϩg (15). This same cross-linked product was present for both the ATP synthase from the wild type and ⌬ATP14 ϩ bF 6 strains (Fig. 5, A and B). However, there was an additional band at about 36 kDa that was obtained upon incubation with 0.2 and 0.5 mM DSP and that was seen with antibodies against either subunit 4 or bF 6 . This band ran slightly ahead of the 4ϩg cross-linked product, and it was absent from the wild type  6 to putative components of the ATP synthase. A, wild type (wt) and ⌬ATP14 ϩ bF 6 mitochondria were dissociated and analyzed by Western blot. Blots were probed with polyclonal antibodies raised against subunits 4 and bF 6 . B, ⌬ATP14 ϩ bF 6 mitochondria (30 g of protein) was cross-linked with 0.2 and 0.5 mM DSP and treated with 2-mercaptoethanol, as indicated, and the blot was probed with antibodies raised against bF 6 . (C) bF 6 was released from F 1 F 0 by sonication (see "Experimental Procedures") and treated with DSP and 2-mercaptoethanol, as indicated, and the blot was probed with polyclonal antibodies raised against bF 6 . sample (Fig. 5B). This band is thus concluded to be a crosslinked product of bF 6 and subunit 4.
Two other bands of low intensity and showing relative molecular masses of 42 and 56 kDa could be heterooligomers containing at least bF 6 and subunit d because the latter crosslinked products were absent from the wild type sample (Fig.  5C). Antibodies raised against subunit 6 and OSCP did not identify any bands that were also detected by antibodies against bF 6 (not shown). The increased intensity of a 26-kDa band upon cross-linking of the ⌬ATP14 ϩ bF 6 mitochondria represents a cross-link involving subunit d and a small component of the yeast ATP synthase (Fig. 5C). The latter band was less apparent in the wild type sample. Thus, these results indicated that bF 6 associated with subunit 4 of the ATP synthase consistent with the interactions of subunit h in the ATP synthase.

DISCUSSION
Subunit h is an essential component of the yeast ATP synthase. It has been described as a supernumerary protein that, until this study, was not apparently related to any subunit described in other ATP synthases (14). Primary structural analysis of nucleotide data banks have identified the existence of an open reading frame in Schizosaccharomyces pombe (TrEMBL accession number 059673) that encodes an hypothetical protein 27% identical to subunit h and another in Botryotinia fuckeliana (GenBank TM accession number AL115386) that also encodes a hypothetical protein 38% identical to subunit h. Yeast subunit h is essential for yeast to grow on a nonfermentable carbon sources, and mitochondria isolated from a yeast strain with a null mutation in subunit h have an ATPase activity that is oligomycin-insensitive and the catalytic sector dissociated from the membrane components (14).
Mammalian coupling factor F 6 is an essential component of the mammalian mitochondrial ATP synthase. F 6 is known to be required for restoration of ATP-P i exchange and oligomycinsensitive ATPase activity to factor 6-depleted ATP synthase (27,28). It is also involved in the binding of F 1 to F 0 (29) and shields F 1 from limited proteolysis (30). As such, F 6 is thought to be required for the coupling of proton translocation to the synthesis of ATP.
The results in this manuscript are quite surprising and have important implications. Both genetic and biochemical data indicate that subunit h of the yeast ATP synthase is the homolog of mammalian coupling factor F 6 . This is a rather surprising because both of these subunits are essential components of their respective multimeric peptide complexes. Despite the apparent functional homology of subunit h with F 6 , primary se-FIG. 6. Primary structure alignment of S. cerevisiae subunit h and bovine F 6 . Sequence alignment of mature subunit h and bF 6 (Swiss-Prot accession numbers Q12349 and P02721, respectively) was performed according to Risler et al. (26). The amino acids conserved are indicated by an asterisk, whereas double and single dots indicate conservative and semi-conservative substitutions, respectively.

FIG. 4. Sedimentation analysis of bF 6 in a sucrose gradient.
