Efficient synthesis of isofagomine and noeuromycin.

Starting from D-arabinose the synthesis of the very strong glycosidase inhibitors isofagomine (2) and noeuromycin (3) was achieved in six and seven steps, respectively. Keystep in the reaction sequence is the application of an efficient C-4 oxidation method to benzyl alpha-D-arabino-pyranoside. Subsequent Henry reaction of the obtained aldoketose with nitromethane provided the required branched carbohydrate precursors, which gave access to 2 and 3 in 17-21% overall yield.

Accurate determination of rate constants of very slow, tight-binding competitive inhibitors by numerical solution of differential equations, independently of precise knowledge of the enzyme concentration.

This paper is concerned with the determination of rate constants characterizing the binding and release of a slow binding inhibitor to and from an enzyme, here almond beta-glucosidase. We demonstrate the inability of the conventional method to yield reliable rate constants when one or more of these is less than 1 x 10(-4) per second. Instead one must use the much more accurate fitting of rate constants of the set of simultaneous differential equations characterizing the kinetic model. This procedure has the added advantage, when properly used, that the rate constants found pertaining to the inhibitor are largely insensitive to the particular value used for the enzyme concentration; i.e., the same data set may be fitted using a range of enzyme concentrations with no change in the resulting parameters. Hence the method can be used when little is known about the enzyme, except for the value of K(m), which is readily determined. Also, we report the somewhat unexpected finding that the association rate constant for the substrate (4-nitrophenyl-beta-d-glucopyranoside) is about one-third of the value of the corresponding rate constant for the inhibitor. The method is used to determine rate constants at several temperatures for the strong, slow binding inhibitor 2-phenethylglucoimidazole 1, enabling us to compute standard thermodynamic functions. The identity of these functions with those of isofagomine (2) reported earlier leads us to argue that the two compounds share a common binding mechanism, involving the same groups, whereas the different stabilities of the enzyme-inhibitor complexes must reside in those parts of the molecules that are not identical.

Synthesis of 5-azacastanospermine, a conformationally restricted azafagomine analogue.

The 5-aza-6-deoxy analogue of castanospermine (+/-)-5a and its 1-epimer (+/-)-5b was synthesized. The synthesis started from the known compound 5-benzyloxy-7-hydroxyhepta-1,3-diene, which was protected and subjected to Diels-Alder reaction with 4-phenyl-1,2,4-triazoline-3,5-dione to give two epimeric adducts. One of these was transformed through epoxidation, acetolysis, a series of side-chain transformations that converted it into a terminally protected aldehyde, deprotection, and hydrogenolysis/reductive amination into 5a. By a similar set of reactions the other adduct epimer was converted into 5b. The castanospermine analogue 5a was a weaker inhibitor of almond beta-glucosidase and rice alpha-glucosidase than castanospermine (2) or 1-azafagomine (4), but was considerably more potent than its epimer 5b. This suggests that these enzymes have a strong preference for binding substrates or azasugars with the 6-OH in an axial conformation.

Slow inhibition of almond beta-glucosidase by azasugars: determination of activation energies for slow binding.

