resides in the cytoplasm (154, 155, 160). This observation, in combination with the five-

transmembrane helices, indicates that the amino-terminus should reside in the

periplasmic space. Polyclonal antibodies against a peptide model of the extreme

carboxyl-terminus as well as antibodies against epitope tags constructed at the carboxyl

terminus of the a subunit revealed this region to be located in the cytoplasm (160).

Moreover, cysteine substitutions at a266 Of G277 were highly reactive on the cytoplasmic

side of the membrane (154, 155). The orientation of the amino and carboxyl-termini was

studied by gene fusion proteins and peptide-directed antibodies, revealing a cytoplasmic

location of both termini (161, 162). Insertion of epitope tags at various positions also

confirmed the cytoplasmic local of both termini, arguing in favor of the controversial six-

transmembrane model of the a subunit (Figure 1-4B) (160). In the five-transmembrane

model a stretch of about 37 amino acids at the amino-terminus resides in the periplasm

with only one transmembrane helix, approximately up to residue a66, preSent (Figure 1-

4A) (154-156, 158). In the six-transmembrane model the amino-terminus resides in the

cytoplasm with two transmembrane helices present before the first cytoplasmic loop,

which range from approximately residues a33-49 and a54-70 (Figure 1-4B) (160).

 A series of a subunit amino-terminal truncations and internal deletions were

constructed and the F1Fo ATP synthase function was tested by growth on a succinate

minimal media. Assembly of intact complexes was tested by membrane-associated

ATPase activity and the presence of the a subunit was analyzed by immunoblot analysis

(163). Four sections were found to be particularly interesting. The first 33 residues at the

amino terminus were shown to be necessary for the insertion of the a subunit into the

membrane. Two internal deletions, from residues a91-99 and al63-177, TOSulted in functional