Identification of N epsilon-carboxymethyllysine as a degradation product of fructoselysine in glycated protein

The chemistry of Maillard or browning reactions of glycated proteins was studied using the model compound, N alpha-formyl-N epsilon-fructoselysine (fFL), an analog of glycated lysine residues in protein. Incubation of fFL (15 mM) at physiological pH and temperature in 0.2 M phosphate buffer resulted in formation of N epsilon-carboxymethyllysine (CML) in about 40% yield after 15 days. CML was formed by oxidative cleavage of fFL between C-2 and C-3 of the carbohydrate chain and erythronic acid (EA) was identified as the split product formed in the reaction. Neither CML nor EA was formed from fFL under a nitrogen atmosphere. The rate of formation of CML was dependent on phosphate concentration in the incubation mixture and the reaction was shown to occur by a free radical mechanism. CML was also identified by amino acid analysis in hydrolysates of both poly-L-lysine and bovine pancreatic ribonuclease glycated in phosphate buffer under air. CML was also detected in human lens proteins and tissue collagens by HPLC and the identification was confirmed by gas chromatography/mass spectroscopy. The presence of both CML and EA in human urine suggests that they are formed by degradation of glycated proteins in vivo. The browning of fFL incubation mixtures proceeded to a greater extent under a nitrogen versus an air atmosphere, suggesting that oxidative degradation of Amadori adducts to form CML may limit the browning reactions of glycated proteins. Since the reaction products, CML and EA, are relatively inert, both chemically and metabolically, oxidative cleavage of Amadori adducts may have a role in limiting the consequences of protein glycation in the body.

Loci of catabolism of beta-very low density lipoprotein in vivo delineated with a residualizing label, 125I-dilactitol tyramine

beta-Very low density lipoprotein (beta-VLDL) may be a major atherogenic lipoprotein, and knowledge of the sites of its catabolism should facilitate elucidation of mechanisms important in the regulation of its plasma concentrations. In this study, catabolic sites of beta-VLDL have been delineated in normolipidemic rabbits with a novel, radioiodinated, residualizing label, 125I-dilactitol tyramine (125I-DLT). Comparative studies of beta-VLDL and low density lipoprotein catabolism were performed with 125I-DLT conjugated to each lipoprotein and with lipoproteins iodine-labeled conventionally. Conjugation did not alter size distributions or charge characteristics of lipoprotein particles. The overall processing (binding and degradation) of lipoproteins by cultured rabbit skin fibroblasts was not influenced by 125I-DLT derivatization, suggesting that attachment of the label did not influence cell receptor-lipoprotein interactions. Furthermore, although degradation products of 125I-lipoproteins leaked out of the cells and into the medium, the degradation products of 125I-DLT lipoproteins were retained by the cells. The principal catabolic site of beta-VLDL in normolipidemic rabbits was found to be the liver with 54 +/- 4% of injected 125I retained in this organ 24 h after injection of 125I-DLT-beta-VLDL. When catabolism was normalized to tissue weight, the liver and adrenals were found to be approximately equally active in the metabolism of beta-VLDL. In agreement with results of other studies with residualizing labels, the principal organ of catabolism of 125I-DLT-LDL in vivo was the liver. The adrenals were the most highly catabolizing organ when results were normalized for tissue weight. The quantitative differences observed in the tissue distributions of injected 125I-DLT-beta-VLDL and 125I-DLT-low density lipoprotein suggested that a significant proportion of beta-VLDL is removed by tissues before conversion to low density lipoprotein.

Glycation of amino groups in protein. Studies on the specificity of modification of RNase by glucose

Ribonuclease A has been used as a model protein for studying the specificity of glycation of amino groups in protein under physiological conditions (phosphate buffer, pH 7.4, 37 degrees C). Incubation of RNase with glucose led to an enhanced rate of inactivation of the enzyme relative to the rate of modification of lysine residues, suggesting preferential modification of active site lysine residues. Sites of glycation of RNase were identified by amino acid analysis of tryptic peptides isolated by reverse-phase high pressure liquid chromatography and phenylboronate affinity chromatography. Schiff base adducts were trapped with Na-BH3CN and the alpha-amino group of Lys-1 was identified as the primary site (80-90%) of initial Schiff base formation on RNase. In contrast, Lys-41 and Lys-7 in the active site accounted for about 38 and 29%, respectively, of ketoamine adducts formed via the Amadori rearrangement. Other sites reactive in ketoamine formation included N alpha-Lys-1 (15%), N epsilon-Lys-1 (9%), and Lys-37 (9%) which are adjacent to acidic amino acids. The remaining six lysine residues in RNase, which are located on the surface of the protein, were relatively inactive in forming either the Schiff base or Amadori adduct. Both the equilibrium Schiff base concentration and the rate of the Amadori rearrangement at each site were found to be important in determining the specificity of glycation of RNase.

