A 'specificity' pocket inferred from the crystal structures of the complexes of aldose reductase with the pharmaceutically important inhibitors tolrestat and sorbinil.

BACKGROUND: Aldose reductase (AR) is an NADPH-dependent enzyme implicated in long-term diabetic complications. Buried at the bottom of a deep hydrophobic cleft, the NADPH coenzyme is surrounded by the conserved hydrophilic residues of the AR active site. The existence of an anionic binding site near the NADP+ has been determined from the structures of the complexes of AR with citrate, cacodylate and glucose-6-phosphate. The inhibitor zopolrestat binds to this anionic site, and in the hydrophobic cleft, after a change of conformation which opens a 'specificity' pocket. RESULTS: The crystal structures of the porcine AR holoenzyme and its complexes with the inhibitors tolrestat and sorbinil have been solved; these structures are important as tolrestat and sorbinil are, pharmaceutically, the most well-studied AR inhibitors. The active site of the holoenzyme was analyzed, and binding of the inhibitors was found to involve two contact zones in the active site: first, a recognition region for hydrogen-bond acceptors near the coenzyme, with three centers, including the anionic site; and second, a hydrophobic contact zone in the active-site cleft, which in the case of tolrestat includes the specificity pocket. The conformational change leading to the opening of the specificity pocket upon tolrestat binding is different to the one seen upon zopolrestat binding; this pocket binds inhibitors that are more effective against AR than against aldehyde reductase. CONCLUSIONS: The active site of AR adapts itself to bind tightly to different inhibitors; this happens both upon binding to the inhibitor's hydrophilic heads, and at the hydrophobic and specificity pockets of AR, which can change their shape through different conformational changes of the same residues. This flexibility could explain the large variety of possible substrates of AR.

Properties of horse liver alcohol dehydrogenase modified by the affinity label 3-chloroacetylpyridine-adenine dinucleotide.

The reactive analogue of NAD+, CPAD+, was incorporated in the horse liver alcohol dehydrogenase (EC 1.1.1.1) linearly with its inactivation, to stoichiometry, with no apparent subunit interaction. No hydride transfer could take place in the modified enzyme, nor the interaction of trans-4-N,N-dimethylaminocinnamaldehyde with its reduced form, indicating an impairment of the accessibility to the catalytic zinc atom. The labeling in the enzyme, alkylated by [carbonyl-14C]CPAD+ was not stable, with a half-life of 32 h.

Mechanism-based inhibition of proline dehydrogenase by proline analogues.

The inactivation of proline dehydrogenase by several L-Pro analogues was investigated with the aim to block the essential metabolic pathway of tsetse flies allowing the degradation of L-Pro to L-Glu. In vitro studies on rat liver mitochondria showed that only 4-methylene-L-proline was able to inactivate proline dehydrogenase. The inactivation kinetics agreed with a mechanism-based inhibition. The other tested analogues E- and Z-4-fluoromethylene-L-proline, and cis and trans-5-ethynyl-D,L-proline were neither substrate nor inactivator of the enzyme. In vivo 4-methylene-L-proline showed no toxicity against Drosophila flies, but was lethal for Glossina pallidipes flies. This result allows the consideration of 4-methylene-L-proline as an attractive compound molecule in the struggle against tsetse flies.

Purification and electrospray mass spectrometry of aldose reductase from pig lens.

Aldose reductase (alditol: NADP+ 1-oxidoreductase, EC 1.1.1.21) has been purified from pig lens to homogeneity by a rapid and efficient three-step procedure involving poly(ethylene glycol) fractionation, ion-exchange chromatography and chromatofocusing. The homogeneity of the purified enzyme was examined by polyacrylamide gel electrophoresis under native and denaturing conditions, by isoelectric focusing and by high-performance liquid chromatography on a size-exclusion column. The highly purified enzyme is a monomeric protein with a molecular mass of 35,775 +/- 3 Da as determined by electrospray mass spectrometry (ESMS). This purification procedure is particularly suited for the preparation of triclinic single crystals of pig lens aldose reductase, which are currently used in X-ray studies of this enzyme.

Crystallization and preliminary X-ray study of pig lens aldose reductase.

Crystals of pig lens aldose reductase have been grown from polyethylene glycol solutions at pH 6.2 and analysed by X-ray diffraction. Two crystal forms were obtained. The first belongs to space group P1 with unit cell dimensions a = 81.3 A, b = 85.9 A, c = 56.6 A, alpha = 102.3 degrees, beta = 103.3 degrees, gamma = 79.0 degrees, with four molecules in the unit cell related by a 222 non-crystallographic symmetry. The second crystal form is hexagonal. The space group is P6(2)22 with a = b = 101 A, c = 257 A and two molecules in the asymmetric unit. Both forms are suitable for X-ray structure analysis to better than 3 A resolution.

