STRUCTURAL AND FUNCTIONAL STUDIES ON TRYPANOSOMA BRUCEI 6-PHOSPHOGLUCONATE DEHYDROGENASE

6-Phosphogluconate dehydrogenase (6PGDH), the third enzyme of the pentose phosphate pathway, catalyzes the NADP-dependent oxidative decarboxylation of 6-phosphogluconate (6PG) to ribulose-5-phosphate (RU5P). It has been validated as drug target for sleeping sickness and veterinary infections caused by protozoan parasites of the Trypanosoma brucei subspecies, in the sub-Saharan Africa. Current treatments are limited in availability, have toxic side effects and are difficult to administer. Out of six clinically approved drugs for the treatment of Human African Trypanosomiasis (HAT), five have been discovered more than 30 years ago, evidencing the need of new drugs for the control and eradication of this disease. Several substrate-related molecules have been tested to find inhibitors able to discriminate between the mammalian and the parasite enzyme, and one of these inhibitors, 4-phospho-erythronohydroxamate, has been modified in a membrane-permeable form, that, at least “in vitro”, kills the parasites. However the stabilities of the protecting groups are poor, thus reducing the bioavailability of this class of compounds. Structure-based molecular design is a powerful tool for developing new inhibitors, nevertheless this technique requires a detailed description of the site where the inhibitor should bind. The fixed crystallographic picture of 6-phosphogluconate dehydrogenase (6PGDH) is a poor and limited tool for a detailed description of the active site Much discrepancy exists with the ligand-induced changes in significant properties of the enzyme. In the first part of this thesis, about structural studies on T. brucei 6PGDH we report that the T. brucei 6PGDH displays a dimer-tetramer equilibrium. Glutaraldehyde cross-linking and dynamic light scattering show that the equilibrium is not perturbed by ligands. However gel filtration and sucrose density gradient evidence a drastic reduction of the association-dissociation rate induced by NADPH, with the substrate 6PG able to antagonize this effect. Instead NADP binding does not modify the oligomerization rate. The different behaviour of the two forms of coenzyme appears related to the large differences in the binding specific heat change (ΔCp), with the NADP binding ΔCp very close to that expected from crystallographic structure, while the NADPH binding ΔCp three times larger than that expected, which is consistent with tetramerization. Physiological significance of this new finding, suggesting alternation between dimer and tetramer during the catalytic cycle, might be to escape NADPH inhibition of this enzyme, whose role is also to deplete 6PG, since its accumulation induces cell senescence [V.P. Sukhatme, B. Chan, 2012]. The thesis second part is about functional studies on the enzyme. One important information that cannot be obtained from X-ray structures is the protonation state of the ionisable residues present in the target site. In the 6PG binding site two conserved residues, K185 and E192, have been shown to be essential for the catalytic activity. These residues are thought to be in the “natural” ionization state in the free enzyme, but with inverted protonation in the enzyme-substrate complex. Accordingly to this mechanism, the binding of 6PG to the enzyme should cause the inversion of the protonation state of K185 and E192. In detail K185 should become unprotonated while E192 protonated. Furthermore, in the proposed mechanism, K185 changes the ionization state in the course of the reaction. We used site-directed mutagenesis and isothermal titration calorimetry to get experimental evidence of the ionisation state and of the pKa of K185 and E192 in the free enzyme and in the enzyme-substrate complex. We further explored the ionisation state of H188 which is a conserved residue at 4-5 Å from E192. The T. brucei 6PGDH mutants were also characterized for cysteine reactivity in the absence and presence of the substrate since in the WT a large difference there is in the two conditions. Our results give experimental support to the proposed mechanism, in fact the lack of the proton-accepting group, E192, causes a greater H+ release, while the lack of the proton releasing group, K185, causes an H+ uptake. The binding of 6PG to 6PGDH shows a pH dependence indicating the requirement of a protonated group with pKa 7.07. On the basis of the observed Kd, no proton release is expected, while a proton uptake is expected at pH higher than 7.0. Instead at all pH tested the binding is accompanied by a release of hydrogen ions. The pKa of 6PG is 6.3, therefore if only the ionized form of the substrate binds to the enzyme, at pH 6.5 the proton release can be totally accounted for by the ionization of 6PG. The pKa decrease at pH lower than 7 suggests that unprotonated 6PG binds to the enzyme. However at pH >7.0, other group(s) must be involved to balance the observed protons release. At least 1.5 H+ are required to account for the proton uptake of the group with pKa 7.07 and the release of H+ observed experimentally. A scheme with three residues involved is proposed as a simplified model, including a residue with a pK of 7.07 binding a proton released by a second residue, while a third residue releases hydrogen ions in the buffer. Data fit with changes in these three pK upon substrate binding. Despite a pKa of 7 appears anomalous for a glutamate, it can be assigned to the glutamate192, in fact mutation of this residue with glutamine causes the disappearance of the pH dependence of the Kd. Furthermore the difference in H+ release between WT and E192Q mutant enzymes fit fairly well with the disappearance of a group whose pKa changes from 7.07 to 9.67, upon substrate binding. The pKa of K185 in the free enzyme can be evaluated from the proton uptake of the K185H mutant. In fact the differences in H+ release/uptake between WT and K185H mutant enzymes should correspond to the H+ released by K185. The data are consistent with a group changing its pKa from 9.3 to 7.17, therefore a pKa 9.3 can be estimated for K185 in the free enzyme. The differences in H+ release between WT and H188L mutant enzymes can be described with a group changing its pKa from 10.5 to 8.65. These pKa values are hardly attributable to the histidine 188, instead it is likely that arise from different small ionization changes due to the conformational changes induced by the substrate. In fact the more interesting property of this mutant is the absence of the effect of 6PG on the reactivity of thiol groups. In the WT enzyme the binding of 6PG suppress the reactivity of all cysteine residues, and this can be explained only by a conformational effect. K185H and E192Q mutants show, with few differences, the same 6PG effect on the reactivity of the cysteine residues, but in the H188L mutant the presence of 6PG does not changes significantly the reactivity of cysteine residues. For this reason we suggest that the differences in H+ release between WT and H188L mutant are due the conformational changes induced by the substrate. It has been shown that the reaction rate is mainly limited by a conformational change, preceding the chemical steps [Hanau S., Rippa M., Dallocchio F. & Barrett; 1996]. The calculated intrinsic binding enthalpy is -4.6 ± 0.23 kcal/mole, and this value, although it includes both experimental and fitting errors, clearly shows that a large part of the binding enthalpy is spent to put the enzyme in the correct protonation form. Accordingly to the proposed mechanism at the end of the reaction the protonation of the enzyme-product complex is the same of that of the free enzyme, therefore the energy spent during the binding of the substrate will be regained during the catalytic cycle. Furthermore in the bell K6PG-pH curve the two pK of 7 and 9.5 have always been attributed to K185 and E192 but, being so different from the natural pK of a lysine and glutamate, it was thought that they were only apparent pK and only a small fraction of the enzyme was in the right protonation state, able to display the catalytic activity. Our results show that the pK of K185 and E192 really can move after the substrate binding to the values of 7 and 9.5, respectively. Thus kinetic parameters with good approximation correspond to the true pK of key aminoacids in catalysis.