125I-glycoconjugate labels for identifying sites of protein catabolism in vivo: effect of structure and chemistry of coupling to protein on label entrapment in cells after protein degradation

Residualizing radioactive labels are designed to remain entrapped within cells following degradation of a carrier protein, and have been used for identification of the tissue and cellular sites of plasma protein catabolism. In this study we describe a convenient synthesis and purification of a series of 125I-labeled glycoconjugates, and an evaluation of their efficiency of retention in liver following degradation of a model carrier protein, asialofetuin. Glycoconjugates were prepared in 65-90% yield by reductive amination of reducing sugars with aromatic amines using NaBH3CN. The products were purified in a single ion-exchange chromatographic step, and then labeled with 125I. The derivatives prepared were mono-and disubstituted lactitol-,cellobiitol-and glucitol-[125I]tyramine and lactitol-[125I]tyrosine. 125I-Glycoconjugates were coupled to asialofetuin using either cyanuric chloride or, for lactose-containing labels, by treatment with galactose oxidase followed by reductive amination with NaBH3CN. Attachment of labels by either procedure did not affect the normal rapid clearance of asialofetuin from the rat circulation nor its uptake and degradation in liver lysosomes. Leakage of 125I-labeled degradation products from cells was measured by following the kinetics of loss of whole-body radioactivity. We observed that degradation products from larger, disubstituted glycoconjugates were retained more efficiently than those from smaller and monosubstituted derivatives, and that glycoconjugates coupled to protein via reductive amination were retained in the body more efficiently than those coupled by cyanuric chloride. Overall, dilactitol-[125I]tyramine coupled to protein by reductive amination was entrapped most efficiently in liver.

Characterization of glycated proteins by 13C NMR spectroscopy. Identification of specific sites of protein modification by glucose

13C NMR spectroscopy has been used to characterize Amadori (ketoamine) adducts formed by reaction of [2-13C]glucose with free amino groups of protein. The spectra of glycated proteins were acquired in phosphate buffer at pH 7.4 and were interpreted by reference to the spectra of model compounds, N alpha-formyl-N epsilon-fructose-lysine and glycated poly-L-lysine (GlcPLL). The anomeric carbon region of the spectrum (approximately 90-105 ppm) of glycated cytochrome c was superimposable on that of N alpha-formyl-N epsilon-fructose-lysine, and contained three peaks characteristic of the alpha- and beta-furanose and beta-pyranose anomers of Amadori adducts to peripheral lysine residues on protein (pK alpha approximately 10.5). The spectrum of GlcPLL yielded six anomeric carbon resonances; the second set of three was displaced about 2 ppm to lower shielding of the first and was assigned to the Amadori adduct at the alpha-amino terminus (pK alpha approximately 7.5). The spectrum of glycated RNase was similar to that of GlcPLL, but contained a third set of three signals attributable to modification of active site lysine 41 (pK alpha approximately 8.8). The assignments for RNase were confirmed by analysis of spectra taken at pH 4 and under denaturing conditions. The spectrum of glycated hemoglobin was comparable to that of GlcPLL, and distinct resonances could be assigned to Amadori adducts at amino-terminal valine and intrachain N epsilon-lysine residues. Chemical analyses were performed to measure the relative extent of alpha- and epsilon-amino group modification in the glycated macromolecules, and the results were compared with estimates based on integration of the NMR spectra.

Changes in blood protein glycosylation during a diabetes summer camp

To evaluate changes in glycemic control during a 2-wk diabetes summer camp program, fasting plasma glucose (FPG), glycosylated hemoglobin (GHb), and glycosylated serum protein ( GSP ) levels were measured in a group of 36 children at the beginning and end of camp. Average FPG and GHb were unchanged during the 2-wk period, but the average decrease in GSP (7%) was significant (P less than 0.005). The results of this study indicate that a measurable improvement in diabetic control occurred in some children during the 2-wk summer camp program.

