the relative uv sensitizer activity of purified advanced glycation endproducts

7
Photochemistry and Photobiology, 1997, 65(4): 666-672 The Relative UV Sensitizer Activity of Purified Advanced Glycation Endproducts B. J. Ortwerth”, Malladi Prabhakaraml, R. H. Nagaraj2 and Mikhail Linetsky’ ‘Mason Eye Institute, University of Missouri, Columbia, MO, USA and ‘Center for Vision Research, Department of Ophthalmology, Case Western Reserve University, Cleveland, OH, USA Received 23 September 1996; accepted 17 December 1996 ABSTRACT The oxidation products of ascorbic acid react with lens proteins to form advanced glycation endproducts (AGE) that are capable of generating reactive oxygen species when irradiated with UVA light. L-Threose, the most ac- tive of these oxidation products, was reacted with N-ace- tyl lysine and six AGE peaks were isolated by RP-HPLC. Each peak exhibited fluorescence and generated super- oxide anion and singlet oxygen in response to UV light. Solutions of these AGE peaks (50 pg/mL) generated 5- 10 nmoVmL of superoxide anion during a 30 min irra- diation. This activity was 100-fold less than the super- oxide anion generated by kynurenic acid and 400-fold less than riboflavin. Ultraviolet irradiation generated from 1.2 to 2.7 pmoVmL of singlet oxygen with the purified threose AGE compounds. This activity was similar to that seen with other purified AGE compounds (pentosidine, LM-1 and Ac-FTP) and with kynurenine and 3-OH kynurenine. This considerable singlet oxygen formation, however, was still 40-fold less than that obtained with kynurenic acid and 100-fold less than riboflavin under the same irradi- ation conditions. In spite of this lower sensitizer efficien- cy, the purified AGE generated 20-60-fold more singlet oxygen on a weight basis than either crude ascorbic acid glycated proteins or a preparation of water-insoluble proteins from aged normal human lenses. On a molar basis, therefore, AGE could account for the sensitizer ac- tivity in these protein preparations if they represented less than 1% of the total amino acids. INTRODUCTION The long-term reaction of sugars with proteins results in the formation of advanced glycation endproducts (AGE).? These *To whom correspondence should be addressed at: Mason Eye In- stitute, University of Missouri, One Hospital Drive, Columbia, MO 6521 2, USA. Fax: 573-882-8474; e-mail: [email protected]. TAbbreviations: Ac-FTP, N-acetyl-formyl-threosyl-pyrrole; AGE, advanced glycation end-products; DTPA, diethylenetri- aminepentaacetic acid; NATA, N-acetyl-tryptophanamide; RNO, N,N-dimethyl-4-nitrosoaniline; ROS, reactive oxygen species; RP, reverse-phase; SOD, superoxide dismutase; TFA, trifluoroacetic acid. 0 1997 American Society for Photobiology 0031-8655/97 $S.OO+O.oO are complex aromatic structures that are responsible for browning, protein-bound fluorophores and protein-protein crosslinks (1-3). While these protein modifications can alter protein structure and the activity of enzymes (43). signifi- cant pathological damage may also result from the formation of reactive oxygen species (ROS) by these compounds (6- 8). Glycation and oxidation-linked pathology have been im- plicated in diabetes (9), atherosclerosis (lo), neuropathy (1 1) and Alzheimer’s disease (12). Glycation of lens proteins has been proposed as a major protein modification reaction contributing to the formation of age-onset and diabetic cataract (13,14). Formation of AGE can occur with either glucose or ascorbic acid (15), but both reactions require oxidative conditions to produce the reacting sugar moeity (16,17). The oxidation of glucose by hydroxyl radical and ascorbic acid by molecular oxygen both produce tetroses (18). mainly threose (19), which may be a common precursor to AGE formation. The model re- action of L-threose with a-N-blocked lysine is capable of producing several AGE molecules, which can be separated by reverse-phase HPLC (RP-HPLC). The synthesis’of AGE may have a fundamental role in brunescent cataract formation (20) because these compounds collectively can absorb UVA light and cause the production of ROS (21,22). Reactive oxygen species not only produce extensive protein damage but also accelerate the oxidation of ascorbic acid, potentially leading to the synthesis of ad- ditional AGE compounds. The formation of AGE by ascor- bic acid glycation may be particularly relevant to human cataract formation. These AGE-modified lens proteins, have the same absorption and fluorescence spectra (23) and pro- duce the same quantities of ROS as those obtained with pro- teins isolated from aged human lenses (24) when irradiated by UVA light. The research here reports on the ability of purified AGE molecules to produce both superoxide anion and singlet ox- ygen and compares these values to those obtained with tryp- tophan oxidation products and riboflavin. MATERIALS AND METHODS Reagents. L-Threose was obtained as a kind gift from the laboratory of Dr. Milton Feather, where it was prepared according to the meth- od of Perlin (25). a-N-acetyl-lysine, N,N-dimethyl-4-nitrosoaniline (RNO). diethylenetriaminepentaacetic acid (DTPA) and cytochrome C (C7752 from horse heart) were obtained from Sigma Chemical Co., St. Louis, MO. Superoxide dismutase (SOD; as a crystalline 666

