Global Analysis of Glycation
The total amount of glycation adducts of different chemically defined structures may be determined by exhaustive enzymatic digestion of protein extracts to hydrolyze protein to their constituent amino acids, followed by stable isotopic dilution analysis LC-MS/MS. This procedure has been automated and is often combined with a screen of protein oxidation and nitration adduct PTMs [2, 83]. Examples of application of this are detection increased FL, CML and MG-H1 in cytoplasmic proteins of the kidney glomeruli, retina and peripheral nerve of rats in experimental diabetes. FL: 0.62-0.96 vs 0.18-0.43 mmol/mol Lys. CML: 0.25-0.46 vs 0.17-0.27 mmol/mol Lys. MG-H1: 4.37-5.98 vs 1.70-2.99 mmol/mol Arg). This provides evidence for increased early glycation and AGE accumulation at tissue sites of development of microvascular complications of diabetes . Genetic and pharmacological prevention of AGE accumulation in Glol transgenic mice and high-dose thiamine-treated rats without change in FL with concomitant prevention of microvascular complications provided supporting evidence that AGEs rather than FL cause damage in diabetes [6, 95]. This was translated clinically to show increased FL and MG-H1 in plasma protein of patients with type 1 diabetes with respect to healthy controls (FL: 3.68 ± 0.86 vs. 1.35 ± 0.16 mmol/mol Lys; and MG-H, 0.99 ± 0.27 vs. 0.31 ± 0.20 mmol/mol Arg) . Skin collagen contents of FL and certain AGEs were found to be risk predictors of development of microvascular complications and cardiovascular disease in clinical diabetes [9, 10].
A great challenge for global screening of glycated proteins by limited proteolysis of cellular and extracellular proteomes and mass spectrometric analysis of related peptides is to maximize sequence coverage of proteins in mass spec- trometric analysis. Leading research teams performing total proteome analysis report a typical median sequence coverage of ca. 20% . A contributory factor to this is the production of short peptides of ambiguous protein origin. This may be improved in MG-modified proteins where MG-H1 formation typically produces missed cleavages with trypsin and Lys-C and longer peptides. A recent computational approach has indicated that with judicious use of proteases the sequence coverage in proteomics analysis may increase to ca. 90% . Until this is routinely implemented, it should be remembered that in glycation proteomics only a minor proportion of glycated proteins are likely to be identified.
For the fructosamine proteome, phenylboronate affinity chromatography was used to enrich glycated proteins and glycated tryptic peptides from both human plasma and erythrocyte membranes. Subsequent analysis by ETD- tandem mass spectrometry identified 76 and 31 proteins with fructosamine modification from human plasma and erythrocyte membranes, respectively. The ETD fragmentation mode permitted the identification of a significantly higher number of glycated peptides (88% of all identified peptides) versus CID mode (17% of all identified peptides) when utilizing enrichment on first the protein and then the peptide level .
In pilot studies using nanoflow liquid chromatography-Orbitrap Fusion™ mass spectrometry with peptide HCD fragmentation, we analyzed cytosolic protein extracts of human aortic endothelial cells in primary culture. In control samples, cell cytosolic protein had a total MG-H1 residue content of ca. 0.4 mmol/mol Arg measured by LC-MS/MS analysis of exhaustive enzymatic digests, while in tryptic digests, Orbitrap Fusion analysis detected 1027 proteins, of which 12 contained MG-H1 or MG-derived dihydroxyimidazolidine residues. After incubation of cytosolic protein extracts with exogenous MG to increase the MG-H1 content ca. 20-fold, we then detected total MG-H1 residue content of ca. 8 mmol/mol Arg and identified 1366 proteins, of which 344 now contained MG-H1 or MG-derived dihydroxyimidazolidine residues . In a recent report , plasma digests were analyzed by nanoflow chromatog- raphy-LTQ Orbitrap XL ETD mass spectrometry and tryptic peptides scanned for m/z 152.1 and 166.1 indicative of glyoxal- and MG-modified proteins. Forty-four peptides representing 42 proteins were annotated. Arginine modifications were mostly represented by glyoxal-derived hydroimidazolones (34 peptides/39 sites) and MG-derived dihydroxyimidazolidine (8 peptides/ 8 sites) and MG-H1 (14 peptides/14 sites). Use of high temperature and pH processing in this study may have compromised the outcome; many glyoxal- modified proteins were detected, whereas LC-MS/MS analysis typically shows very low amounts of glyoxal-derived-AGEs, hydroimidazolone, and Na- carboxymethylarginine in plasma protein .