13C NMR investigation of nonenzymatic glucosylation of protein. Model studies using RNase A

Nonenzymatic glucosylation of protein is initiated by the reversible condensation of glucose in its open chain form with the amino groups on the protein. The initial product is an aldimine (Schiff base) which cyclizes to the glycosylamine derivative. The aldimine can undergo a slow Amadori rearrangement to yield the relatively stable ketoamine adduct which is structurally analogous to fructose. 13C NMR has been used to characterize these early products of nonenzymatic glucosylation, using RNase A as a model protein. C-1 of the beta-pyranose anomer of the glycosylamine was identified at 88.8 ppm in the spectrum of RNase glucosylated approximately 1:1 with D-[1-13C]glucose. C-1 of the Amadori product was also apparent in this spectrum, resonating as a pair of intense peaks at 52.7 and 53.1 ppm. The anomeric (C-2) resonances of the Amadori adduct were seen in the spectrum of RNase glucosylated approximately 1:1 with [U-13C]glucose. This spectrum was interpreted by comparison to the spectra of reference compounds: D-fructose, fructose-glycine, N alpha-formyl-N epsilon-fructose-lysine, and glucosylated poly-L-lysine. In the protein spectrum, the most intense of the C-2 resonances was that of the beta-fructopyranose anomer at 95.8 ppm. The alpha- and beta-fructofuranose anomers were also observed at 101.7 and 99.2 ppm, respectively. One unidentified signal in the anomeric region was observed in the spectra of poly-L-lysine and RNase, both glucosylated with [U-13C]glucose; no comparable resonances were observed in the spectra of the model compounds.

Determination of the specific activity of NaB3H4

A method for the rapid determination of the specific activity of NaB3H4 is presented. NaB3H4 is used to reduce NAD+ to [3H]NADH, which is then isolated by anion exchange chromatography. The specific activity of the NaB3H4 is calculated from measurements of radioactivity and absorbance (340 nm) in the [3H]NADH fractions.

A rapid micro-modification of the quantitative immunoprecipitation assay

A micro-modification of the quantitative immunoprecipitin test is described which permits rapid and sensitive titering of antisera, or quantitation of antigen, within a few hours. The procedure requires less than a microgram of radiolabeled antigen and can be used to titer antisera containing less than 1 microgram/ml of precipitating antibody. Radiotiters obtained for commercial antisera agree closely with results obtained using the classical quantitative immunoprecipitation procedure. The radiotiter assay procedure has been adapted for quantitating albumin in human urine by radioimmunoassay. The analysis can be performed in approximately 1 h, is sensitive to less than 50 ng of albumin, and requires no special apparatus or separation techniques.

Nonenzymatic glucosylation and glucose-dependent cross-linking of protein

A model system using RNase A has been established for studying the nonenzymatic glucosylation and glucose-dependent cross-linking of protein (Maillard reaction) under physiological conditions in vitro. The rate of glucosylation of RNase was first order in glucose. Glucosylation was accompanied by a comparable decrease in primary amino groups in the protein and lysine recoverable by amino acid analysis. Analysis of glucosylation reaction mixtures by sodium dodecyl sulfate-polyacrylamide gel electrophoresis in the presence of mercaptoethanol revealed the time-dependent formation of RNase dimer and trimer. The polymerization reaction was mixed order with respect to glucose concentration, but was approximately first order with respect to protein concentration. When glucosylated protein was separated from glucose, the protein continued to polymerize even in the absence of glucose. Under these conditions, the primary cross-linking reaction occurred by condensation of a glucosylated amino acid on one RNase molecule with a free amino group on another. Lysine efficiently inhibited cross-linking between glucosylated and native RNase in the absence of glucose. An attempt to model the cross-linking reaction was made by studying the incorporation of [3H]lysine and N alpha-formyl-[3H]lysine into glucosylated RNase. Both were incorporated covalently into glucosylated but not native protein. However, free lysine was the major product recovered following NaBH4 reduction and amino acid analysis of the lysine derivative of glycosylated protein. The data are discussed in terms of the mechanism of protein cross-linking by glucose and the relevance of this reaction to the pathophysiology of diabetes.

Albumin catabolism in diabetic rats

The kinetics of albumin catabolism were studied in normal rats and rats with streptozotocin induced diabetes (blood glucose greater than 500 mg%). Whether determined from the clearance of 125I-albumin from plasma or from the whole body, after 10 days of severe, uncontrolled diabetes there was a 30-35% decrease in the catabolic rate for albumin in the diabetic rats compared to normals. There was also about a 35% contraction of the relative extravascular distribution volume for albumin in the diabetic rats, and about a 25% decrease in the total body mass of albumin. However, the concentration of albumin in the circulation was the same in normal and diabetic animals. We conclude that when the rate of albumin synthesis is substantially depressed in diabetes, the rat maintains normal plasma albumin concentration both by decreasing albumin's fractional catabolic rate and by shifting albumin from the extravascular to the vascular compartment.

