The oxidative mechanism of heparin interferes with radical production by glucose and reduces the degree of glycooxidative modifications on human serum albumin

Eur. J. Biochem. 268, 2193±2200 (2001) q FEBS 2001

P R I O R I T Y PA P E R

The oxidative mechanism of heparin interferes with radical productionby glucose and reduces the degree of glycooxidative modifications onhuman serum albumin

Paola Finotti1, Andrea Pagetta1 and Tony Ashton2

1Department of Pharmacology and Anaesthesiology, University of Padova, Italy; 2School of Applied Sciences, University of Glamorgan,

Pontypridd, UK

Among substances which may prove useful in preventing or

reducing the progression of glycooxidative modifications of

proteins, heparin plays a unique role. To elucidate the mech-

anism whereby heparin may favourably influence the protein

structure during glycation, human serum albumin (HSA)

was glycated with both 25 and 50 mm glucose in the

absence and presence of 12 mg´mL21 low-molecular-mass

heparin. Glycation caused: (a) modifications of fluores-

cence emission and excitation spectra consistent with the

covalent attachment of glucose to protein; (b) a significant

increase in the esterase activity of HSA on p-nitrophenyl

acetate; (c) a reduced susceptibility to tryptic digestion and

(d) enhanced formation of high-molecular mass aggregates

of HSA. These alterations were accompanied by oxidative

reactions, as the EPR spectra showed a clear-cut radical

signal, dependent on glucose concentration, further confirmed

by measurement of the carbonyl content of HSA, as an

indirect proof of oxidative damage. In the presence of heparin

all the above alterations, especially at 25 mm glucose,

turned out to be antagonized. The effects of heparin were

dependent on its specific binding to HSA, which triggered

an oxidative mechanism strikingly different from that

caused by glucose. In the presence of heparin, only the

radical species catalyzed by heparin was detected across all

samples of glycated HSA, irrespective of glucose concen-

tration. In addition, at 25 mm glucose, enhancement of the

oxidative capacity of heparin was also observed. The results

demonstrate that the oxidative mechanism sustained by

heparin mediates biological effects that may be beneficial

in reducing the extent of glycooxidative damage on HSA.

Keywords: serum albumin; heparin; protein structure;

glycation; oxidation.

Nonenzymatic glycosylation (glycation) of both circulatingand structural proteins is a process of particular physio-pathological relevance for the development and progressionof vascular complications in both ageing and various patho-logical conditions such as diabetes mellitus [1,2]. Glycationpathways involve a series of complex, multistep reactionswhich, through glucose condensation with free aminogroups of lysine residues on protein, lead to the formationof not yet fully characterized products of advancedglycosylation (AGEs) [3]. Although initial reactions ofthe glycation process appear to be reversible, subsequentreactions give rise to irreversible cross-linked rearrange-ment products of glucose with proteins, whose structuraland functional properties undergo profound alterations[4±6]. The pathological consequences of these alterationscrucially depend on the nature of proteins involved as wellas on their function and concentration in specific anatomi-cal districts. Among the circulating proteins, serum albumin

is the most intensively studied, given its abundant plasmaconcentration and unusually numerous and differentiatedfunctions. Glycated and AGE-modified albumins not onlydisplay a characteristic novel fluorescence [2,7] but alsoshow strikingly different functional and immunologicalproperties [7±9]. A specific receptor for AGE-modifiedproteins has been identified, and reported to mediate thesynthesis and activation of different cellular factors, whichin the endothelial cells turn out to be responsible foralterations of vascular permeability [10±12]. In the kidney,defects in vascular permeability may lead to leakage ofalbumin in the urine, which represents an early marker ofthe vascular complications in insulin-dependent diabetesmellitus [13].

Evidence has accumulated which supports the hypothesisthat glycation and oxidation are closely connected processes[14]. Alterations induced by AGE-modified proteins aftertheir binding to specific vascular receptors are known to bemediated by oxidant stress [2,15] and the presence of O2 isrequired for the formation of some AGEs which in turnyield radicals responsible for acceleration of glucose-mediated protein cross-linking [16,17]. The cross-linkedSchiff base generated by protein glycation has also beendemonstrated to be an active center capable of catalyzingoxidation-reduction reactions [18]. Although there is nodirect proof that the glycooxidative modifications of pro-teins lead to the degenerative processes characterizing

Correspondence to P. Finotti, Department of Pharmacology

University of Padova, Largo E. Meneghetti 2, 35131 Padova,

Italy. Fax: 1 39 0498275093, Tel.: 1 39 0498275088,

E-mail: [email protected]

Abbreviations: HSA, human serum albumin; AGEs, advanced

glycosylation end-products; PBN, a-phenyl N-tert-butylnitrone.

(Received 22 February 2001, accepted 28 February 2001)

various pathological conditions, several observations indir-ectly demonstrated that oxidative modifications of proteinsare implicated in ageing and numerous diseases, and areworsened by hyperglycaemia [4,19]. Thus, there is a linkbetween radical formation and progression of variousdiabetic complications [20,21], and furthermore reactivespecies have been implicated as mediators of early islet bcell damage in animal models of insulin-dependent diabetesmellitus [22].

