Formation of AGEs (advanced glycation end-products) and ALEs (advanced lipoxidation end-products) on proteins is associated with aging and various diseases of oxidative stress, notably diabetes and its complications. Modification of protein to AGE/ALEs is known to be site-directed and this has potential implications for protein functionality and design of AGE/ALE inhibitors. Determination of the site-specificity of modification is achieved most efficiently by MS. The present paper summarizes some of the challenges that need to be addressed when determining the site-specificity of AGE/ALE formation on protein by MS, using the protein RNase as an example. The following topics are discussed: formation and significance of AGE/ALEs, location of glycated peptides, enzymic digestion of glycated peptides and selection of mass spectrometric settings of analysis for glycated peptides.
AGE (advanced glycation end-product) formation
Components possessing a free carbonyl group react with the side chains of reactive amino acid residues (especially lysine and arginine) within proteins to give compounds known as AGEs and ALEs (advanced lipoxidation end-products) . The carbonyl compounds include reducing sugars, e.g. lactose or glucose, leading to AGEs, and lipid oxidation products, e.g. HNE (4-hydroxy-2-nonenal) or MDA (malondialdehyde), giving rise to ALEs, as well as oxidation products of both sugars and lipids, e.g. GO (glyoxal), MGO (methylglyoxal) and glycolaldehyde [1–3]. Reactions involving the ϵ-amino group of lysine residues and an aldose initially lead to an early glycation product known as the ARP (Amadori rearrangement product). When glucose is the aldose, the ARP is known as FL (fructoselysine). The ARP subsequently oxidizes to form AGEs, including CML [Nϵ-(carboxymethyl)lysine] . Dicarbonyls such as GO and MGO attack predominantly the guanidino group of arginine residues to yield AGE/ALEs known as DHs (dihydroxyimidazolidines) and HIs (hydroimidazolones) [2,3]. MGO can also modify arginine residues to THP (tetrahydropyrimidines) and argpyrimidine. The structures of some AGE/ALEs are shown in Figure 1.
Structures of selected AGE/ALEs
Significance of AGE/ALEs in aging and disease
AGE/ALEs are irreversible chemical modifications and cross-links in proteins, and CML and MG-HI (MGO-derived HI) are, quantitatively, some of the major AGE/ALEs detected in tissue proteins [3,5]. Increased chemical modification of proteins by oxidation, glycoxidation and lipoxidation reactions is implicated in the pathogenesis of many chronic age-related diseases, including arthritis, atherosclerosis, diabetes and Alzheimer's disease [6,7]. AGE/ALEs increase naturally in proteins with age, but this process is accelerated during hyperglycaemia in diabetes . The increase in AGE/ALE formation is implicated in the development of diabetic complications, including vascular, renal and retinal disease.
The structural features of any particular protein are likely to determine the distribution of AGE/ALEs among the potential reactive sites, and localization of a high proportion of AGE/ALEs at a few reactive sites might accentuate the effect of AGE/ALE formation on protein functionality. The relationship between sites of glycation and AGE/ALE formation has been studied only recently, and there is little understanding of the structural features in proteins that catalyse AGE/ALE formation . Information concerning the distribution of protein AGE/ALEs, as well as protein modifications more generally, is important because modification at one or two key amino acid residues would exacerbate effects of AGE/ALEs on protein functionality in aging and diabetes. Knowledge of the specificity of AGE/ALE formation may also have implications for the design of AGE/ALE inhibitors, e.g. the merits of chelating activity or anionic compared with cationic inhibitors.
Our recent studies using RNase A as a model protein indicate, for example, that Lys-41 is the primary site of early glycation and carboxymethylation , whereas Arg-39 and Arg-85 are the main sites of reaction with GO and MGO [10,11]. We have also shown Met-29 to be the principal site of methionine sulfoxide formation . Peptide mapping using MS is the most efficient means of determining the nature of AGE/ALEs and their distribution on protein. [7,9,10,13,14]
MS for probing AGE/ALE-modified protein
Location of glycated peptides
Most of our experiments concerning the site-specificity of modification of protein to AGE/ALEs and pre-AGEs has involved the used of LC (liquid chromatography) coupled to a triple-quadrupole mass spectrometer. Although the mass accuracy of such equipment is relatively low, compared with a TOF (time-of-flight) MS, since we were working with enzymic digests of a single well-characterized protein (RNase), the masses of all of the predicted peptides (unmodified and modified) could be readily calculated. Our first work in this area concerned the location of peptides modified to specific glycation adducts, i.e. FL and CML . Location of FL– and CML–peptides, as well as unmodified peptides, was relatively straightforward and involved extracting ion chromatograms at m/z values that corresponded to the different charge states of the peptides of interest (Figure 2). The detection of two or more charge states of the FL– or CML–peptides in enzymic digests of the modified protein that were absent from the unmodified protein confirmed the presence of the modified peptide of interest .
