ATP sulfurylase from Penicillium chrysogenum was purified to homogeneity. The enzyme binds 8 mol of free ATP ( K s = 0.53 mM) or AMP ( K s = 0.50 mM) per 440,000 g. The results are consistent with our earlier report that the enzyme is composed of eight identical subunits of M r 55,000 ( J. W. Tweedie and I. H. Segel, 1971, Prep. Biochem. 1, 91–117; J. Biol. Chem. 246, 2438–2446 ). In the absence of cosubstrates, the purified enzyme catalyzes the hydrolysis of MgATP (to AMP and MgPP i) and adenosine 5′-phosphosulfate (APS) (to AMP and SO 4 2−). MgATP hydrolysis is inhibited by nonreactive sulfate analogs such as nitrate, chlorate, and formate (uncompetitive with MgATP). In spite of the hydrolytic reactions it is possible to observe the binding of MgATP and APS to the enzyme in a qualitative (nonequilibrium) manner. Neither inorganic sulfate (the cosubstrate of the forward reaction) nor formate or inorganic phosphate (inhibitors competitive with sulfate) will bind to the free enzyme in detectable amounts in the absence or in the presence of Mg 2+, Ca 2+, free ATP, or a nonreactive analog of MgATP such as Mg-α,β-methylene-ATP. Similarly, inorganic pyrophosphate (the cosubstrate of the reverse reaction) will not bind in the absence or in the presence of Mg 2+ or Ca 2+. The induced binding of 32P i (presumably to the sulfate site) can be observed in the presence of MgATP. The results are consistent with the obligately ordered binding sequence deduced from the steady-state kinetics ( J. Farley et al., 1976, J. Biol. Chem. 251, 4389–4397 ) and suggest that the subsites for SO 2− 4 or MgPP i appear only after nucleotide cleavage to form E~AMP · MgPP i or E~AMP · SO 4 complexes. The suggestion is supported by the relative values of K ia (ca. 1 m m for MgATP) and K iq (ca. 1 α m for APS) and by the inconsistent value of k −1 calculated from V f K i a K m A (The value is considerably less than V r) Purified ATP sulfurylase will also catalyze a Mg 32PP i-MgATP exchange in the absence of SO 4 2−. A 35SO 4 2−-APS exchange could not be demonstrated in the absence or presence of MgPP i. This result was not unexpected: The rate of APS hydrolysis (or conversion to MgATP) is extremely rapid compared to the expected exchange rate. Also, the pool of APS at equilibrium is extremely small compared to the sulfate pool. The V values for molybdolysis, APS hydrolysis (in the absence of PP i), ATP synthesis (from APS + MgPP i), and Mg 32PP i-MgATP exchange at saturating sulfate are all about equal (12–19 μmol × min −1 × mg of enzyme −1). The rates of Mg 32PP i-MgATP exchange in the absence of sulfate, APS synthesis (from MgATP + sulfate), and MgATP hydrolysis (in the absence of sulfate) are considerably slower (0.10 – 0.35 μmol × min −1 × mg of enzyme −1). These results and the fact that k 4 calculated from V r K i q K m Q is considerably larger than V f suggest that the rate-limiting step in the overall forward reaction is the isomerization reaction E~AMP-SO 2− 4 → E APS. In the reverse direction the rate-limiting step may be SO 2− 4 release or isomerization of the E~AMP · MgPP i · SO 4 2− complex. (The reaction appears to be rapid equilibrium ordered.) Reactions involving the synthesis or cleavage of APS are specific for Mg 2+. Reactions involving the synthesis or cleavage of ATP will proceed with Mg 2+, with Mn 2+, and, at a lower rate, with Co 2+. The results suggest that the enzyme possesses a Mg 2+-preferring divalent cation (activator) binding site that is involved in APS synthesis and cleavage and is distinct from the MeATP or MePP i site. The equilibrium binding of about one atom of 45Ca 2+ per subunit (possibly to the activator site) could be demonstrated ( K s = 1.4 mM).