The overall oxidative damage not only depends on theformation and accumulation of oxidation endproducts, butalso on the efficiency of specific systems aimed at prevent-ing and eliminating them [4,17]. Defects of antioxidativedefenses are reported in both ageing and diabetes [23],and accumulation of oxidized proteins is inversely relatedto the activity of alkaline proteases responsible for eliminat-ing damaged proteins [4]. Although oxidized proteins arereported to be a good substrate for proteases, oxidativedamage renders proteins more resistant to the proteolyticattack [4,6,24]. Much effort has thus been spent searchingfor substances capable of either preventing or arresting theprogression of glycooxidation-dependent complications.Heparin may be one of the most promising substances, asit has been shown to antagonize glucose at glycosylationbinding sites on proteins [25,26]. Heparin administration todiabetic patients has also been observed to lead to amelior-ation of hyperglycaemia-related alterations of coagulationparameters and a reduction in urinary loss of albumin [27].Recently, it has been demonstrated that heparin binds toalbumin in a specific manner and mutually affects bothglucose binding and glucose-dependent structural altera-tions of nonglycated albumin [28]. Because heparin hasboth reducing [29] and oxidizing properties [28,30], andbecause glycated albumin has been shown to displayoxidation-reduction properties in relation to free radicalgeneration [18], it was of interest to investigate the effectsof heparin on glycated albumin, more specifically todetermine whether heparin was able to interfere withglucose-dependent oxidative modifications of the protein.To this purpose, the effects of low-molecular-mass heparin,used at a concentration falling in the range of those foundduring a common therapeutic regimen in vivo, were testedon HSA during an in vitro glycation process in the presenceof glucose at concentrations which simulate the hyper-glycaemic condition.

M A T E R I A L S A N D M E T H O D S

Materials

Globulin-free HSA (A-8763), 2,4-dinitrophenylhydrazine,TPCK-treated bovine trypsin, a-phenyl N-tert-butylnitrone(PBN) were from Sigma Chemicals Co. (St Louis, MO,USA). Low-molecular-mass heparin (4.6 kDa) was a gener-ous gift of B. Casu (`G. Ronzoni' Institute of Chemistry andBiochemistry, Milan). Its purity was assessed as reportedpreviously [31]. All other chemicals were of the highestavailable grade.

Solutions of glycated HSA

Solutions of 0.68 mg´mL21 (10.4 mm) HSA were made upin 20 mm sodium phosphate buffer, pH 7.0, filtered through

0.2-mm Millipore filters before use and incubated understerile conditions in capped vials for 4.5 and 8 days at37 8C in a stirring bath, without and with d-glucose (25 and50 mm, final concentrations) in the absence and presence of12 mg´mL21 heparin. While 25 mm glucose was chosen asit represents the concentration encountered during an acutehyperglycaemic condition, 50 mm was used as a model forlong-term effects of hyperglycaemia on HSA. Unboundglucose and heparin were always removed after incubationby exhaustive dialysis against pure distilled water usingSlide-A lyser dialysis cassettes of 10 000 MWCO (Pierce).Solutions were stored at 280 8C before being analyzed.Protein concentration was measured spectrophotometricallyat 280 nm using an absorption coefficient :1%

280nm �5.3 m21´cm21 [32]. In separate experiments, solutions ofHSA glycated with 25 and 50 mm glucose, as specifiedabove, were incubated again at 37 8C for 2, 4 and 6 h in thepresence of 12 mg´mL21 heparin. At the end of eachincubation interval, aliquots were drawn for fluorescencespectra measurements and electrophoretic analyses. For theelectron paramagnetic resonance (EPR) analysis, solutionswere incubated in the presence of 25 mm PBN as spintrapping agent.

Spectroscopic analyses

Ultraviolet and fluorescence spectra were measured on aShimadzu UV-160 visible recording spectrophotometer anda PerkinElmer LS-3 spectrofluorimeter, respectively. UVspectra were recorded from 350 to 200 nm. Emissionspectra were in the range from 310 to 400 nm afterexcitation at both the wavelengths of 295 and 280 nm;excitation spectra were recorded from 250 to 410 nm at thefixed emission wavelength of 440 nm, in a 1.0±cmrectangular quartz cuvette at room temperature.

For the EPR analysis, 1.0 mL of HPLC grade toluene(Sigma Ltd) was added to 0.5 mL of each solution andvortex mixed for one minute. 300 mL of toluene waspipetted into a precision bore quartz EPR sample tube andvacuum degassed using a turbo and rotary pump system(West Technologies) in a freeze-pump-thaw procedure to1023 Torr (Pirani 14, Edwards Ltd) for three consecutive10 minute cycles. The sample was immediately analyzedunder vacuum at room temperature using a Bruker X-bandEMX spectrometer (Karlsruhe, Germany) under the follow-ing conditions: incident microwave power, 20.117 mW;frequency, 9.722 GHz; modulation amplitude, 1.50 G; modu-lation frequency, 100.00 kHz; time constant, 163.84 ms;sweep time, 167.77 s; receiver gain, 1.00 � 1025; sweepwidth, 100.00 G; field centre, 3300.00 G. EPR controlexperiments were also performed, for example, analyzing ablank sample of PBN and toluene, for the presence ofartefactual radicals. No EPR signals were detected in any ofthe reagents used in EPR analysis, additionally all spectrawere initially analyzed using 1000 G scan width which wasthen narrowed to focus in on the PBN adducts detected.

Electrophoresis and immunoblotting

Unless otherwise specified, 10 mg of HSA glycated with 25and 50 mm glucose was processed in SDS/PAGE on 10%acrylamide gel, according to the method of Laemmli [33] inboth the absence and presence of reducing treatment with

2194 P. Finotti et al. (Eur. J. Biochem. 268) q FEBS 2001

2-mercaptoethanol and boiling. In native electrophoresis,SDS was replaced by Triton X-100 (10%) in both runningand sample buffers. Gels were stained with Coomassiebrilliant blue. Immunoblotting was performed on 4 mgsample proteins electroblotted from the gel on poly(vinyli-dene difluoride) membrane (Immobilon, Millipore) usingan anti-HSA serum (1 : 130 dilution in 10 mm Tris/HClwith 0.15 m NaCl) as first antibody, and a donkey anti-(sheepIgG) Ig±alkaline phosphatase conjugate (The Binding Site,Birmingham, UK) as the second antibody (at a 1 : 5000dilution).