Upload: b-j-ortwerth

Post on 21-Jul-2016

213 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: The Relative UV Sensitizer Activity of Purified Advanced Glycation Endproducts

Photochemistry and Photobiology, 1997, 65(4): 666-672

The Relative UV Sensitizer Activity of Purified Advanced Glycation Endproducts

B. J. Ortwerth”, Malladi Prabhakaraml, R. H. Nagaraj2 and Mikhail Linetsky’ ‘Mason Eye Institute, University of Missouri, Columbia, MO, USA and ‘Center for Vision Research, Department of Ophthalmology, Case Western Reserve University, Cleveland, OH, USA

Received 23 September 1996; accepted 17 December 1996

ABSTRACT The oxidation products of ascorbic acid react with lens proteins to form advanced glycation endproducts (AGE) that are capable of generating reactive oxygen species when irradiated with UVA light. L-Threose, the most ac- tive of these oxidation products, was reacted with N-ace- tyl lysine and six AGE peaks were isolated by RP-HPLC. Each peak exhibited fluorescence and generated super- oxide anion and singlet oxygen in response to UV light. Solutions of these AGE peaks (50 pg/mL) generated 5- 10 nmoVmL of superoxide anion during a 30 min irra- diation. This activity was 100-fold less than the super- oxide anion generated by kynurenic acid and 400-fold less than riboflavin.

Ultraviolet irradiation generated from 1.2 to 2.7 pmoVmL of singlet oxygen with the purified threose AGE compounds. This activity was similar to that seen with other purified AGE compounds (pentosidine, LM-1 and Ac-FTP) and with kynurenine and 3-OH kynurenine. This considerable singlet oxygen formation, however, was still 40-fold less than that obtained with kynurenic acid and 100-fold less than riboflavin under the same irradi- ation conditions. In spite of this lower sensitizer efficien- cy, the purified AGE generated 20-60-fold more singlet oxygen on a weight basis than either crude ascorbic acid glycated proteins or a preparation of water-insoluble proteins from aged normal human lenses. On a molar basis, therefore, AGE could account for the sensitizer ac- tivity in these protein preparations if they represented less than 1% of the total amino acids.

INTRODUCTION The long-term reaction of sugars with proteins results in the formation of advanced glycation endproducts (AGE).? These

*To whom correspondence should be addressed at: Mason Eye In- stitute, University of Missouri, One Hospital Drive, Columbia, MO 6521 2, USA. Fax: 573-882-8474; e-mail: [email protected].

TAbbreviations: Ac-FTP, N-acetyl-formyl-threosyl-pyrrole; AGE, advanced glycation end-products; DTPA, diethylenetri- aminepentaacetic acid; NATA, N-acetyl-tryptophanamide; RNO, N,N-dimethyl-4-nitrosoaniline; ROS, reactive oxygen species; RP, reverse-phase; SOD, superoxide dismutase; TFA, trifluoroacetic acid.

0 1997 American Society for Photobiology 0031-8655/97 $S.OO+O.oO

are complex aromatic structures that are responsible for browning, protein-bound fluorophores and protein-protein crosslinks (1-3). While these protein modifications can alter protein structure and the activity of enzymes (43). signifi- cant pathological damage may also result from the formation of reactive oxygen species (ROS) by these compounds (6- 8). Glycation and oxidation-linked pathology have been im- plicated in diabetes (9), atherosclerosis (lo), neuropathy (1 1) and Alzheimer’s disease (12).

