Glutathionylation of lens proteins through the formation of thioether bond.

Formation of lanthionine, a dehydroalanine crosslink, is associated with aging of the human lens and cataractogenesis. In this study we investigated whether modification of lens proteins by glutathione could proceed through an alternative pathway: that is, by the formation of a nonreducible thioether bond between protein and glutathione. Direct ELISA of the reduced water-soluble and water-insoluble lens proteins from human cataractous, aged and bovine lenses showed a concentration-dependent immunoreactivity toward human nonreducible glutathionyl-lens proteins only. The reduced water-insoluble cataractous lens proteins showed the highest immunoreactivity, while bovine lens protein exhibited no reaction. These data were confirmed by dot-blot analysis. The level of this modification ranged from 0.7 to 1.6 nmol/mg protein in water-insoluble proteins from aged and cataractous lenses. N-terminal amino acid determination in the reduced and alkylated lens proteins, performed by derivatization of these preparations with dansyl chloride followed by an exhaustive dialysis, acid hydrolysis and fluorescence detection of dansylated amino acids by RP-HPLC, showed that N-terminal glutamic acid was present in concentration of approximately 0.2 nmol/mg of lens protein. This evidence points out that at least some of the N-terminal amino groups of nonreducible glutathione in the reduced human lens proteins are not involved in a covalent bond formation. Since disulfides were not detected in the reduced and alkylated human lens proteins, GSH is most likely attached to lens proteins through thioether bonds. These results provide, for the first time, evidence that glutathiolation of human lens proteins can occur through the formation of nonreducible thioether bonds.

Glucose-derived Amadori compounds of glutathione.

Under the chromatographic conditions used in these studies we observed time- and concentration-dependent formation of N-1-Deoxy-fructos-1-yl glutathione as the major glycation product formed in the mixtures of GSH with glucose. N-1-Deoxy-fructos-1-yl glutathione had a characteristic positively charged ion with m/z=470 Th in its LC-MS spectra. Mixtures of glutathione disulfide and glucose generated two compounds: N-1-Deoxy-fructos-1-yl GSSG (m/z=775 Th) as major adduct and bis di-N, N'-1-Deoxy-fructos-1-yl GSSG (m/z=937 Th) as the minor one. All three compounds showed a resonance signal at 55.2 ppm in the 13C-NMR spectra as C1 methylene group of deoxyfructosyl, which represents direct evidence that they are Amadori compounds. All three compounds purified from GSSG/Glc or GSH/Glc mixtures also showed LC-MS/MS fragmentation patterns identical to those of the synthetically synthesized N-1-Deoxy-fructos-1-yl glutathione, N-1-Deoxy-fructos-1-yl GSSG and bis di-N, N'-1-Deoxy-fructos-1-yl GSSG. N-1-Deoxy-fructos-1-yl glutathione was shown to be a poor substrate for glutathione peroxidase (6.7% of the enzyme's original specific activity) and glutathione-S-transferase (25.7% of the original enzyme's specific activity). Glutathione reductase failed to recycle the disulfide bond within the structure of di-substituted bis di-N, N'-1-Deoxy-fructos-1-yl GSSG. It showed only 1% of the original enzyme's specific activity, but retained its ability to reduce the disulfide bond within the structure of N-1-Deoxy-fructos-1-yl GSSG by 57% of its original specific activity. Since the GSH concentration in diabetic lens is significantly decreased and the glucose concentration can increase 10-fold and higher, the formation of Amadori products of the different forms of glutathione with this monosaccharide may be favored under these conditions and could contribute to a lowering of glutathione levels and an increase of oxidative stress observed in diabetic lens.

Dehydroalanine crosslinks in human lens.

This study was conducted to develop a methodology for the purification and detection of histidinoalanine, lanthionine and lysinoalanine in the lens tissue. Cataractous and aged human lens proteins were hydrolysed and fractionated by using anion-exchange chromatography. The fraction containing the bulk of dehydroalanine crosslinks was derivatized with dansyl chloride and then separated and quantified by means of RP-HPLC. The spectral and chromatographic properties of all three substances purified and quantified in this study were identical to those of their synthesized counterparts. Histidinoalanine and lanthionine were the most abundant dehydroalanine crosslinks in both water-soluble and water-insoluble lens proteins. Histidinoalanine levels in water-soluble proteins from the cataractous lenses of Indian origin were 6.2-fold higher than those in water-soluble proteins from normal lenses (1.68+/-0.75 vs 0.26+/-0.06 nmol/mg protein; p<0.001). In water-insoluble proteins, they were 2.2-fold higher in cataractous lenses compared with normal lenses (1.59+/-0.76 vs 0.73+/-0.17 nmol/mg protein; p<0.01). Lanthionine levels were significantly higher in water-insoluble proteins of cataractous lenses when compared to non-cataractous lenses (2.5+/-1.68 vs 0.95+/-0.08 nmol/mg protein; p<0.03). Unlike histidinoalanine, this crosslink appears to accumulate in relatively high concentrations in water-soluble lens proteins; its concentration was 9-fold higher than histidinoalanine from the same proteins (0.26+/-0.06 HAL vs 2.34+/-0.76 LAN nmol/mg protein; p<0.0004). The concentration of lysinoalanine was in the picomolar range and in cataractous lens proteins only.