Tryptic digestion of HSA

200 mg of HSA glycated with both 25 and 50 mm glucosewere submitted to tryptic digestion by incubating thesamples at the final protein-to-trypsin weight ratio of200 : 1 in 0.1 mL 0.1 m NH4HCO3, pH 7.83, at 37 8C for15, 30, 60 and 120 min. Proteolysis was terminated byboiling in SDS sample buffer and the degree of HSAdigestion evaluated by SDS/PAGE on 15% gel acrylamidefollowed by immunoblotting.

Measurement of carbonyl content

This was performed by using 2,4-dinitrophenylhydrazine ascarbonyl reagent [34]. The carbonyl content, taken as ameasure of oxidation extent, was calculated from averageabsorbances at both 360 and 390 nm with a molarabsorption coefficient of 22 000 m21´ cm21. Values areexpressed as mol´CO mol protein21. Group means werecompared by means of one-way anova and pairwisecomparisons were performed according to Bonferroni'smultiple comparison test (Graph Pad prism, version 2.0,San Diego, CA, USA).

Esterolytic activity

The activity was measured on p-nitrophenyl acetate sub-strate by monitoring spectrophotometrically at 400 nm theformation of p-nitrophenol. The reaction mixtures con-tained, in a 1.5-mL final volume, 2.5 mm substrate and8.6 mm sample albumin in 20 mm sodium phosphate buffer(pH 7.0). Under these conditions, first-order rate constantcan be determined [35]. The reactions were started by theaddition of substrate and were followed at 37 8C. Non-linear regression analysis was used to calculate pseudofirst-order association kinetics parameters (Graph Padprism).

R E S U LT S

At variance with the UV spectra which did not show anysignificant modifications, after glycation the fluorescenceemission intensity of both tyrosine (Fig. 1, upper panels)and tryptophan residues of HSA (not shown) was found tobe reduced (5% and 10% reduction at 25 and 50 mmglucose, respectively), without any shift in the wavelength.In the presence of heparin, fluorescence quenching wasantagonized at 25 mm but not at 50 mm glucose (Fig. 1,upper panels, spectra c). In the excitation spectra, a novelfluorescence with an emission maximum at 440 nm forexcitation at 325 nm was apparent, consistent with the

formation of covalent adducts of glucose with albumin,whose intensity appeared to be dependent on glucoseconcentration (Fig. 1, lower panels). It is worth noting thatthe same pattern was observed in the excitation spectra ofHSA samples submitted to glycation for up to 8 days (notshown); this indicates that no further increase in AGEformation occurs when the incubation was prolonged from4.5 to 8 days (for this reason the shorter incubation timewas taken thereafter). In the presence of heparin, again thefluorescence was completely abolished at 25 mm glucose,whereas it was only partially reduced at 50 mm glucose(Fig. 1, lower panels, spectra c).

The electrophoretic pattern of glycated HSA did notshow any substantial modification compared to the control(nonglycated HSA) under denaturing conditions (Fig. 2A:lanes 2 and 3 vs. 1), whereas under nondenaturing electro-phoresis high-molecular-mass aggregates of albumin werevisible in a band near the cathode (arrow), especially whenglycation was performed with 50 mm glucose (Fig. 2B:lane 3). This picture was further confirmed by a moreintense immunoprecipitation with anti-HSA IgG in the

Fig. 1. Emission and excitation spectra of HSA glycated with 25

and 50 mm glucose, in the absence and presence of heparin. Spectra

were recorded as specified in Materials and methods on incubated HSA

solutions after dilution of proteins to 30 mg´mL21. Upper panels:

emission spectra recorded at the indicated wavelength range of

emission after excitation at 280 nm. For clarity, the spectra obtained

after excitation at 295 nm for specific Trp emission are omitted, but the

pattern observed is similar to that shown for Tyr emission. Lower

panels: excitation spectra recorded at the indicated wavelength range of

excitation for emission at 440 nm. In each panel spectra are: (a) control

HSA; (b) HSA glycated with the specified glucose concentration; (c)

glycated HSA with heparin.

q FEBS 2001 Heparin antagonizes albumin glycooxidative damage (Eur. J. Biochem. 268) 2195

immunoblot analysis (data not shown). In the presence ofheparin, albumin aggregates almost disappeared, especiallyat 25 mm glucose (Fig. 2B: lane 4) and two novel bands, ofapparent molecular masses of 74 and 54 kDa, were alsovisible in SDS/PAGE (Fig. 2A: lanes 4 and 5). Densitometricanalysis showed that the band at 54 kDa displayed similarintensity irrespective of glucose concentration, whereas the74-kDa band appeared to be slightly more intense at 25 thanat 50 mm glucose (data not shown). In nondenaturingelectrophoresis, only one novel band, running above theHSA monomer, was apparent with 25 mm glucose(Fig. 2B: lane 4, arrow). It is worth noting that in theabsence of glucose, heparin was able to produce the samequalitative effect on incubated HSA (data not shown),similar to previously reported observations on nonglycatedHSA [28], but the intensity of the effect was shown to behigher in the presence of glucose, especially 25 mm.