The thermodynamic and activation energies of the slow inhibition of almond beta-glucosidase with a series of azasugars were determined. The inhibitors studied were isofagomine ((3R,4R,5R)-3,4-dihydroxy-5-hydroxymethylpiperidine, 1), isogalactofagomine ((3R,4S,5R)-3,4-dihydroxy-5-hydroxymethylpiperidine, 2), (-)-1-azafagomine ((3R,4R,5R)-4,5-dihydroxy-3-hydroxymethylhexahydropyridazine, 3), 3-amino-3-deoxy-1-azafagomine (4) and 1-deoxynojirimycin (5). It was found that the binding of 1 to the enzyme has an activation enthalpy of 56.1 kJ/mol and an activation entropy of 25.8 J/molK. The dissociation of the enzyme-1 complex had an activation enthalpy of -2.5 kJ/mol and an activation entropy of -297 J/molK. It is suggested that the activation enthalpy of association is due to the breaking of bonds to water, while the large negative activation entropy of dissociation is due at least in part to the resolvation of the enzyme with water molecules. For the association of 1 DeltaH(0) is 58.6 kJ/mol and DeltaS(0) is 323.8 J/molK. Inhibitor 3 has an activation enthalpy of 39.3 kJ/mol and an activation entropy of -17.9 J/molK for binding to the enzyme, and an activation enthalpy of 40.8 kJ/mol and an activation entropy of -141.0 J/molK for dissociation of the enzyme-inhibitor complex. For the association of 3 DeltaH(0) is -1.5 kJ/mol and DeltaS(0) is 123.1 J/molK. Inhibitor 5 is not a slow inhibitor, but its DeltaH(0) and DeltaS(0) of association are -30 kJ/mol and -13.1 J/molK. The large difference in DeltaS(0) of association of the different inhibitors suggests that the anomeric nitrogen atom of inhibitors 1-4 is involved in an interaction that results in a large entropy increase.

Specific glycosidase activity isolated from a random phage display antibody library.

Carbohydrates serve as key receptor sites in various cellular events such as viral attachment, tumor formation, and tissue inflammation. A potential route to control these events is to manipulate targeted carbohydrate structures in vivo using specifically designed glycohydrolases. Here we show that a stereospecific catalytic activity designed toward a particular sugar and linkage can be readily isolated from a phage display antibody library derived from a nonimmunized host. The activity was isolated using a transition-state analogue mimicking an alpha-glucosidasic linkage as antigen and showed a 20-fold specificity for that sugar and linkage. The DNA sequence, however, contains a large deletion in the antibody gene, which also changes the downstream reading frame, resulting in a translated sequence containing only 57 amino acids that has a predominantly hydrophobic amino terminal and a strongly hydrophilic carboxy terminal. The isolated catalytic activity has a strong pH dependence, attributable to one or more of the numerous potentially charged groups in the carboxyl terminal. While the protein readily forms more stable multimers, the 7.3-kD monomer represents by far the smallest glycosidase enzyme reported to date and can provide substantial new information toward understanding and modifying glycosidase activity.

Chemoenzymatic synthesis of isogalactofagomine.

A new chemoenzymatic synthesis of optically pure isogalactofagomine 2 starting from achiral starting materials is presented. Dimethyl 4-hydroxypyridine-3,5-dicarboxylate (7) was synthesized and converted to the corresponding saturated piperidine 8. Then the key step of the synthesis was carried out: Lipase M catalyzed hydrolysis of the prochiral diester 8 to cause formation of an asymmetric monoacid with at least 98% enantiomeric excess. Reduction of the acid, saponification of the remaining ester, and radical iododecarboxylation gave an iodide that after substitution with silver trifluoroacetate and hydrolysis gave 2.

Processing glucosidase inhibition by 1-azafagomine.

Natural azasugars have the ring oxygen substituted by nitrogen. They show potent inhibitory activity against glycosidases. The effect of substituting the ring carbon with nitrogen was examined with 1-azafagomine. 1-Azafagomine exhibited similar activity against processing glucosidase to that of fagomine.

Investigation of the slow inhibition of almond beta-glucosidase and yeast isomaltase by 1-azasugar inhibitors: evidence for the 'direct binding' model.

(-)-1-Azafagomine [(3R,4R,5R)-4,5-dihydroxy-3-hydroxymethylhexahydropyridazine; inhibitor 1] is a potent glycosidase inhibitor designed to mimic the transition state of a substrate undergoing glycoside cleavage. The inhibition of glycosidases by inhbitor 1 and analogues has been found to be a relatively slow process. This 'slow inhibition' process was investigated in the inhibition of almond beta-glucosidase and yeast isomaltase by inhibitor 1 and analogues. Progress-curve experiments established that the time-dependent inhibition of both enzymes by inhibitor 1 was a consequence of relatively slow dissociation and association of the inhibitor from and to the enzyme, and not a result of slow interchanges between protein conformations. A number of hydrazine-containing analogues of inhibitor 1 also inhibited beta-glucosidase and isomaltase slowly, while the amine isofagomine [(3R,4R,5R)-3,4-dihydroxy-5-hydroxymethylpiperidine; inhibitor 5] only inhibited beta-glucosidase slowly. Inhibitor 1 and related inhibitors were found to leave almond beta-glucosidase with almost identical rate constants, so that the difference in K(i) values depended almost entirely on changes in the binding rate constant, k(on). The same trend was observed for the inhibition of yeast isomaltase by inhibitor 1 and a related inhibitor. The values of the rate constants were obtained at 25 degrees C and at pH 6.8.

