be a substrate of the protease FtsH, which will rapidly degrade the subunit if it is not in

its native state (151). It was readily labeled with TID, which is a hydrophobic carbene

generator that is believed to react from the nonpolar region of the lipid bilayer, but its

solubility properties made it unsuitable for analysis as was done with the c and b subunit

(106). Consequently, the amino acids in contact with the lipid phase of the bilayer were

not identified. Due to difficulties in obtaining high-resolution structural data, much of

what is known of the a subunit arises from mutational studies.

 Topology. Hydropathy analyses indicated five definite membrane-spanning

regions and one putative membrane span (121, 140, 152, 153). Much of what is known

of the a subunit structure and has come from the analysis of cysteine mutagenesis.

Greater than 50 cysteine substitutions, which resulted in a functional F1Fo ATP synthase,

were used in two kinds of experiments (154-158). Various maleimide derivatives were

used to search for the surface-accessible regions (154, 155). And double cysteine

mutations were used to search for disulfide formation between a-a and a-c (153). The

results supported the model in which the a subunit spans the membrane five times and the

fourth span, which includes aarg210 iS in contact with the second transmembrane ot-helix of

the c subunit (Figure 1-4A). Additionally, residues that were originally thought to be

located in the cytoplasm were not labeled, indicating that the six-membrane span model

was incorrect (132, 159).

 The location of the amino-terminus of the a subunit has also been very

controversial. A substantial amount of evidence indicates that the carboxyl terminus