Specific inhibition of an alpha-galactosyltransferase from Trypanosoma brucei by synthetic substrate analogues.

Since the alpha-D-galactose-(1-->3)-D-galactose epitope has been identified to be the major target in the process of hyperacute rejection of xenografts transplanted from nonprimate donors to humans, specific inhibitors of alpha-galactosyltransferases are of broad interest. Using Trypanosoma brucei, a protozoan parasite causing sleeping sickness and Nagana, we have a very useful model system for the investigation of alpha-galactosyltransferase inhibitors, since the variant surface glycoprotein (VSG) accounts for about 10% of the total cell protein an this parasite expresses many different galactosyltransferases including the one catalysing the formation of the Galalpha1-->3Gal epitope. In order to study inhibition of galactosylation on the VSG from Trypanosoma brucei, we designed, synthesized and tested substrate analogues of trypanosomal alpha-galactosyltransferases. Effective inhibitors were a pair of diastereoisomeric UDP-galactose analogs, in which the galactose residue is linked to UDP via a methylene bridge rather than an ester linkage. Hence, galactose cannot be transferred to the respective acceptor substrate VSG or the synthetic acceptor substrate Manalpha1-->6Manalpha1S-(CH2)7-CH3, which was previously proven to replace VSG effectively [Smith et al. (1996) J Biol Chem 271:6476-82]. Inhibitors have been prepared starting from 1-formyl galactal. The final condensation was performed using UMP morpholidate leading to a pair of diastereomeric compounds in 39% or 30% yield, respectively. These compounds were tested using alpha-galactosyltransferases prepared from T. brucei membranes and lactose synthetase from bovine milk. While the K(M)-value for UDP-galactose was determined as 59 microM on bovine lactose synthetase, the K(I)-values for both inhibitors were 0.3 mM and 1.1 mM respectively, showing that these inhibitors are unable to inhibit enzyme activity significantly. However, using the N-glycan specific alpha-galactosyltransferase from trypanosomes, the K(M)-value was determined as 20 microM, while the K(I)-values were 34 microM and 21 microM respectively. Interestingly, other trypanosomal alpha-galactosyltransferases, which modify the GPI membrane anchor, are 2 orders of magnitude less effected by the inhibitor.

The sulfhydryl group microenvironment of lactose synthase from bovine milk.

Galactosyltransferase from bovine milk was inactivated by a series of sulfhydryl group specific reagents of different structures and sizes. The inactivation rate constants suggest that the thiol is located in a nonpolar microenvironment. The ESR spectrum of a spin labeled galactosyltransferase showed that the sulfhydryl group is in a region of non-restricted rotation, consistent with its broad reactivity towards various thiol reagents. Galactosyltransferase immobilized onto agarose through its sulfhydryl group retained its ability to catalyze the synthesis of N-acetyllactosamine and lactose. Thus the residual activity of the sulfhydryl group modified enzyme is not due to an isozyme lacking such a group. In addition, the active thiol can not be located at the active site nor the protein-protein interaction site between galactosyltransferase and alpha-lactalbumin.

Photoaffinity labeling of lactose synthase with a UDP-galactose analogue.

A photoaffinity analogue of UDP-galactose, 4-azido-2-nitrophenyluridylyl pyrophosphate (ANUP), has been synthesized for the investigation of the binding topography of alpha-lactalbumin on galactosyltransferase. Results obtained from steady state kinetics show that ANUP is an effective competitive inhibitor against UDP-galactose in the reactions of lactose and N-acetyllactosamine syntheses. The specific binding of ANUP to the UDP-galactose-binding site is further demonstrated by its ability to facilitate the formation of the lactose synthase complex on solid supports, either alone or in the presence of glucose or N-acetyl-glucosamine. ANUP inactivates galactosyltransferase on irradiation. One mole of ANUP was incorporated per mol of enzyme inactivated. This process is Mn2+-dependent and can be prevented by UDP-galactose. Glucose and N-acetylglucosamine render only partial protection. Photoaffinity labeling of lactose synthase either free in solution or immobilized on Sepharose does not result in any reduction of the alpha-lactalbumin modifier activity. In addition, no incorporation of radioactivity into alpha-lactalbumin was observed when radioactive ANUP was used, whereas galactosyltransferase was labeled. These data indicate that alpha-lactalbumin does not bind to galactosyltransferase in the region of the ANUP site, suggesting that the location of protein-protein interaction between the two subunits of lactose synthase may be removed from the UDP-galactose-binding domain.

Synthesis of 4-deoxy-D-xylo-hexose and 4-azido-4-deoxy-D-glucose and their effects on lactose synthase.

Syntheses are reported of 4-deoxy-D-xylo-hexose and 4-azido-4-deoxy-D-glucose as potential inhibitors for lactose synthase [uridine 5'-(alpha-D-galactopyranosyl pyrophosphate):D-glucose 4-beta-D-galactopyranosyltransferase, EC 2.4.1.22]. These syntheses involved SN2 displacement of the 4-methylsulfonyloxy group of methyl 2,3,6-tri-O-benzoyl-4-O-methylsulfonyl-alpha-D-galactopyranoside by iodide and azide ions. In both cases, inversion in configuration was observed. The resulting intermediates, methyl 2,3,6-tri-O-benzoyl-4-deoxy-4-iodo-alpha-D-glucopyranoside and methyl 4-azido-2,3,6-tri-O-benzoyl-4-deoxy-alpha-D-glucopyranoside, were obtained in crystalline form. Both 4-deoxy-D-xylo-hexose and 4-azido-4-deoxy-D-glucose were found to be inhibitors for lactose synthase in the presence of alpha-lactalbumin, but had no effect in the absence of alpha-lactalbumin. Both D-glucose analogues bind to the enzyme system far more weakly than D-glucose, suggesting that the recognition of the 4-OH group of the acceptor substrate is an important factor in binding.