A new method for the deprotection of benzyl ethers or the selective protection of alcohols

A new selective method for the deprotection of benzyl ethers situated next to alcohols in the alpha, beta, or gamma position is presented which uses either NIS or DIB/I2 as a reagent. After initial formation of a hypoiodite intermediate, the reaction is believed to follow a radical pathway to resemble the Hoffman-Loffler-Freytag reaction. The formation of the intermediate hypoiodite is suggested on the basis of NMR studies. Depending on the substrate, the corresponding benzylidene derivatives or diols are isolated.

Carbohydrate Symposium--20th International Meeting. 27 August-1 September 2000, Hamburg, Germany.

The symposium was attended by experts of carbohydrate science and encompassed a wide spectrum of carbohydrate chemistry and biology. Data presented included novel methods of synthesizing carbohydrates, as well as the identification and elucidation of new carbohydrates. Clinical aspects of the symposium concentrated on carbohydrate vaccines, influenza treatment and diagnosis, and antidiabetics.

Direct NMR-spectroscopic determination of active-enzyme concentration by titration with a labeled inhibitor: determination of the k(cat) value of almond beta-glucosidase.

A new method for the determination of active-enzyme concentration of a glucosidase by using (13)C NMR spectroscopy is reported. The method consists of quantifying the binding between a (13)C-labelled, strong competitive inhibitor, [5-(13)C]-1-azafagomine (1), and the enzyme. The concentration of free inhibitor 1 is measured in a series of binding experiments from the intensity of its NMR signal relative to that of a reference. From a plot of the concentrations of bound vs. free inhibitor 1, the amount of specifically bound 1, that is, the amount of active sites, is determined. From this value, active-enzyme concentration and k(cat) value can be calculated.

Enantiospecific synthesis of 1-azafagomine.

For the first time the two enantiomeric forms of the glycosidase inhibitor 1-azafagomine have been synthesised starting from D- and L-xylose. D-Xylose was converted to the 2,3,5-tribenzylfuranose, which upon reductive amination with tert-butyl carbazate gave the protected 1-hydrazino-1-deoxypentitol in high yield. N-acetylation, mesylation of the 4-OH, removal of the Boc group, cyclisation and deprotection gave (+)-1-azafagomine ((+)-1). By a similar sequence of reactions, L-xylose was converted to (-)-1-azafagomine ((-)-1). Enzymatic and other routes to optically pure 1-azafagomine were also studied. Compound (-)-1 is a potent competitive glycosidase inhibitor, while (+)-1 has no biological activity. The inhibition of almond beta-glucosidase by (-)-1 was found to be slow owing to a slow binding step of inhibitor to enzyme, with no subsequent conformational rearrangement. The rate constants for binding and release were found to be 3.3 x 10(4)M(-1)s(-1) and 0.011 s(-1), respectively, yielding Ki = 0.33 microM.

Synthesis and deconvolution of the first combinatorial library of glycosidase inhibitors.

A combinatorial library of 125 compounds with a structure consisting of 1-azafagomine linked at N-1 via an acetic acid linker to a variable tripeptide was synthesised. The library was synthesised by Merrifield split and mix synthesis of the peptide, followed by capping with chloroacetate, regioselective nucleophilic substitution with 1-azafagomine and cleavage from the polymeric support. The library was screened for inhibition of beta-glucosidase, alpha-glucosidase and glycogen phosphorylase and found to display beta-glucosidase inhibition. Deconvolution of the library revealed that some inhibition was caused by all library members but the strongest inhibitor was clearly a compound having three hydroxyproline residues in the peptide fragment. This compound was a weaker but more selective inhibitor than 1-azafagomine itself.

