On the intermediacy of carboxyphosphate in biotin-dependent carboxylations

In the ATP-dependent carboxylation of biotin that is catalyzed by most biotin-dependent carboxylases, a fundamental mechanistic question is whether the ATP activates bicarbonate (via the formation of carboxyphosphate as an intermediate) or whether the ATP activates biotin (via the formation of O-phosphobiotin). We have resorted to three mechanistic tests using the biotin carboxylase subunit of acetyl-CoA carboxylase from Escherichia coli: positional isotope exchange, intermediate trapping, and 18O tracer experiments on the ATPase activity. First, no catalysis of positional isotope exchange in adenosine 5'-[( alpha, beta-18O, beta, beta-18O2]triphosphate) was observed when either biotin or bicarbonate was absent, nor was any exchange seen in the presence of both N-1-methylbiotin and bicarbonate. Second, the putative carboxyphosphate intermediate could not be trapped as its trimethyl ester, under conditions of incubation and analysis where the authentic triester was shown to be adequately stable. In the third test, however, we showed that the ATPase activity of biotin carboxylase that is seen in the absence of biotin, an activity that is known to parallel the normal carboxylase reaction when biotin is present, occurs with the transfer of an 18O label directly from [18O]bicarbonate into the product Pi. This result suggests that the bicarbonate-dependent biotin-independent ATPase reaction catalyzed by biotin carboxylase goes via carboxyphosphate and that the carboxylation of biotin itself may proceed analogously.