Given the role of trypanothione in the redox defenses of pathogenic trypanosomal and leishmanial parasites, in contrast to glutathione for their mammalian hosts, selective inhibitors of trypanothione reductase are potential drug leads against trypanosomiasis and leishmaniasis. In the present study, the rational drug design approach was used to discover tricyclic neuroleptic molecular frameworks as lead structures for the development of inhibitors, selective for trypanothione reductase over host glutathione reductase. From a homology-modeled structure for trypanothione reductase, replaced in the later stages of the study by the X-ray coordinates for the enzyme from Crithidia fasciculata, a series of inhibitors based on phenothiazine was designed. These were shown to be reversible inhibitors of trypanothione reductase from Trypanosoma cruzi, linearly competitive with trypanothione as substrate and noncompetitive with NADPH, consistent with ping-pong bi bi kinetics. Analogues, synthesized to define structure-activity relationships for the active site, included N-acylpromazines, 2-substituted phenothiazines, and trisubstituted promazines. Analysis of Ki and I50 data, on the basis of calculated log P and molar refractivity values, provided evidence of a specially favored fit of small 2-substituents (especially 2-chloro and 2-trifluoromethyl), with a remote hydrophobic patch on the enzyme accessible for larger, hydrophobic 2-substituents. There was also evidence of an additional hydrophobic enzymic region available to suitable N-substituents of the promazine nucleus. Ki data also indicated that the phenothiazine nucleus can adopt more than one inhibitory orientation in its binding site. Selected compounds were tested for in vitro activity against Trypanosoma brucei, T. cruzi, and Leishmania donovani, with selective activities in the micromolar range being determined for a number of them.

Trypanothione reductase, an essential component of the anti-oxidant defences of parasitic trypanosomes and Leishmania, differs markedly from the equivalent host enzyme, glutathione reductase, in the binding site for the disulphide substrate. Molecular modelling of this region suggested that certain tricyclic compounds might bind selectively to trypanothione reductase without inhibiting host glutathione reductase. This was confirmed by testing 30 phenothiazine and tricyclic antidepressants, of which clomipramine was found to be the most potent, with a K(i) of 6 microM, competitive with respect to trypanothione. Many of these compounds have been noted previously to have anti-trypanosomal and anti-leishmanial activity and thus they can serve as lead structures for rational drug design.

Pyrimethamine acts by selectively inhibiting malarial dihydrofolate reductase-thymidylate synthase (DHFR-TS). Resistance in the most important human parasite, Plasmodium falciparum, initially results from an S108N mutation in the DHFR domain, with additional mutation (most commonly C59R or N51I or both) imparting much greater resistance. From a homology model of the 3-D structure of DHFR-TS, rational drug design techniques have been used to design and subsequently synthesize inhibitors able to overcome malarial pyrimethamine resistance. Compared to pyrimethamine (Ki 1.5 nM) with purified recombinant DHFR fromP. falciparum, the Ki value of the m-methoxy analogue of pyrimethamine was 1.07 nM, but against the DHFR bearing the double mutation (C59R + S108N), the Ki values for pyrimethamine and the m-methoxy analogue were 71.7 and 14.0 nM, respectively. The m-chloro analogue of pyrimethamine was a stronger inhibitor of both wild-type DHFR (with Ki 0.30 nM) and the doubly mutant (C59R +S108N) purified enzyme (with Ki 2.40 nM). Growth of parasite cultures of P. falciparum in vitro was also strongly inhibited by these compounds with 50% inhibition of growth occurring at 3.7 microM for the m-methoxy and 0.6 microM for the m-chloro compounds with the K1 parasite line bearing the double mutation (S108N + C59R), compared to 10.2 microM for pyrimethamine. These inhibitors were also found in preliminary studies to retain antimalarial activity in vivo in P. berghei-infected mice.