In order to test whether heparin changed the pattern ofproteolytic fragmentation of glycated HSA, experimentswere set up in which protein digestion was performed withtrypsin at different incubation times. After glycation, HSAunderwent a reduced proteolytic pattern (Fig. 3, lanes 2 and3 vs. 1), in accord with the notion that glycated proteins aremore resistant to proteolysis than nonglycated ones [4,6,24].Interestingly, heparin appeared to reverse this picture,yielding bands of digestion which were already visible atshorter incubation times (15 min, not shown, and 30 min,Fig. 3) which differed in mobility and intensity of stainingfrom those of the control. This suggests that in the presenceof heparin, glycated HSA undergoes an increased trypticdigestion at sites which do not overlap exactly the trypticcleavage points in control HSA. The effects of heparin weremore evident when glycation was performed with 25 than50 mm glucose, differences being mostly related to thebands in the region of molecular masses below 45 kDa(Fig. 3, lane 4 vs. 5). Immunoblot analysis confirmed thispicture, further evidencing the reduction of thickness ofthe main tryptic bands of HSA and a blurred increase ofstaining around the 14-kDa molecular mass region in thepresence of heparin consistent with a more intense diges-tion (data not shown). Moreover, while in the absence ofheparin the differences in the digestion pattern betweenglycated and nonglycated HSA were blunted after 120 minincubation with trypsin, they were still visible in thepresence of heparin.

Fig. 3. SDS/PAGE of HSA glycated in the

absence and presence of heparin after

tryptic digestion. Samples were treated as

specified in Materials and methods. 10 mg

proteins were loaded in each well without

boiling and reducing conditions. A 15%

polyacrylamide gel was used. Lanes are:

(1), control HSA; (2) and (3), HSA glycated

with 25 and 50 mm glucose, respectively;

(4) and (5), HSA glycated with 25 and

50 mm glucose, respectively, in the presence

of heparin. Only patterns after 30 and

120 min incubation with trypsin are shown.

Fig. 2. SDS and nondenaturing electrophoreses of HSA glycated in

the absence and presence of heparin. Electrophoreses were

performed as specified in Materials and methods. Both in (A) (SDS/

PAGE) and (B) (nondenaturing PAGE) sample proteins were 10 mg per

lane without boiling and reducing conditions on 10% polyacrylamide

gel. Lanes are: (1), control HSA; (2) and (3), HSA glycated with 25

and 50 mm glucose, respectively; (4) and (5), HSA glycated with 25

and 50 mm glucose, respectively, in the presence of heparin.


Measurement of CO content was performed to deter-mine the extent of oxidative modification of the protein.The basal value of incubated, nonglycated HSA was0.131 ^ 0.0086 mol CO per mol protein (mean ^ SD offour independent measurements carried out in duplicate),which is similar to that reported previously for non-glycated HSA [28]. Importantly, because the CO contentof incubated HSA overlapped that measured on fresh HSA(0.134 ^ 0.0077), the implication is that incubation on itsown does not lead to any oxidative modification of theprotein. In the presence of glucose there was a dose-dependent increase in the CO content (0.155 ^ 0.0079and 0.173 ^ 0.015 at 25 and 50 mm glucose, respectively)which reached the statistical significance at 50 mmglucose (P , 0.05). In the presence of heparin, the COcontent of HSA reached similar values regardless of theconcentration of glucose (0.160 ^ 0.0159 at 25 mm glucoseand 0.159 ^ 0.0126 at 50 mm glucose). This caused similarpercent increases over the control (22.0% and 21.4%, at25 and 50 mm glucose, respectively). In the presence ofheparin alone the CO content was 0.149 ^ 0.026 (anincrease of 14% over the control value). Apparently thus,the degree of HSA oxidation by glucose turns out to beeither enhanced or slightly reduced by heparin dependingon the glucose concentration.

The EPR spectra showed the presence of radical speciesin the samples of glycated HSA (Fig. 4). The hyperfinecoupling constants recorded from the spectra of HSAincubated with 25 mm glucose were aN � 14.04 G andaH � 2.24 G, while those recorded from the EPR spectra ofHSA incubated with 50 mm glucose were aN � 13.73 Gand aH � 2.05 G. These values suggest that the trappedadducts may be carbon-centred and originate from theoxidative modification of protein. Furthermore, the verysimilar hyperfine coupling constants imply that the speciestrapped are the same irrespective of glucose concentration.There is a 100% increase in the concentration of freeradicals in the 50 mm glucose solution compared to the25 mm glucose solution (36.4 � 103 vs. 18.3 � 103, peakheights in arbitrary units for HSA in 50 mm and 25 mmglucose, respectively). This clearly suggests that glycationof HSA leads to increased free radical production in aconcentration-dependent manner, which further supportsthe observed increase in protein carbonyl content. Additionof heparin alone to HSA resulted in production of radicalspecies (4.5 � 103 peak heights in arbitrary units) withidentical hyperfine coupling constants (aN � 13.50 G andaH � 1.90 G) to all other samples containing heparin,irrespective of glucose concentration (Fig. 4). The differentcoupling constants obtained from heparinized samplescompared to those without heparin implies that the speciesdetected in heparinized samples are different from thoseproduced by glycation alone. Addition of heparin to the25 mm-glucose HSA solution resulted in an increased EPRsignal intensity (18.3 � 103 vs. 25.6 � 103 peak heights inarbitrary units for HSA in 25 mm glucose without and withheparin, respectively). The converse was true in the 50 mmglucose HSA solution in which heparin addition caused a

Fig. 4. EPR spectra of nonglycated and glycated HSA in the

absence and presence of heparin. Spectra were measured as specified

in Materials and methods on HSA samples after incubation for 4.5 days

at 37 8C. Spectra are: (a), control HSA; (b) and (c), HSA glycated with

25 and 50 mm glucose, respectively; (d), HSA with heparin; (e) and

(f ), HSA glycated with 25 and 50 mm glucose, respectively, in the

presence of heparin.