Modification of alpha-amylase from Bacillus licheniformis by the polyaldehyde derived from beta-cyclodextrine and alpha-amylase thermostability.

The cleavage of beta-cyclodextrine by sodium periodate at the seven 2-3 diols of the glucose unit gives rise to the polyaldehyde 1, used to modify alpha-amylase. The reductive modification of alpha-amylase from Bacillus licheniformis reduced the number of reactive lysine groups from 8 to 3.5 per mol of enzyme with an activity loss of 25% and increased the half-life at 80 degrees C from 4.7 to 7.0 minutes.

Enzyme and organic solvents: horse liver alcohol dehydrogenase in non-ionic microemulsion: stability and activity.

In a microemulsion made with Triton X-100, the stability of the enzymatic activity was higher than in ionic microemulsions. The stability increased with water content. The kinetic constants (Michaelis constant of NAD+ and maximum velocity) were close to those found in the previously studied microemulsions. The Michaelis constant of NAD+ expressed with respect to the buffer volume was higher than in water. The pH dependence of the kinetic constants in this microemulsion was determined. The activity determined by NAD+ reduction decreased with water content, whereas the redox activity determined via butanol oxidation coupled to retinal reduction was only slightly reduced.

Designed inhibitors of horse liver alcohol dehydrogenase. Effect of active-site metal ion substitution.

3-(p-Butoxyphenyl)propionamide, -thioamide and -hydrazide and the formamide of p-butoxybenzylamine were tested as inhibitors of cadmium(II) and cobalt(II) active-site substituted alcohol dehydrogenase. The results agree with a direct coordination of these inhibitors except for the hydrazide to the active-site metal ion, in the enzyme-NADH-inhibitor complex. The hydrazide might be situated at some distance from the metal ion without a direct coordination bond.

Squalene epoxide cyclase and microemulsion.

Squalene epoxide cyclase was extracted from microsomal preparations of rat liver using anionic, cationic and non-ionic microemulsions. The anionic microemulsion was the best with respect to protein solubilisation, extracted cyclase activity and stability of this activity over time. The activity assay showed cyclase activity to be higher in anionic microemulsion than in buffer in the presence of surfactant. Calcium chloride in the anionic microemulsion had a stabilising effect and less total protein seemed to be extracted.

Inactivation of yeast alcohol dehydrogenase by a reactive coenzyme analogue: 3-chloroacetyl pyridine adenine dinucleotide.

Yeast alcohol dehydrogenase is very rapidly and irreversibly inactivated by 3-chloroacetyl pyridine adenine dinucleotide, a reactive NAD+-analogue (Biellmann et al., 1974, FEBS Lett. 40, 29-32). Kinetic investigations with this compound, and structurally related compounds, show that this inactivation, against which NAD+ provides a complete protection, corresponds to an affinity label. The incorporation of the coenzyme analogue correlates linearly with the enzyme inactivation, the total inactivation corresponding to one mole of inactivator per coenzyme binding site. The pH-dependence of the inactivation rates of the enzyme by this coenzyme analogue and by its reduced form reflects exactly the pH variation of their respective dissociation constants. In spite of a good stability of the label in the non denatured inactivated enzyme, no modified amino-acid residue could be identified. Considering the affinity of this analogue for yeast alcohol dehydrogenase and the strict steric requirements of this enzyme towards its ligands, the nature of the inactivation reaction as well as different possibilities of the loss of the label in the inactivated enzyme are discussed.

Beta-hydroxysteroid dehydrogenase: activity in microemulsion and extraction from Pseudomonas testosteroni cells with microemulsion.

The stability of purified beta-hydroxysteroid dehydrogenase activity measured as a function of time was good in buffered cationic and non-ionic microemulsions. The use of 1-pentanol and 1-hexanol in place of 1-butanol as cosurfactant gave increased activity and stability. The NAD+ Michaelis constant was 0.22 mM in buffer and 3.5 mM in waterpool concentration in microemulsion. Proteins, among them beta-hydroxysteroid dehydrogenase, were extracted from Pseudomonas testosteroni with cationic microemulsion, thus indicating that microemulsions may be utilized in protein release from cells.

3-Chloroacetylpyridine adenine dinucleotide phosphate, an alkylating analogue of NADP+.

An alkylating analogue of NADP+ the 3-chloroacetylpyridine adenine dinucleotide phosphate was prepared from 3-diazoacetylpyridine adenine dinucleotide phosphate which was obtained by enzymatic transglucosidation of NADP+. The 3-diazoacetylpyridine adenine dinucleotide phosphate proved to be more unstable when compared to the corresponding NAD+ analogue. The alkylation of several dehydrogenases using this alkylating analogue is mentioned.