No abstract

Trypanothione reductase, the enzyme which in trypanosomal and leishmanial parasites catalyses the reduction of trypanothione disulphide to the redox-protective dithiol and has been identified as a potential target for rational antiparasite drug design, has been found to be strongly inhibited by tricyclic compounds containing the saturated dibenzazepine (imipramine) nucleus, with Ki values in the low micromolar range. This drug lead structure was designed by molecular graphics analysis of a three-dimensional homology model, focussing on the active-site. Inhibition studies were carried out to determine the effect of inhibitor structure on the inhibitory strength towards recombinant trypanothione reductase from Trypanosoma cruzi. Hansch analysis showed that inhibitory strength depended on terms in pi, pi 2 and sigma m indicating dependence on both lipophilicity and inductive effect for ring-substituted analogues of imipramine. The side-chain omega-aminoalkyl chain had to be longer than 2-carbon units for inhibition. The effect on inhibition strength of the substituent at the omega-amino position on the side-chain of the central ring nitrogen atom depended markedly on the detailed substitution pattern of the rest of the molecule. This provides kinetic evidence studies of multiple binding modes within a single, blanket binding site for the inhibitor with the tricyclic ring system in the general region of the hydrophobic pocket lined by Trp21, Tyr110, Met113 and Phe114. This aspect of the structural sensitivity of the precise active-site triangulation adopted by the inhibitor is probably a function of the use of hydrophobic interactions of low directional specificity in this pocket combined with an electrostatic anchoring by the omega-N+ HMe2 function of the inhibitor, presumably with a glutamate side-chain, such as Glu-18, Glu-466' and/or Glu-467'.

We have used one-dimensional (1D) and two-dimensional (2D) proton nuclear magnetic resonance spectroscopy at 600 MHz for structural analysis of the complex formed between d(CGCGAATTCGCG)2 and 2-[2-(3-hydroxyphenyl)-6-benzimidazoyl]-6-(1-methyl-4-piperazinyl) benzimidazole (meta-Hoechst). This analogue differs from Hoechst 33258 only in the location of its meta rather than para phenolic hydroxyl group and was designed to introduce the possibility of intermolecular hydrogen bonding to DNA via the phenol. Complex formation was shown to be 1:1 at 25 degrees C in phosphate buffer in D2O by 1D NMR spectroscopic titration of a solution of d(CGCGAATTCGCG)2 with meta-Hoechst. From 1D NMR spectroscopy the observed perturbations of the assigned chemical shifts of the oligonucleotide observed on binding meta-Hoechst could be used to locate the ligand in the central AATT stretch. By means of 2D NMR spectroscopic techniques, over 400 proton-proton NOEs were defined within the complex. DNA nonexchangeable resonance assignments were made using the sequential assignment method and NOESY. Binding the unsymmetrical ligand lifted the C2v symmetry of the DNA. Exchangeable hydrogens were assigned from NOESY data acquired in 85% H2O/15% D2O medium for the complex and showed differences between the Hoechst 33258 and meta-Hoechst complexes with d(CGCGAATTCGCG)2. The location of meta-Hoechst in the minor-groove AATT region was triangulated using 32 intermolecular NOEs determined for the complex. From the intermolecular NOEs involving the aromatic C-H protons of the phenolic ring of meta-Hoechst, it was clear that this region of the molecule did not rotate freely within the minor groove on the NMR time scale and was oriented with its hydroxyl group toward the floor of the minor groove, in line with the occurrence of the predicted hydrogen bonding between it and the DNA. The pKa of the N3H proton of meta-Hoechst in its bound state in this complex was measured as 6.1 by NMR spectroscopy, a value slightly elevated relative to estimates (approximately 5.2) of the pKa of this proton for the free ligand. Molecular mechanics and the distance restraints provided by the intermolecular NOEs were used in molecular modeling of the meta-Hoechst/d(CGCGAATTCGCG)2 complex, and the distances in the model were consistent with the formation of hydrogen bonds involving the m-OH group of meta-Hoechst and the DNA.

By introducing cationic charge sites novel peptide lead inhibitor structures for trypanothione reductase have been designed using molecular modelling methods. The inhibitors showed reversible, linear competitive inhibition and the strongest peptide inhibitor to date was found to be N-benzyloxycarbonyl-Ala-Arg-Arg-4-methoxy-beta-naphthylamide with a Ki value of 2.4 microM and a selectivity for parasitic enzyme (trypanothione reductase) over the host enzyme (human glutathione reductase) of over 3 orders of magnitude.

No abstract