Cytidine deaminases from B. subtilis and E. coli: compensating effects of changing zinc coordination and quaternary structure.

Cytidine deaminase from E. coli is a dimer of identical subunits (M(r) = 31 540), each containing a single zinc atom. Cytidine deaminase from B. subtilis is a tetramer of identical subunits (M(r) = 14 800). After purification from an overexpressing strain, the enzyme from B. subtilis is found to contain a single atom of zinc per enzyme subunit by flame atomic absorption spectroscopy. Fluorescence titration indicates that each of the four subunits contains a binding site for the transition state analogue inhibitor 5-fluoro-3,4-dihydrouridine. A region of amino acid sequence homology, containing residues that are involved in zinc coordination in the enzyme from E. coli, strongly suggests that in the enzyme from B. subtilis, zinc is coordinated by the thiolate side chains of three cysteine residues (Cys-53, Cys-86, and Cys-89) [Song, B. H., and Neuhard, J. (1989) Mol. Gen. Genet. 216, 462-468]. This pattern of zinc coordination appears to be novel for a hydrolytic enzyme, and might be expected to reduce the reactivity of the active site substantially compared with that of the enzyme from E. coli (His-102, Cys-129, and Cys-132). Instead, the B. subtilis and E. coli enzymes are found to be similar in their activities, and also in their relative binding affinities for a series of structurally related inhibitors with binding affinities that span a range of 6 orders of magnitude. In addition, the apparent pK(a) value of the active site is shifted upward by less than 1 unit. Sequence alignments, together with model building, suggest one possible mechanism of compensation.

Complementary truncations of a hydrogen bond to ribose involved in transition-state stabilization by cytidine deaminase.

The crystal structure of the complex formed between Escherichia coli cytidine deaminase and the transition-state analogue inhibitor 3,4-dihydrouridine [Betts, L., Xiang, S., Short, S. A., Wolfenden, R., & Carter, C. W. (1994) J. Mol. Biol. 235, 635] shows the presence of an H-bond between Glu-91 and the 3'-OH group of substituent ribose, a part of the substrate that is not directly involved in its chemical transformation. To test the contribution of this interaction to transition-state stabilization, Glu-91 was converted to alanine. The mutant enzyme is very much less active than the wild-type enzyme, with a 500-fold increase in Km and a 32-fold reduction in kcat using cytidine as substrate. No change in secondary structure is evident in the circular dichroic spectrum. As measured by kcat/Km, Glu-91 thus appears to stabilize the transition state for cytidine deamination by an overall factor of 1.7 x 10(4), equivalent to -5.8 kcal/mol in free energy. To test the contribution of this interaction in the opposite sense, the 3'-OH group of the substrate was replaced by a hydrogen atom. Comparing 3'-deoxycytidine with cytidine, the native enzyme shows a 17-fold increase in Km and a 400-fold decrease in kcat, indicating that the 3'-hydroxyl group of cytidine stabilizes the transition state for deamination by an overall factor of 6.3 x 10(3), equivalent to -5.2 kcal/mol in free energy, as measured by kcat/Km. After one binding partner has been removed, however, the effect of removing the remaining partner is relatively slight. For the mutant enzyme E91A, removal of the 3'-hydroxyl group from substrate cytidine reduces kcat/Km by a factor of only 3. Complete removal of substituent ribose reduces the wild-type enzyme's kcat/Km by a factor of more than 10(8); thus, substituent ribose, although distant from the site of chemical transformation of the substrate, contributes at least 11 kcal to the free energy of stabilization of the transition state for cytidine deamination, matching the apparent contribution to transition state binding made by the 4-OH group of the pyrimidine ring, which is at the site of substrate transformation [Frick, L., Yang, C., Marquez, V. E., & Wolfenden, R. (1989) Biochemistry 28, 9423].

Role of glutamate-104 in generating a transition state analogue inhibitor at the active site of cytidine deaminase.