1-Azaribofuranoside analogues as designed inhibitors of purine nucleoside phosphorylase. Synthesis and biological evaluation.

Pyrrolidine analogues of 2-deoxyribofuranose, having nitrogen in place of anomeric carbon, have been synthesised as potential transition state analogues of enzymatic nucleoside cleavage. Efficient synthetic methods were developed that allowed the synthesis of a wide range of 4-substituted 3-hydroxypyrrolidines starting from pyrroline and using opening of the pyrrolidine 3,4-epoxide, with carbon nucleophiles. Among the compounds synthesised were the 4-cyano- [(+/-)-16], 4-hydroxymethyl [(+/-)-22] and 4-carboxymethyl derivates [(+/-)-18]. From the hydroxymethyl derivative [(+/-)-22] N-alkylation with chloromethyluracil gave an inosine analogue [(+/-)-23]. The new compounds were tested for inhibition of human erythrocyte purine nucleoside phosphorylase. Compound (+/-)-22 was found to show non-competitive inhibition of the enzyme with a Ki of 160 microM. This suggested that (+/-)-22 binds to the ribofuranose portion of the active site. Furthermore, a solid-phase synthesis of 1'-azanucleoside analogues was developed.

A study of baker's yeast reduction of piperidone-carboxylates.

The stereoselective baker's yeast reduction of various N-protected piperidone-carboxylic acids have been studied, and the enantioselectivity was found to be widely dependent on whether fermenting or non-fermenting conditions were employed. Thus reaction of N-tert-butoxycarbonyl-4-oxopiperidine-3-carboxylic acid ethyl ester (6) with fermenting baker's yeast gave almost racemic N-tert-butoxycarbonyl-4-hydroxypiperidine-3-carboxylic acid ethyl ester (7), however, with complete diastereoselectivity. Reduction of 6 with non-fermenting yeast gave 7 with a 24-41% enantiomeric excess. Similarly, reduction of N-tert-butoxycarbonyl-3-oxopiperidine-4-carboxylic acid ethyl ester (17) with fermenting baker's yeast gave racemic N-tert-butoxycarbonyl-3-hydroxypiperidine-4-carboxylic acid ethyl ester [(+/-)-18] diastereoselectively. A convenient method for determining the enantiomeric excess of the hydroxypiperidine carboxylic acids derivatives was found in the reaction with Sanger's reagent followed by HPLC on a chiral column.

Synthesis of substituted chiral piperazones resembling aza-sugars.

(6R)-6-(Hydroxymethyl)piperazin-2-one (1), (6S)-6-(hydroxymethyl)piperazin-2-one (2) and (6S)-6-[(1R, 2S, 3R)-1,2,3,4-tetrahydroxybutyl]piperazin-2-one (3) have been prepared in optically pure forms starting from D-glucosamine hydrochloride (4). The compounds (1-3) were tested for glycosidase inhibition.

Synthesis of kojitriose using silicon-tethered glycosidation.