Mitochondrial Triton X-100 ) extracts (Ext.) were separated on a 10 -30% sucrose gradient. Protein from fractions (30 l) was separated by SDS-polyacrylamide gel electrophoresis, electroblotted, and probed with antibodies raised against the ␣-subunit, bF 6 and subunit 4, as indicated. A, wild type mitochondrial Triton X-100 extract. B and C, Triton X-100 extracts of ⌬ATP14 ϩ bF 6 mitochondria were treated with 0.2 mM DSP (B) and 0.5 mM DSP (C). D and E, bF 6 was released from F 1 F 0 by sonication and incubated without (D) or with 0.5 mM DSP (E). Note that ␣* is a degradation product of the ␣-subunit. quence alignment of both subunits shows a very low sequence identity of just 14.5% and, when allowing for amino acid replacements, a 54% similarity (Fig. 6). This low level of sequence identity and homology is at the level seen between two random peptides. In contrast, most of the remaining subunits of the ATP synthase demonstrate a high degree of identity (31). For instance, the ␣-and ␤-subunits of the yeast ATP synthase are highly conserved with percent identities of about 72 and 75%, respectively. Thus, it is surprising that bF 6 is able to replace subunit h and form a functional enzyme, even now, knowing that they are functional homologs.
The biochemical studies here demonstrate that the complementation by bF 6 is due to the direct replacement of subunit h with bF 6 and not due to a secondary mechanism. The biochemical studies indicate that bF 6 occupies that same spatial relationship in the yeast enzyme as subunit h. Cross-linking products involving subunit h and subunit 4, a component of the second stalk, were obtained from positions K98C (15) and Q203C of the latter subunit, 2 which are two positions located in the hydrophilic part of subunit 4, thus suggesting that subunit h also participates to form the second stalk or stator. Bovine F 6 is a component of F 0 (11,32), and it is associated with the stalk as shown by reconstitution experiments (33). Nearest neighbors relationships have been demonstrated, by cross-linking experiments, between F 6 and the b-subunit (subunit 4) (34) and bF 6 with the ␣and ␤-subunits (35). The data in this study indicate that subunit 4 is the major peptide cross-linked with bF 6 using DSP within the yeast enzyme missing subunit h. Thus, bF 6 and subunit h appear to occupy the same spatial arrangement in the enzyme, and bF 6 can directly replace subunit h in the yeast enzyme.
Despite the lack of primary structural similarity, subunit h and bF 6 must share some structural features that define them as functional homologs, and indeed, there are some features that are conserved. First, they are both relatively small peptides with calculated molecular masses of 10,408 and 8,958 Da for subunit h and bF 6 , respectively. Second, both subunit h and bF 6 are acidic proteins showing pI of 4.06 and 5.23, respectively. Third, secondary structural computer analyses predict two ␣-helix regions for both subunits: the first ␣-helix formed by residues 6 -23 for bF 6 and by residues 2-13 for subunit h and the second ␣-helix formed by residues 34 -50 for bF 6 and residues 47-64 for subunit h. Fourth, they both have an acidic tail, although it is slightly longer in subunit h. Thus, these conserved features may form some of the basis required for the functional replacement of bF 6 for subunit h. Of course, beyond these features, it is possible that despite low primary sequence identity, that these peptides fold into similar three-dimensional structures.
The partial complementation of the null mutant atp14 by the bovine coupling factor 6 is more consistent with the fact that F 6 or h are important for stabilizing the ATP synthase than for a mechanistic role. Of all the F 0 components, bF 6 and subunit h are the only acidic proteins. One possible hypothesis is that these highly negative charged proteins help in the association of other positively charged components of the second stalk, such as subunits 4, d, and OSCP, three proteins with calculated pIs of 7.83, 8.92, and 9.3, respectively.
Expressions of other subunits of the mammalian ATP synthase have been demonstrated to complement the corresponding null mutations in yeast. Expression of bovine ␣-, ␤-, ␥-, and ⑀-subunits (17) and rat OSCP (36) have all complemented the corresponding null mutant strains in yeast. However, although some of these homologous peptides did not show a large amount of identity, it was always high enough to suggest them as homologs by simple primary structural analysis. Subunit h and bF 6 are so divergent that even a one on one comparison of their primary structure provided no clue that they were indeed homologs. The results of this study are even more startling because these peptides are not the sole peptide in an enzyme complex, but must interact within a heterosubunit multimeric enzyme complex. The implications of this are significant because they indicate that primary structural analysis cannot be used as the sole evidence that functional peptide homologs do not exist between two species. This is true even when the peptide is within a multimeric peptide complex that otherwise might be highly conserved.