The immunoglobulin-binding protein in vitro autophosphorylation site maps to a threonine within the ATP binding cleft but is not a detectable site of in vivo phosphorylation.

In vitro incubation of immunoprecipitated immunoglobulin-binding protein (BiP) complexes with calcium and [gamma-32P]ATP resulted in the phosphorylation of BiP on a threonine residue. This autophosphorylation activity did not occur in the presence of magnesium but had the same pH optimum as reported for its magnesium-dependent ATPase activity. This suggested the possibility that both activities could occur through ATP hydrolysis at the same site. In support of this, mutation of either Thr-37 or Thr-229 to a glycine eliminated both autophosphorylation and ATPase activities, and mutation of either residue to a serine significantly reduced both activities. Glutamic acid 175 in HSC71 has been hypothesized to flank the divalent cation complexed with ATP. Mutation of the analogous glutamic acid, Glu-201, in BiP abolished ATPase activity but still supported some autophosphorylation. The in vitro phosphorylation site was mapped to Thr-229 by mutational analysis. This threonine has been hypothesized to interact with the gamma-phosphate of ATP through a polarized water molecule and would be in a position to act as a phosphate acceptor in the ATP hydrolysis reaction. These data imply that both ATPase and autophosphorylation result from ATP hydrolysis at the same site and that the cation associated with BiP determines which activity is observed. Comparison of partial protease digestion or cyanogen bromide cleavage products of in vitro and in vivo phosphorylated BiP demonstrated that Thr-229 is not a detectable site of phosphorylation in cells. Therefore, whatever functional role phosphorylation may have in vivo, it cannot be attributed to autophosphorylation of Thr-229.