Chemical modification of tyrosyl and lysyl residues in goat alpha lactablbumin and the effect on the interaction with the galactosyl transferase.

The effect of acylation of goat alpha-lactalbumin on lactose synthetase activity and the ability of alpha-lactalbumin to inhibit the transfer of galactose to N-acetylglucosamine is biphasic. Approx. 15% of the lactose synthase activity of goat alpha-lactalbumin and 10% of its inhibitory power is lost in the initial phase, with corresponding losses of 65 and 30% in the second phase. Deacylation of reacted tyrosyl groups with hydroxylamine restored inhibitory power completely in the initial phase and partially in the second phase. Removal of acyl groups in the initial phase decreased lactose synthase activity, but had no effect in the second phase. The differential effect of acylation of alpha-lactalbumin on lactose synthase and inhibitory properties appears to be the result of differential changes in the affinity of the UDP-Gal-galactosyl-transferase-alpha-lactalbumin ternary complex for monosaccharides.

The binding of glycoconjugates to human-milk D-galactosyltransferase.

Through the use of affinity chromatography, a homogeneous preparation of human beta(1 leads to 4)-D-galactosyltransferase (the A protein of lactose synthase) was obtained. The specificity of this protein for glycoconjugates was studied in the presence and absence of human alpha-lactalbumin. A kinetic analysis of the transfer of D-galactose to N-acetyl-D-glucosamine and the beta(1 leads to 4) linked N-acetylglucosamine oligomers, suggested that the active site region of the enzyme contains more than one binding site for acceptor moleucles. Furthermore, experiments with Na-acetylglucosamine-beta(1 leads to4)-N-acetylmuramic-pentapeptide isolated from Micrococcus luteus indicated that the presence of a peptide chain does not enhance enzymic activity, as compared with the corresponding free disaccharide. Similar results were obtained using ovalbumin and the ovalbumin glycopeptide (which have similar apparent Km values for A protein) as galactose acceptors. In contrast to its ability to inhibit N-acetyllactosamine production, alpha-lactalbumin did not inhibit the transfer of D-galactose to the N-acetylglucosamine oligomers or the glycopeptides. Although alpha-lactalbumin can switch the specificity of A protein from N-acetyl-D-glucosamine to D-glucose resulting in the production of lactose, no transfer of galactose was observed to beta(1 leads to 4)-linked glycose oligomers or to a collagen glycopeptide, D-glycopyranosyl-alpha(1 leads to 2)-D-galactopyranosyloxy-beta(1 leads to 5)-lysine. IT therefore appears that alpha-lactalbumin can only modify human A protein for monosaccharide acceptors.

Evidence for specific transport of uridine diphosphate galactose across the Golgi membrane of rat mammary gland.

The inhibition of lactose synthesis by UDP-glucose, UDP-glucuronate and, less so, by UDP-N-acetylglucosamine was markedly smaller in preparations of "intact" than of lysed vesicles derived from the Golgi apparatus of lactating rat mammary gland. This constitutes evidence for a specific, probably facilitated, transport of UDP-galactose across the Golgi membrane.

Kinetic studies on the effect of uridine diphosphate galactose and manganous ions on the reaction between lactose synthetase A protein from human milk and p-hydroxymercuribenzoate.

The inhibition of lactose synthetase A protein by p-hydroxymercuribenzoate at pH7.5 and 25 degrees C, which involves the reaction of one molecule of inhibitor with each molecule of enzyme, was decreased in rate by UDP-galactose, especially in the presence of Mn(2+). Pseudo-first-order rate constants for the reaction between 0.1mm-p-hydroxymercuribenzoate and free enzyme, the enzyme-UDP-galactose complex and the enzyme-Mn(2+)-UDP-galactose complex were 4.4x10(-2), 1.9x10(-2) and 0.3x10(-2)min(-1) respectively. The results also indicated that dissociation constants for UDP-galactose in the enzyme-UDP-galactose and enzyme-Mn(2+)-UDP-galactose complexes were 313 and 16mum respectively, the latter value being similar to the K(m) for UDP-galactose in the lactose synthetase reaction. The protective effect of UDP-galactose and the role of Mn(2+) ions in lactose synthetase are discussed.

Studies on the thiol group of lactose synthetase A protein from human milk and on the binding of uridine diphosphate galactose to the enzyme.

The lactose synthetase activity of A protein from human milk was much decreased but not abolished by reaction with thiol-group reagents. Protection experiments indicated that a free thiol group on the enzyme is situated near the UDP-galactose binding site and inactivation of the enzyme with p-hydroxymercuribenzoate was probably due to prevention of UDP-galactose binding. Affinity chromatography showed that the mercuribenzoate substituent also decreased the affinity of A protein for N-acetylglucosamine but complex-formation between A protein-N-acetylglucosamine and alpha-lactalbumin was relatively unaffected. UDP-galactose appears to be bound to the enzyme mainly through its pyrophosphate group with Mn(2+) ion and through the cis hydroxyls of ribose, whereas its hexose moiety has little if any affinity for the enzyme. Lactose synthetase activity remaining after the reaction with thiol-group reagents indicates that a free thiol group is not an essential part of the A protein active site.