Fig. 5. Esterase-like activity of nonglycated and glycated HSA in

the absence and presence of heparin. The activity was measured as

specified in Materials and methods following the reaction of

p-nitrophenol production from degradation of p-nitrophenyl acetate

substrate at 400 nm. Nonlinear regression was used to fit the data to the

equation describing pseudo first-order kinetics. Data points are

reported every 20 s up to 600 s, and points only every 100 s are

displayed thereafter. Curves are: X, control HSA; B, V: HSA glycated

with 25 and 50 mm glucose, respectively; W, HSA with heparin; A, K:

HSA glycated with 25 and 50 mm glucose, respectively, in the

presence of heparin.


significant decrease in the EPR signal intensity (from36.4 � 103 to 12.3 � 103 peak heights in arbitrary units, inthe absence and presence of heparin, respectively). Theseresults fit those obtained in the measurement of thecarbonyl content.

It is known that HSA, besides its wide ligand capabili-ties, possesses enzymic properties mostly exerted as anesterase activity on the substrate p-nitrophenyl acetate[36,37]. Recently, it has been confirmed that Tyr411together with Arg410 are crucial for the esterase activityof HSA [38]. Although it is known that the protease activityof plasma is increased in diabetes [39], and that glycosyla-tion of some serine proteases enhances their esteraseactivity [40], to our knowledge no study has so far beenreported demonstrating that glycation of HSA leads to anincreased esterase activity. Results of measurements of theesterolytic activity of glycated HSA, with and withoutheparin, are shown in Fig. 5. The observed pseudo first-order rate constant (kobs) calculated with the exponentialequation for nonlinear regression analysis, gave values of1.01 � 103 s21 for nonglycated, incubated HSA (control),and 1.27 � 103 and 1.25 � 103 s21 after glycation with25 mm and 50 mm glucose, respectively. The difference inthe reaction velocity between glycated and nonglycatedHSA was statistically significant (P , 0.05). In the presenceof heparin, velocity decreased significantly to similarvalues at both glucose concentrations (kobs � 0.882 � 103

and 0.872 � 103 s21 in the presence of 25 and 50 mmglucose, respectively) and almost overlapped that observedwith heparin alone (0.884 � 103 s21). Thus, whereasglycation, irrespective of glucose concentration, increasedthe esterolytic activity of HSA, addition of heparincompletely reversed this effect, driving the velocity ofreaction to values even lower than those of the controlitself.

D I S C U S S I O N

In a previous paper we demonstrated that heparin causedoxidative modifications of HSA and that this effect was dueto a specific binding of heparin to sites on HSA distinctfrom those to which glucose binds [28]. In that study effectsof both heparin and glucose were measured on HSA aftera short incubation time which did not cause any non-enzymatic glycosylation of the protein. It was thus ofinterest to verify whether heparin could also affect theglycooxidative modifications occurring in the long-termprocess of condensation of glucose with HSA. This issuehas been addressed in the present work which shows thatnot only is the glycation inextricably linked to oxidation,offering for the first time a direct proof of glucose-inducedradical production, but also that heparin does interfere withglucose, reversing the structural and functional alterationscaused by glycation.

The fluorescence spectra showed the expected time- andconcentration-dependent effects of glucose on the HSAstructure [7,14,41]. Fluorescence emission quenchingoccurred together with appearance of novel fluorescencein the excitation spectra, which strongly suggests that theformation of fluorophores is coupled with conformationalchanges of the indole ring of the Trp residue and thearomatic ring of nearby Tyr residues which become more

exposed to the solvent [42]. Quenching of the emissionduring glycation may also be due to a process of energytransfer between the chromophore residues (mostly Trp)and newly formed glycation products without significantmodifications in the protein structure [7]. Heparin antagon-ized the changes of fluorescence due to glycation with25 mm glucose, whereas, at 50 mm glucose, it was still ableto reduce the fluorescence in the excitation spectra, but didnot so for emission quenching. Similar findings, i.e. a reduc-tion in intensity of AGE-dependent fluorescence associatedwith emission quenching, was also observed after heparinwas incubated for a 6-h interval with HSA previouslyglycated with 25 mm glucose (data not shown). Theseeffects of heparin may be explained by considering thatheparin binding to HSA involves the chromophore residues,yielding prevalently quenching of the emission the intensityof which depends on the concentration of both heparin andglucose, and time of incubation [28]. Thus, while heparinbinding to HSA may partly concur to the emissionquenching at 50 mm glucose, the significant antagonismby heparin on the structural modifications due to covalentlinkages of glucose with the protein, implies that thisbinding is specific and drives changes in the structure whichovercome those of glucose. Indeed, the concentration ofheparin used here was four orders of magnitude lowerthan the lowest concentration of glucose and furthermore itfalls in the lower range of concentrations found after acommon therapeutic regimen with heparin. Thus, ourresults may have relevance for the effects observed in vivoespecially when heparin is administered in hyperglycaemicconditions.

Measurements of the esterase activity of HSA allowed usto define further the characteristics of binding and mech-anism of action of heparin. To our knowledge, this is thefirst report showing that glycation leads to a significantincrease in the esterase activity of HSA. The finding thatthe reaction velocity at 50 mm glucose was only slightlyhigher than that seen at 25 mm glucose, proves that themaximal effect is attained by a glucose concentration farlower than that responsible for the fluorophore formation.This means that a limited number of sites with highspecificity are involved in the glucose-dependent increaseof the esterase activity of HSA. Recently, Watanabe et al.confirmed that both Tyr411 and Arg410 in subdomain III ofHSA play a crucial role for the esterase activity [38]. Aplausible explanation for our results is that glycation ofsites nearby Tyr411, more specifically Lys199 and Lys525,which are recognized as nonenzymatic glycosylation sitesboth in vivo and in vitro [43,44], influence the conforma-tion of the immediate surroundings of Tyr411, leading to anaccelerated nucleophile attack by its reactive hydroxyl inthe hydrolytic reaction with the substrate. The physio-pathological relevance of this result becomes clear if oneconsiders that the plasma of insulin-dependent diabetics hasa higher than normal proteolytic activity [39], and thata proteinaceous fraction purified from the same plasmashows an increased esterolytic activity [45]. Glycation ofplasma proteins, especially albumin, which is the mostabundant among the circulating proteins, may be thusdirectly responsible for increased esterolytic activity ofdiabetic plasma, representing an accelerating factor for thedevelopment of vascular alterations in this pathologicalcondition [46].