Binding of aldose reductase inhibitors: correlation of crystallographic and mass spectrometric studies.

Aldose reductase is a NADP(H)-dependent enzyme, believed to be strongly implicated in the development of degenerative complications of Diabetes Mellitus. The search for specific inhibitors of this enzyme has thus become a major pharmaceutic challenge. In this study, we applied both X-ray crystallography and mass spectrometry to characterize the interactions between aldose reductase and four representative inhibitors: AminoSNM, Imirestat, LCB3071, and IDD384. If crystallography remains obviously the only way to get an extensive description of the contacts between an inhibitor and the enzymatic site, the duration of the crystallographic analysis makes this technique incompatible with high throughput screenings of inhibitors. On the other hand, dissociation experiments monitored by mass spectrometry permitted us to evaluate rapidly the relative gas-phase stabilities of the aldose reductase-inhibitor noncovalent complexes. In our experiments, dissociation in the gas-phase was provoked by increasing the accelerating voltage of the ions (Vc) in the source-analyzer interface region: the Vc value needed to dissociate 50% of the noncovalent complex initially present (Vc50) was taken as a gas-phase stability parameter of the enzyme-inhibitor complex. Interestingly, the Vc50 were found to correlate with the energy of the electrostatic and H-bond interactions involved in the contact aldose reductase/inhibitor (Eel-H), computed from the crystallographic model. This finding may be specially interesting in a context of drug development. Actually, during a drug design optimization phase, the binding of the drug to the target enzyme is often optimized by modifying its interatomic electrostatic and H-bond contacts; because they usually depend on a single atom change on the drug, and are easier to introduce than the hydrophobic interactions. Therefore, the Vc50 may help to monitor the chemical modifications introduced in new inhibitors. X-ray crystallography is clearly needed to get the details of the contacts and to rationalize the design. Nevertheless, once the cycle of chemical modification is engaged, mass spectrometry can be used to select a priori the drug candidates which are worthy of further crystallographic investigation. We thus propose to use the two techniques in a complementary way, to improve the screening of large collections of inhibitors.

Evaluation of the antitumor activity of 1-(3-C-ethynyl-beta-D-ribofuranosyl) (PJ272), a recent ribonucleoside analogue.

The antiproliferative properties of a new ribonucleoside derivative, 1-(3'-C-ethynyl-beta-D-ribofuranosyl)uracil (PJ 272) that we synthesized a few years ago, were investigated in vitro on a variety of tumor cell lines from human and murine origins and in vivo, in tumor bearing mice. Using the 3(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay, we showed the ability of this compound to depress, at nanomolar concentrations, the growth of leukemia and lymphoma cultured cells. In 7 out of 8 tumor cell lines tested concentration of 50% inhibition (IC50) was found to be less than 25 nM. PJ 272 was also shown to present the same cytotoxicity against K562 Adriamycin-resistant cell line, which express a multi-drug resistance (MDR) phenotype, and its Adriamycin-sensitive parent cell line. Moreover, when injected intraperitoneally at 20 mg/kg every three days, PJ 272 was found to significantly increase the survival rate (T/C = 149%) of DBA/2 mice injected intraperitoneally with L1210 leukemic cells. Taken together, these results suggest that PJ 272 could be considered as a potentially very active drug against lymphoma and leukemia.

Di-tert-butyl diethylphosphoramidite as the phosphitylating reagent in the preparation of 3-deoxy-3-C-methylene-D-ribo-hexose-6-phosphate and 3-deoxy-3-C-methylene-D-erythro-pentose-5-phosphate.

3-Deoxy-3-C-methylene-D-ribo-hexose-6-phosphate and 3-deoxy-3-C-methylene-D-erythro-pentose-5-phosphate were prepared from a common intermediate 3-deoxy-3-C-methylene-1,2-O-isopropylidene-alpha-D-ribo-hexofuranose. The preparation of the phosphorylated unsaturated sugars employed di-tert-butyl diethylphosphoramidite as the phosphitylating reagent. The removal of all the protecting groups was done under acidic conditions in the ultimate step. The unsaturated sugar phosphates were competitive inhibitors but neither substrates nor inactivators of glucose-6-phosphate and ribose-5-phosphate isomerases.

Preparation of 1-(3-C-(propa-1,2-dienyl)-D-ribo-pentofuranosyl)uracil, an allenic nucleoside.

The Crabbé reaction was extended to the preparation of C-3'-allenyl-uridine. The effects of solvent and protecting group on the reaction were studied. The conversion in refluxing dioxan of disilyl either 3 proceeds to the corresponding allenic nucleoside 7; whereas, in refluxing THF the Mannich base 5 was obtained. Fully deprotected Mannich base and allenic uridines 6 and 9 were tested for their antitumor activity.