Extracted ion chromatograms of peptide Cys-40–Lys-61 containing Lys-41 modified to FL
In later work, we were interested in using the triple-quadrupole MS data to obtain a more global view of the AGE/ALE-modified peptides, regardless of the nature of the specific AGE/ALE modifications. A rapid means of achieving this was to reconstruct a single mass spectrum for the entire chromatogram for a sample . An example for MGO-modified RNase is shown in Figure 3. By comparing such regenerated mass spectra for unmodified and modified protein digests over 100 a.m.u. (atomic mass unit) increments, we were able to locate ions that were present in only one or the other digest . By scrutinizing the masses of these ions, it frequently became clear that some were likely to be due to different charge states of the same peptide. Our approach was then to sequence these peptides to confirm the amino acid sequence and deduce the nature of the modification .
Single combined mass spectra for tryptic digests of RNase incubated with MGO (A) and native RNase (B)
When protein reacts with carbonyl compounds, the side chains of the lysine and arginine residues are most susceptible to modification. Trypsin is the enzyme most commonly used to digest protein before analysis by MS. Trypsin cleaves lysine and arginine residues C-terminal within the peptide chain, but not if the side chain of the lysine or arginine residue is modified. This lack of cleavage of modified protein by trypsin leads to some challenges. One of these is the very large modified peptides that can result in some cases. For example, when RNase reacts with MGO, Arg-85 is a favoured site of modification . However, since Lys-91 is resistant to digestion by trypsin , the resulting modified peptide contains the amino acid sequence from Asn-67 to Lys-98, i.e. 32 amino acid residues. In our hands, attempts to analyse the most abundant (3+) charge state of this peptide by MS/MS on the QTOF (quantitative TOF) provided little sequencing information . We therefore digested the protein using endoprotease Glu-C and trypsin in sequence. Glu-C cleaves C-terminally to glutamic acid residues (and, in the presence of phosphate, aspartic acid residues). Digestion with Glu-C for 24 h in the absence of phosphate followed by trypsin for 5 h resulted in a peptide containing Arg-85 that contained the amino acid sequence from Asn-67 to Glu-86 . The 3+ charge state of this peptide was most amenable to sequencing using QTOF, and highly informative spectra resulted . The location of Arg-85 adjacent to the C-terminal amino acid residue resulted in a prominent series of y-ions for each peptide that was most helpful in confirming the nature of the modification .
Choice of appropriate MS settings
The sequencing spectrum we obtained for one of our MGO-modified Arg-85-containing Glu-C/tryptic peptides (named UK85) was initially difficult to interpret . The mass of the peptide (3+ parent ion mass=886.0) was 90 a.m.u. higher than that of the mass of peptide Asn-67–Glu-86 containing Arg-85 modified to MG-HI . However, the masses of the y-ions for UK85 were only 62 a.m.u. higher than the corresponding y-ions for peptide Asn-67–Glu-86 containing Arg-85 modified to MG-HI. Also, there were additional series of ions in the spectrum obtained for UK85. Careful study of the spectrum for UK85 suggested that the difference in mass between the parent ions of UK85 and peptide Asn-67–Glu-86 containing Arg-85 modified to MG-HI (90 a.m.u.) and their corresponding y-ions (28 a.m.u.), i.e. 62 a.m.u., might be accounted for by the loss of water and carbon dioxide in the collision cell of the MS. By performing a series of experiments involving varying the collision energy, we proved this to be the case. In addition, by varying the cone voltage, we showed that loss of water can also occur (although not so readily) in the ionization source . Furthermore, we performed a series of experiments to demonstrate that dehydration of MG-DH (MGO-derived DH) to MG-HI occurs in the source and increases with cone voltage over the range 40–80 eV. We also established that MG-HI itself can degrade in the source using MS conditions commonly used to sequence peptides . These findings are important in two respects. First, they draw attention to the need to carefully select the correct MS conditions for analysis when the aim is to quantify compounds, especially those that may readily undergo dehydration or decarboxylation. Secondly, it is essential to be aware of the possibility of such neutral losses when attempting to sequence peptides and identify modifications within proteins.
Conclusions and future work
Carbonyl compound modification of protein is site-directed. Certain AGE/ALEs, such as those formed by MGO attack on arginine residues, are not completely stable when peptides are sequenced by MS. Unless care is taken, erroneous conclusions may be drawn regarding the nature of the modifications and the amounts formed. In the future, MS methods developed to determine site-specificity of modification of protein should be refined and applied to ex vivo samples from patient studies to facilitate understanding of the relationships between protein glycation and aging/disease.
Bioanalysis in Oxidative Stress: A Biochemical Society Focused Meeting held at the University of Exeter, U.K., 2–3 April 2008. Organized and Edited by John Moody (Plymouth, U.K.) and Paul Winyard (Peninsula Medical School, Exeter, U.K.).
advanced glycation end-product
advanced lipoxidation end-product
atomic mass unit
Amadori rearrangement product