Crystallographic studies revealed that Tyr411 and Lys199are located in the primary binding pockets of HSA, wherethey affect the binding process of several ligands [47]. Ourresult that the esterase activity was steadily inhibited byheparin, irrespective of concomitant addition of overwhelm-ing concentrations of glucose, besides being a novel effect,strongly supports the hypothesis that heparin binding toHSA involves Tyr411, either directly or indirectly throughbinding to nearby residues in the binding pocket of III Asubdomain. The possibility that heparin binding involvesboth the glycosylable residues (likely Lys525 and Lys199)and chromophore ones which all concur to form andstabilize the interface between the II A and III A sub-domains of HSA [47] fits the result that heparin was able toantagonize the fluorescence due to glycation while sustain-ing emission quenching.

This hypothesis is further corroborated by results of theEPR spectra which, together with electrophoretic analyses,shed further light on the mechanism whereby heparininterferes with glucose in the glycation process. The EPRspectra clearly showed that free radicals form during glyca-tion in a concentration-dependent manner, the result beingsupported by the increase in protein carbonyl content as afurther measure of oxidative stress. Arguably, the mostimportant aspect of the EPR data is the clear difference inthe hyperfine coupling constants between samples withheparin and those without heparin, and the finding that theconstants recorded from the spectra of samples containingheparin were identical irrespective of glucose concentra-tion. This demonstrates not only the ability of heparin toproduce free radicals independently of glycation, but thatthe species are different compared to those produced byglucose alone. While oxidative modifications accompany-ing glycation lead to the formation of high-molecular-massaggregates of the protein, which also displays an increasedresistance to the proteolytic attack, the oxidation catalysedby heparin appears to counterbalance the above structuralchanges, mostly by causing cleavage of peptide bonds,reducing protein cross-linking and rendering the proteinmore susceptible to the proteolytic digestion. Becauseglycooxidative modifications of HSA are known to beresponsible for triggering a series of biochemical reactionson vascular cell membrane which eventually lead to thedevelopment of vascular complications, these effects ofheparin may prove to be potentially beneficial as they maylead to an overall reduction in the number of AGE-modifiedHSA molecules.

Both the results that only the heparin-catalysed radicalspecies was detected in the presence of glucose, indepen-dently of its concentration, and that in the presence ofheparin the intensity of the radical signal of HSA with25 mm glucose was always more intense than that observedwith 50 mm glucose, are intriguing findings which needfurther investigation. We can only speculate about themechanism underlying these effects as the results of ourwork do not permit us to draw any definitive conclusion. Itis possible that heparin binding can either quench radicalsformed by glucose or even prevent their formation byinterfering with glucose binding. Heparin binding to HSAmay also change the reaction potential of the protein withglucose, similar to that observed with heparin-cytochrome ccomplexes, which display a modified reactivity with otherredox agents [48]. Whatever the mechanism, our results are

in keeping with numerous reports in the literature stressingboth the antioxidative properties of heparin [49] and itscapacity to enhance the oxidative modifications of protein[50] including a higher susceptibility to proteolytic digestion.Recently, it has been observed that some polymericcarbohydrates, which show similarity in structure toheparin, display both antioxidant and free radical stimulat-ing activity, depending on the concentration [51].

In conclusion, our results demonstrate the potentialtherapeutic benefit of low heparin concentration in antagon-izing both the structural and functional alterations ofprotein due to oxidative damage by glucose. Heparinbehaves as an effective antioxidative agent while displayingits own oxidative properties. However, oxidation by heparinturns out to be profoundly different from that caused byglycation, as it yields completely opposite biological effectson the protein.

A C K N O W L E D G E M E N T S

This work was supported by MURST (Ministero dell'UniversitaÁ e della

Ricerca Scientifica e Tecnologica) grants of 60% and by CNR

(Consiglio Nazionale delle Ricerche) targeted project n898.0110 CT14.

R E F E R E N C E S

1. Keaney, J.F. & Loscalzo, J. (1999) Diabetes, oxidative stress, and

platelet activation. Circulation 99, 189±191.

2. Wolff, S.P., Jiang, Z.Y. & Hunt, J.V. (1991) Protein glycation and

oxidative stress in diabetes mellitus and ageing. Free Radic. Biol.

Med. 10, 339±352.

3. McCance, D.R., Dyer, D.G., Dunn, J.A., Bailie, K.E., Thorpe,

S.R., Baynes, J.W. & Lyons, T.J. (1993) Maillard reaction products

and their relation to complications in insulin±dependent diabetes

mellitus. J. Clin. Inv. 91, 2470±2478.

4. Berlett, B.S. & Stadtman, E.R. (1997) Protein oxidation in aging,

disease, and oxidative stress. J. Biol. Chem. 272, 20313±20316.

5. Giardino, I., Edelstein, D. & Brownlee, M. (1994) Nonenzymatic

glycosylation in vitro and in bovine endothelial cells alters basic

fibroblast growth factor activity. J. Clin. Inv. 94, 110±117.