Measurement of nonenzymatically glucosylated serum protein by an improved thiobarbituric acid assay

We describe here some useful modifications of the thiobarbituric acid (TBA) assay for measurement of nonenzymatic glucosylation of serum protein. The modified assay minimizes interference by glucose without a lengthy dialysis step, and does not require an independent blank determination. These modifications should make the TBA assay more convenient for evaluating glycemic control in diabetes. Serum protein is first precipitated with cold ethanol to remove endogenous glucose. The protein is then hydrolyzed in an oxalic acid solution to release glucose as hydroxymethylfurfural (HMF). The HMF is reacted with TBA to form a chromophore which is extracted into isobutanol for spectrophotometric analysis (lambda max = 435 nm). The absorbance at 435 nm is corrected by subtracting a blank reading at 500 nm, and the nmol HMF released is determined using a standard curve prepared with pure HMF. Normal values of this assay for both adults and children are 0.38 +/- 0.10 nmol HMF/mg serum protein (means +/- 2 SD). When the assay was applied to serum samples from a group of 39 Type I diabetic children more than 90% of the children exceeded the normal range of the assay.

Carbohydrate-mediated clearance of immune complexes from the circulation. A role for galactose residues in the hepatic uptake of IgG-antigen complexes

The purpose of these experiments was to evaluate the hypothesis that galactose residues on IgG mediate the clearance of IgG.antigen complexes from the circulation. After 28 days of immunization of rats with bovine serum albumin (BSA), approximately 90% of anti-BSA antibody was IgG; the circulating half-life of trace amounts of BSA antigen in immunized rats was 6 min, compared to 24 h in nonimmunized rats. Similarly, soluble IgG.125I-BSA complexes formed in vitro, under conditions of antibody excess, had a circulating half-life of 4 min in normal rats. For both antigen in immunized rats, or IgG.125I-BSA complexes in normal animals, clearance was markedly inhibited by pre- or co-injection of asialofetuin, but was insensitive to large doses of fetuin, ovalbumin, or mannan. Liver parenchymal cells were the major site of uptake of complexes formed in vivo or in vitro. In vitro binding of IgG.125I-BSA complexes by isolated hepatocytes was effectively competed by asialofetuin, asialo-orosomucoid, galactose, and N-acetylgalactosamine, but was unaffected by fetuin, orosomucoid, ovalbumin, mannan, or mannose. These data suggest that antigen-induced conformational changes in IgG result in both recognition of galactose residues on IgG and clearance of IgG-immune complexes from the circulation by the galactose-specific receptor in hepatic parenchymal cells.

Nonenzymatic glucosylation of serum proteins and hemoglobin: response to changes in blood glucose levels in diabetic rats

The relationship between concentrations of blood glucose and nonenzymatically glucosylated serum proteins was studied in rats with alloxan-induced diabetes of varying severity. Fasting serum glucose correlated strongly with both glucosylated albumin (r = 0.91, P less than 0.001) and glucosylated serum protein (r = 0.93, P less than 0.001). The relative rates of response of serum protein and hemoglobin glycosylation to changes in blood glucose were also compared. Following withdrawal of insulin from diabetic rats, the half-times to reach new steady state levels of blood glucose, glucosylated serum proteins, and glycohemoglobins were about 2, 3, and 8 days, respectively. Similarly, on reinstitution of insulin therapy, the half-times for these same indices to return to baseline values were 2, 3.5, and 15 days, respectively. Changes in glucosylated albumin were more sensitive than glycohemoglobins to changes in serum glucose, consistent with the observation that albumin was glucosylated at about 10 times the rate for hemoglobin in incubations in vitro. These data indicate that glucosylated serum protein measurements can serve as a sensitive, short-term integrator of blood glucose homeostasis in diabetes.

Carbohydrate-mediated clearance of antibody . antigen complexes from the circulation. The role of high mannose oligosaccharides in the hepatic uptake of IgM . antigen complexes

After immunization of rats with bovine serum albumin (BSA) for a 12-day period, approximately 90% of anti-BSA antibody was IgM. The circulating half-life of limiting amounts of 125I-BSA in immunized and control rats was 6 min and 24 h, respectively. The rapid clearance of 125I-BSA was inhibited by pre- or co-injection of mannan and ovalbumin, but not by asialofetuin, rat serum albumin, carbon particles, dextran, or depletion of serum complement. Soluble IgM . 125I-BSA complexes, formed in vitro under conditions of antibody excess, were rapidly cleared from the circulation of nonimmunized rats, and clearance was also inhibited by ovalbumin but not by asialofetuin. Immune complexes formed in vivo or in vitro were recovered primarily (approximately 60% of dose) in hepatic nonparenchymal cells and in other organs of the reticuloendothelial system. In experiments in vitro, IgM was bound tightly by concanavalin A only when complexed with antigen. Digestion of IgM . 125I-BSA complexes with alpha-mannosidase abolished both binding by concanavalin A and rapid clearance in normal rats. These data suggest that antigen-induced conformational changes can result in exposure of high mannose oligosaccharides on IgM which signal the clearance of soluble immune complexes from the circulation.