Reaction of 3,4,6-tri-O-acetyl-beta-D-glucopyranosyl chloride (1) with potassium phenylselenate gave phenyl 3,4,6-tri-O-acetyl-1-seleno-alpha, beta-D-glucopyranoside (2) in 59% yield. Silylation of benzyl 3,4,6-tri-O-benzyl-beta-D-glucopyranoside (4) with ethyl 3,4,6-tri-O-benzyl-2-O-chlorodimethylsilyl-1-thio-beta -D-glucopyranoside gave benzyl 2-O-(3,4, 6-tri-O-benzyl-1-S-ethyl-1-thio-beta-D-glucopyranos-2-O-yldimet hylsilyl)-3,4,6,-tri-O-benzyl-beta-D-glucopyranoside (5) in 35% yield. Reaction of 5 with N-iodosuccinimide in nitromethane gave benzyl 2-O-(3,4, 6-tri-O-benzyl-alpha-D-glucopyranosyl)-3,4, 6-tri-O-benzyl-beta-D-glucopyranoside (6) in 45% yield. Chlorodimethylsilylation of phenyl 3,4, 6-tri-O-acetyl-1-seleno-alpha-D-glucopyranoside (2 alpha) and reaction with 6 gave benzyl 2-O-[2-O-(3,4,6 -tri-O-acetyl-1-Se-phenyl-1-seleno-alpha-D-glucopyranos-2-O-yld imethylsilyl) -3,4,6-tri-O-benzyl-alpha-D-glucopyranosyl]-3,4,6-tri-O-benzyl-beta-D- glucopyranoside (7) in 82% yield. Intramolecular glycosidation of 7 using N-iodosuccinimide in nitromethane gave benzyl 2-O-[2-O-(3,4,6-tri-O-acetyl-alpha-D-glucopyranosyl)-3, 4,6-tri-O-benzyl-alpha-D-glucopyranosyl]-3, 4,6-tri-O-benzyl-beta-D-glucopyranoside (8) in 45% yield. Deprotection of 8 gave kojitriose (9) in quantitative yield. Chlorodimethylsilylation of 1,3,4,6-tetra-O-benzyl-alpha, beta-D-fructofuranose (10) with dimethyldichlorosilane and pyridine followed by reaction with ethyl 3,4, 6-tri-O-benzyl-1-thio-beta-D-glucopyranoside (3) gave ethyl 2-O-(1,3,4,6-tetra-O-benzyl-alpha, beta-D-fructofuranosyloxydimethylsilyl)-3,4, 6-tri-O-benzyl-1-thio-beta-D-glucopyranoside (11) in 85% yield. Chlorodimethylsilylation of 1,3,4, 6-tetra-O-benzoyl-alpha-D-fructofuranose (12) with dimethyldicholorosilane and triethylamine followed by reaction with phenyl 3,4, 6-tri-O-acetyl-1-thio-alpha-D-glucopyranoside (13) gave phenyl 2-O-(1,3,4, 6-tetra-O-benzoyl-alpha-D-fructofuranosyloxydimethylsilyl)-3, 4,6-tri-O-acetyl-1-thio-alpha-D-glucopyranoside (14) in 62% yield. Both 11 and 14 failed to undergo intramolecular glycosidation.

Evaluation of isofagomine and its derivatives as potent glycosidase inhibitors.

A pseudo-aza-monosaccharide and several pseudo-aza-disaccharide compounds were constructed based on replacement of the anomeric carbon with a nitrogen and the ring oxygen with a carbon. The inhibition constants of these compounds toward five different glycosidases, alpha-glucosidase, beta-glucosidase, isomaltase, alpha-mannosidase, and glucoamylase, were obtained. Isofagomine, the pseudo-aza-monosaccharide, shows a broad spectrum of strong inhibition against glycosidases. It is the most potent inhibitor of beta-glucosidase from sweet almonds reported to date and also a strong inhibitor of glucoamylase, isomaltase, and alpha-glucosidase. Isofagomine inhibits beta-glucosidase, glucoamylase, and isomaltase more strongly than 1-deoxynojirimycin where the ring oxygen has been replaced with a nitrogen. The alpha-1,6- linked pseudo-disaccharide showed very strong inhibition toward glucoamylase, being nearly as potent an inhibitor as acarbose. Pseudo-disaccharides in which the anomeric nitrogen was methylated to favor formation of either the alpha or beta substrate linkage generally had weakened inhibition for the glycosidases studied most likely due to steric interference with the various active sites. These results indicate that the presence of a basic group at the anomeric center is important for carbohydrase inhibition. The presence of a charged carboxylate group near the anomeric carbon which interacts with the basic nitrogen is suggested for these enzymes, particularly for beta-glucosidase. The presence of a second alpha-linked glucosyl residue is also critical for strong inhibition of glucoamylase.