6. Anderson, S.S., Tsilibary, F.C. & Charonis, A.S. (1993) None-

nzymatic glycosylation±induced modifications of intact bovine

kidney tubular basement membrane. J. Clin. Inv. 92, 3045±3052.

7. Coussons, P.J., Jacoby, J., McKay, A., Kelly, S.M., Price, N.C. &

Hunt, J.V. (1997) Glucose modification of human serum albumin:

a structural study. Free Radic. Biol. Med. 22, 1217±1227.

8. Matsuda, T., Ishiguro, H., Ohkubo, I., Sasaki, M. & Nakamura, R.

(1992) Carbohydrate binding specificity of monoclonal antibodies

raised against lactose±protein Maillard adducts. J. Biochem. 111,

383±387.

9. Miyata, S. & Monnier, V. (1992) Immunohistochemical detection

of advanced glycosylation end products in diabetic tissues using

monoclonal antibody to pyrraline. J. Clin. Inv. 89, 1102±1112.

10. Schmidt, A.M., Vianna, M., Gerlach, M., Brett, J., Ryan, J., Kao,

J., Esposito, C., Hegarty, H., Hurley, W. & Clauss, M. (1992)

Isolation and characterization of two binding proteins for

advanced glycosylation end products from bovine lung which

are present on the endothelial cell surface. J. Biol. Chem. 267,

14987±14997.

11. Yan, S.D., Schmidt, A.M., Anderson, G.M., Zhang, J., Brett, J.,

Zou, Y.S., Pinsky, D. & Stern, D. (1994) Enhanced cellular oxidant

stress by the interaction of advanced glycation end products with

their receptors/binding proteins. J. Biol. Chem. 269, 9889±9897.

12. Khechai, F., Ollivier, V., Amar, M., Hakim, J. & de Prost, D.

(1997) Effect of advanced glycation end product-modified albumin


on tissue factor expression by monocytes. Arterioscler. Thromb.

Vascul. Biol. 17, 2885±2890.

13. Dubrey, S.W., Beetham, R., Miles, J., Noble, M.I.M., Rowe, R. &

Leslie, R.D.G. (1997) Increased urinary albumin and retinol±

binding protein in type I diabetes. Diabetes Care 20, 84±89.

14. Traverso, N., Menini, S., Cottalasso, D., Odetti, P., Marinari, U.M.

& Pronzato, M.A. (1997) Mutual interaction between glycation

and oxidation during non-enzymatic protein modification. Bio-

chim. Biophys. Acta 1336, 409±418.

15. Hunt, J.V., Bottoms, M.A. & Mitchinson, M.J. (1993) Oxidative

alterations in the experimental glycation model of diabetes

mellitus are due to protein±glucose adduct oxidation. Biochem.

J. 291, 529±535.

16. Sakurai, T. & Tsuchiya, S. (1988) Superoxide production from

nonenzymatically glycated protein. FEBS Lett. 236, 406±410.

17. Beckman, K.B. & Ames, B.N. (1998) The free radical theory of

aging matures. Physiol. Rev. 78, 547±581.

18. Lee, C., Yim, M.B., Chock, P.B., Yim, H.S. & Kang, S.O. (1998)

Oxidation±reduction properties of methylglyoxal-modified

protein in relation to free radical generation. J. Biol. Chem. 273,

25272±25278.

19. Butterfield, D.A., Howard, B.J., Yatin, S., Allen, K.L. & Carney,

J.M. (1997) Free radical oxidation of brain proteins in accelerated

senescence and its modulation by N-tert-butyl-a-phenylnitrone.

Proc. Natl Acad. Sci. USA 94, 674±678.

20. Trachtman, H., Futterweit, S., Prenner, J. & Hanon, S. (1994)

Antioxidants reverse the antiproliferative effect of high glucose

and advanced glycosylation end products in cultured rat mesangial

cells. Biochem. Biophys. Res. Commun. 199, 346±352.

21. Mullarkey, C.J., Edelstein, D. & Brownlee, M. (1990) Free radical

generation by early glycation products: a mechanism for

accelerated atherogenesis in diabetes. Biochem. Biophys. Res.

Commun. 173, 932±939.

22. Ho, E., Chen, G. & Bray, T.M. (2000) Alpha-phenyl-tert-

butylnitrone (PBN) inhibits NFkB activation offering protection

against chemically induced diabetes. Free Radic. Biol. Med. 28,

604±614.

23. Asayama, K., Uchida, N., Nakane, T., Hayashibe, H., Dobashi, K.,

Amemiya, S., Kato, K. & Nakazawa, S. (1993) Antioxidants in the

serum of children with insulin-dependent diabetes mellitus. Free

Radic. Biol. Med. 15, 597±602.

24. Qin, K., Yang, D.-S., Yang, Y., Chishti, M.A., Meng, L.-J.,

Kretzschmar, H.A., Yip, C.M., Fraser, P.E. & Westaway, D. (2000)

Copper (II)-induced conformational changes and protease resist-

ance in recombinant and cellular PrP. J. Biol. Chem. 275,

19121±19131.

25. BioÈrk, J., YlinenjaÈrvi, K., Olson, S.T., Hermetin, P., Conradt, H.S.

& Zettlmeissl, G. (1992) Decreased affinity of recombinant

antithrombin for heparin due to increased glycosylation. Biochem.

J. 286, 793±800.

26. Giardino, I., Edelstein, D. & Brownlee, M. (1994) Nonenzymatic

glycosylation in vitro in bovine endothelial cells alters basic

fibroblast growth factor activity. J. Clin. Inv. 94, 110±117.

