was unclear (52). The frequent reports of "two nucleotide" structures have bewildered

scientists due to the vast body of biochemical data from numerous laboratories, using a

variety of techniques, which establish indisputably that all three catalytic sites are readily

filled with nucleotide (32, 53). It is possible that the enzyme preferentially crystallizes in

a ground state intermediate which may occur after the release of product, leaving one site

empty and opened (54). Nevertheless, the accumulating structural data may be indicative

of several intermediary steps that may form during the synthesis of ATP. A detailed

account of the mechanism of ATP synthesis follows later in this chapter.

 The crystal structure does offer some insight of the chemical mechanism of ATP

synthesis (20). In the P subunit, 4.4 A+ from the terminal phosphate of the bound

nucleotide triphosphate, there is clearly a density for a water molecule hydrogen bonded

to the carboxylate of PglU1SS. This carboxylate is positioned to allow an inline

nucleophilic attack of the water molecule on the terminal phosphate. The guanidinium of

a neighboring residue, aarg373, iS thought to help stabilize the negative charge on the

terminal phosphate during the transition state (20). This same arrangement can be found

in the catalytic site of transducin-a (42). The crystal structure also provides some insight

as to why the nucleotide binding sites in the a subunit are noncatalytic. There is no

spacial equivalent of the carboxylate of PglU1SS in the a subunit. The spatial equivalence

in the a subunit is filled by a agln208, with the side chain pointed away from the terminal

phosphate (20). The binding of the adenine to the noncatalytic site of the a subunit is

highly specific, unlike the P subunit nucleotide binding site, which can accommodate

GTP, ITP as well as ATP (55, 56). This specificity is due to several hydrogen bonds as

well as the presence of the Ptyr368 ClOse to the 2-position of the adenine ring in the a