Nonenzymatic glucosylation of serum proteins in diabetes mellitus

The extent of nonenzymatic glucosylation of serum protein in control and diabetic subjects was measured by a chemical procedure using thiobarbituric acid. A mean value of 0.81 (+/- 0.21 SD) nmol glucose per milligram serum protein was observed in the control group. Diabetics displayed elevated levels of glucosylated serum proteins, up to 4 nmol glucose per milligram protein. Glucosylation of serum protein correlated strongly with fasting blood sugar (r = 0.71), percent hemoglobin A1 (r = 0.79), and percent glucosylated albumin (r = 0.99). There was no overlap between control and diabetic groups, i.e., within 3 SD of the mean of controls. These studies indicate that the assay for glucosylated serum protein appears to be an especially sensitive indicator of the degree of hyperglycemia in diabetes.

Enhanced nonenzymatic glucosylation of human serum albumin in diabetes mellitus

Use of an ion exchange chromatographic method and a colorimetric method with thiobarbituric acid showed that levels of nonenzymatically glucosylated serum albumin were increased in patients with poorly controlled diabetes mellitus compared to controls. The two methods correlated well (r = 0.99) and clearly discriminated between normal and poorly controlled diabetic populations. The levels of glycosylated hemoglobin were also measured in both populations. Several patients apparently in good control based on glycosylated hemoglobin measurements were found to have increased levels of glycosylated albumin. Because albumin has a shorter circulating half-life than does the human erythrocyte, the plasma concentration of glucosylated albumin should be expected to reflect short-term control of hyperglycemia in diabetes. The studies reported here suggest that the level of glucosylated albumin may indeed be a sensitive indicator of moderate hyperglycemia and of early glucose intolerance.

Nonenzymatically glucosylated albumin. In vitro preparation and isolation from normal human serum

Incubation of human serum with D-[6-3H]glucose resulted in the gradual accumulation of radioactivity in acid-precipitable material. Upon chromatography on Sephadex G-200, radioactivity was found associated with each of the major molecular weight classes of serum protein. Purified human serum albumin was also glucosylated in vitro upon exposure to D-[6-3H]glucose in phosphate-buffered saline. The glucosylated and unmodified albumins were separated by ion exchange chromatography. The physiological significance of these observations in vitro was confirmed by the isolation and quantitation of glucosylated albumin from normal human serum. Glucosylated albumin represents approximately 6 to 15% of total serum albumin in normal adults. The post-translational modification appears to occur by a nonenzymatic process analogous to that responsible for glucosylation of hemoglobin A to hemoglobin AIc, i.e. through Schiff base formation and Amadori rearrangement to a ketoamine derivative.

Enzyme therapy. V. In vivo fate of erythrocyte-entrapped beta-glucuronidase in beta-glucuronidase-deficient mice

The use of erythrocyte entrapment as a strategy to deliver and protect exogenously administered enzymes for replacement therapy in selected genetic diseases has been evaluated in a mammalian model system. The uptake, tissue distribution, intracellular localization, and in vivo lifetime of erythrocyte-entrapped bovine beta-glucuronidase were determined by a selective thermal inactivation assay after intravenous administration into beta-glucuronidase-deficient mice. The exogenous activity was cleared from the circulation with a half-life of about 20 min and was no longer detectable at 2 hr. A concomitant uptake of the injected enzyme was observed in murine tissues, primarily the liver; approximately 30% of the bovine activity was recovered at 30 min and maximal hepatic uptake, 71% of dose, was detected at 2 hr. Hepatic recovery of the bovine activity was observed to decrease in a biphasic pattern to nondetectable levels by 5 days. The recovery of the entrapped activity was characterized by a latency of detection in hepatic tissue up to 13 hr postinjection. At each time point more than 80% (84-100%) of the recovered bovine activity was detected in the lysosomally enriched hepatic subcellular fraction. Maximal recoveries of 10% and 15% of administered dose were observed in splenic and renal tissues, respectively, soon after enzyme administration. In comparison to results obtained after intravenous administration of unentrapped bovine beta-glucuronidase, erythrocyte-entrapped activity was retained fourfold longer in the circulation, fivefold longer in hepatic tissue, and was more efficiently delivered to a variety of tissues.