27. Myrup, B., Hansen, P.M., Jensen, T., Kofoed-Enevoldsen, A.,

Feldt-Rasmussen, B., Gram, J., Kluft, C., Jespersen, J. & Deckert,

T. (1995) Effects of low-dose heparin on urinary albumin excretion

in insulin-dependent diabetes mellitus. Lancet 345, 421±422.

28. Finotti, P. & Pagetta, A. (1997) Heparin-induced structural

modifications and oxidative cleavage of human serum albumin

in the absence and presence of glucose. Implications for trans-

capillary leakage of albumin in hyperglycaemia. Eur. J. Biochem.

247, 1000±1008.

29. Armbruster, D.A. (1987) Fructosamine: structure, analysis, and

clinical usefulness. Clin. Chem. 33, 2153±2163.

30. Finotti, P. (1997) Biphasic pattern of heparin-induced oxidative

degradation of trypsin in the presence of glucose. Biochimie 79,

351±358.

31. Casu, B. (1989) Methods of structural analysis. In Heparin:

Chemical and Biological Properties, (Lane, D.A. & Lindhal, U.,

eds), pp. 25±49. Edward Arnold, London.

32. Bradford, M.M. (1976) A rapid and sensitive method for the

quantitation of microgram quantities of protein utilizing the

principle of protein-dye binding. Anal. Biochem. 74, 248±254.

33. Laemmli, U.K. (1970) Cleavage of structural proteins during the

assembly of the head of bacteriophage T4. Nature 227, 680±685.

34. Levine, R.L., Garland, D., Oliver, C.N., Amici, A., Climent, I.,

Lenz, A.G., Ahn, B.W., Shaltiel, S. & Stadtman, E.R. (1990)

Determination of carbonyl content in oxidatively modified

proteins. Methods Enzymol. 186, 464±478.

35. Segel, I.M., ed. (1975) Kinetics of unireactant enzymes. In Enzyme

Kinetics, pp. 18±99. John Wiley & Sons, New York.

36. Ozeki, Y., Kurono, Y., Yotsuyanagi, T. & Ikeda, K. (1980) Effects

of drug binding to the esterase activity of human serum albumin:

inhibition modes and binding sites of anionic drugs. Chem. Pharm.

Bull. 28, 535±540.

37. Salvi, A., Carrupt, P.-A., Mayer, J.M. & Testa, B. (1997) Esterase-

like activity of human serum albumin toward prodrug esters of

nicotinic acid. Drug Metab. Disp. 25, 395±398.

38. Watanabe, H., Tanase, S., Nakajou, K., Maruyama, T., Kragh-

Hansen, U. & Otagiri, M. (2000) Role of Arg-410 and Tyr-411 in

human serum albumin for ligand binding and esterase-like activity.

Biochem. J. 349, 813±819.

39. Finotti, P., Piccoli, A. & Carraro, P. (1992) Alteration of plasma

proteinase-antiproteinase system in type 1 diabetic patients.

Influence of sex and relationship with metabolic control. Diabetes

Res. Clin. Pract. 18, 35±42.

40. Lloyd, R.C., Davis, B.G. & Jones, J.B. (2000) Site-selective

glycosylation of subtilisin Bacillus lentus causes dramatic

increases in esterase activity. Biorg. Med. Chem. 8, 1537±1544.

41. Shaklai, N., Garlick, R.L. & Bunn, H.F. (1984) Nonenzymatic

glycosylation of human serum albumin alters its conformation and

function. J. Biol. Chem. 259, 3812±3817.

42. Schmid, F.X. (1990) Spectral Methods of Characterizing Protein

Conformation and Conformational Changes In Protein Structure:

a Practical Approach (Creighton, T.E., ed.) pp. 251±285. IRL

Press, Oxford.

43. Garlick, R.L. & Mazer, J.S. (1983) The principal site of non-

enzymatic glycosylation of human serum albumin in vivo. J. Biol.

Chem. 258, 6142±6146.

44. Iberg, N. & FluÈckiger, R. (1986) Nonenzymatic glycosylation of

albumin in vivo. J. Biol. Chem. 261, 13542±13545.

45. Finotti, P. & Verbaro, R. (1987) Identification and partial

purification of a (Na-K) ATPase stimulating serine protease

from plasma of insulin-dependent diabetics. Clin. Chim. Acta 170,

121±134.

46. Esposito, C., Gerlach, H., Brett, J., Stern, D. & Vlassara, H. (1989)

Endothelial receptor-mediated binding of glucose-modified

albumin is associated with increased monolayer permeability and

modulation of cell surface coagulant properties. J. Exp. Med. 170,

1387±1407.

47. He, X.M. & Carter, D.C. (1992) Atomic structure and chemistry of

human serum albumin. Nature 358, 209±215.

48. Petersen, L.C. & Cox, R.P. (1980) The effect of complex-

formation with polyanions on the redox properties of cytochrome

c. Biochem. J. 192, 687±693.

49. Arai, H., Kashiwagi, S., Nagasaka, Y., Uchida, K., Hoshii, Y. &

Nakamura, K. (1999) Oxidative Modification of apolipoprotein E

in human very-low-density lipoprotein and its inhibition by

glycosaminoglycans. Arch. Biochem. Biophys. 367, 1±8.

50. Upritchard, J.E. & Sutherland, W.H.F. (1999) Oxidation of

heparin-treated low-density lipoprotein by peroxidases. Athero-

sclerosis 146, 211±219.

51. Tsiapali, E., Whally, S., Kalbfleisch, J., Ensley, H.E., Browdler, W.

& Williams, D.L. (2001) Glucans exhibit weak antioxidant activity

but stimulate macrophage free radical activity. Free Radic. Biol.

Med 30, 393±402.


The oxidative mechanism of heparin interferes with radical production by glucose and reduces the degree of glycooxidative modifications on human serum albumin

Documents