The 19F-NMR resonance of 5-[19F]fluoropyrimidin-2-one ribonucleoside moves upfield when it is bound by wild-type cytidine deaminase from Escherichia coli, in agreement with UV and X-ray spectroscopic indications that this inhibitor is bound as the rate 3,4-hydrated species 5-fluoro-3,4-dihydrouridine, a transition state analogue inhibitor resembling an intermediate in direct water attack on 5-fluorocytidine. Comparison of pKa values of model compounds indicates that the equilibrium constant for 3,4-hydration of this inhibitor in free solution is 3.5 x 10(-4) M, so that the corrected dissociation constant of 5-fluoro-3,4-dihydrouridine from the wild-type enzyme is 3.9 x 10(-11) M. Very different behavior is observed for a mutant enzyme in which alanine replaces Glu-104 at the active site, and kcat has been reduced by a factor of 10(8). 5-[19F]Fluoropyrimidin-2-one ribonucleoside is strongly fluorescent, making it possible to observe that the mutant enzyme binds this inhibitor even more tightly (Kd = 4.4 x 10(-8) M) than does the native enzyme (Kd = 1.1 x 10(-7) M). 19F-NMR indicates, however, that the E104A mutant enzyme binds the inhibitor without modification, in a form that resembles the substrate in the ground state. These results are consistent with a major role for Glu-104, not only in stabilizing the ES++ complex in the transition state, but also in destabilizing the ES complex in the ground state.

Major contribution of a carboxymethyl group to transition-state stabilization by cytidine deaminase: mutation and rescue.

The crystal structure of an inhibitory complex formed between Escherichia coli cytidine deaminase and the transition-state analog 3,4-dihydrouridine indicates the presence of a short H-bond between Glu-104 and the inhibitor. To test the possibility that analogous H-bonds might play a significant role in stabilizing the hydrated substrate in the transition state for deamination, we replaced Glu-104 by alanine. Compared with the wild-type enzyme, the mutant enzyme's affinities for substrate cytidine and product uridine were found to have increased, whereas kcat for deamination of cytidine had been reduced by 8 orders of magnitude. By its presence, the carboxymethyl group of Glu-104 appears to minimize the activation barrier for deamination, not only by stabilizing the altered substrate in the transition state but also by destabilizing the enzyme-substrate and enzyme-product complexes. In the presence of added formate ion, but not in the presence of bulkier carboxylic acids, the low catalytic activity of the mutant enzyme was enhanced substantially.

Mutations affecting transition-state stabilization by residues coordinating zinc at the active site of cytidine deaminase.

Cytidine deaminase from Escherichia coli contains 1 mol of tightly bound zinc per enzyme subunit (Yang, C., Carlow, D., Wolfenden, R., & Short, S.A. (1992) Biochemistry 31, 4168-4174). When the metal liganding residues Cys-129 and Cys-132 were replaced by Ala, and His-102 was replaced by Ala, Asn, or Gln, deaminase activities of cell extracts containing these mutant enzymes were decreased by several orders of magnitude relative to that of the wild-type enzyme. After purification, each mutant protein was found to contain less than 0.2 mol of zinc per enzyme subunit, except mutant H102Q, which contained 1 mol of zinc per subunit. The activity of each mutant enzyme increased in the presence of added zinc but never attained wild-type activity. Mutant H102N was unique in that this protein could be purified as a stable apoenzyme, activated by added zinc, and then inhibited by EDTA. This mutant enzyme bound zinc with an apparent Kd value of 6.0 x 10(-10) M and regained maximal activity in the presence of 1 mol of zinc per subunit. Affinities of the mutant cytidine deaminases for the transition-state analogue, 5-fluoropyrimidin-2-one ribonucleoside (3,4) hydrate, were found to decrease in rough proportion to kcat/Km over a range spanning several orders of magnitude. This variation in catalytic efficiency arose mainly from effects on kcat, indicating the involvement of zinc coordination in the catalytic process rather than in substrate binding.