Glycation of lens proteins has been proposed as a major protein modification reaction contributing to the formation of age-onset and diabetic cataract (13,14). Formation of AGE can occur with either glucose or ascorbic acid (15), but both reactions require oxidative conditions to produce the reacting sugar moeity (16,17). The oxidation of glucose by hydroxyl radical and ascorbic acid by molecular oxygen both produce tetroses (18). mainly threose (19), which may be a common precursor to AGE formation. The model re- action of L-threose with a-N-blocked lysine is capable of producing several AGE molecules, which can be separated by reverse-phase HPLC (RP-HPLC).

The synthesis’of AGE may have a fundamental role in brunescent cataract formation (20) because these compounds collectively can absorb UVA light and cause the production of ROS (21,22). Reactive oxygen species not only produce extensive protein damage but also accelerate the oxidation of ascorbic acid, potentially leading to the synthesis of ad- ditional AGE compounds. The formation of AGE by ascor- bic acid glycation may be particularly relevant to human cataract formation. These AGE-modified lens proteins, have the same absorption and fluorescence spectra (23) and pro- duce the same quantities of ROS as those obtained with pro- teins isolated from aged human lenses (24) when irradiated by UVA light.

The research here reports on the ability of purified AGE molecules to produce both superoxide anion and singlet ox- ygen and compares these values to those obtained with tryp- tophan oxidation products and riboflavin.

MATERIALS AND METHODS Reagents. L-Threose was obtained as a kind gift from the laboratory of Dr. Milton Feather, where it was prepared according to the meth- od of Perlin (25). a-N-acetyl-lysine, N,N-dimethyl-4-nitrosoaniline (RNO). diethylenetriaminepentaacetic acid (DTPA) and cytochrome C (C7752 from horse heart) were obtained from Sigma Chemical Co., St. Louis, MO. Superoxide dismutase (SOD; as a crystalline

666

Page 2: The Relative UV Sensitizer Activity of Purified Advanced Glycation Endproducts

Photochemistry and Photobiology, 1997, 65(4) 667

powder) and catalase (as a crystalline suspension) were purchased from Boehringer Mannheim Biochemicals, Indianapolis, IN. N-ace- tyltryptophanamide (NATA) and the tryptophan oxidation products, L-kynurenine, 3-hydroxy-~~-kynurenine and kynurenic acid, were obtained from Sigma Chemical Co. Purified AGE products were prepared and purified as described previously: pentosidine (26). LM I (27) and N-acetyl-formyl-threosyl-pyrrole (Ac-FTP) (28).

Isolation of AGE compounds synthesized with L-threose and a-N- acetyl-lysine. L-Threose (1.6 g) was incubated with 5.0 g of a-N- acetyl-lysine in 0.1 M phosphate buffer containing 1 .O mM DTPA for 1 week during which time extensive browning was observed. The mixture was applied in 50 mg aliquots to a 22 X 250 mm Vydac C18 column and the column eluted isocratically with 5% acetonitrile in 0.01% trifluoroacetic acid (TFA). The six major peaks were pooled and subjected to repeated HPLC chromatography under the above conditions. Peaks 3A, 3B and 4 were chromatographed on a Bio-Gel P2 column and separated once more on the C-18 Vydac column. The final preparations were lyophilized and weighed, and both absorption and fluorescence spectra were determined.

UVA irradiation. Light from a 1000 W Hg/Xe (Oriel Corp. no. 6293) lamp was filtered through a 5 % CuSO, solution and focused on a 0.25 mL quartz cuvette. This provided broadband radiation from 300 to 900 nm. Both the filter and the cuvette holder were cooled to 17°C during the irradiation. Fluence was measured to be 0.22 W/cm’ by a black block placed directly behind the cuvette holder and connected to a YSI detection meter (23).

ROS assays. Superoxide anion was determined by the SOD-in- hibitable reduction of cytochrome C (22). Reaction mixtures (0.20 mL) contained 10 pg of the compound to be assayed with 30 pg of cytochrome C and 50 units of catalase in 50 mM Chelex-treated phosphate buffer, pH 7.0 containing 0.05 mM DTPA. The absor- bance at 550 nm was followed over 30 min of irradiation (0.22 W/cm2) (20.21) and the nmoles of superoxide calculated (AEM sso ,,,,, = 2.1 X lo4 M I cm-I). Catalase was initially added at 10 UlmL and then replenished by adding 10 U/mL every 5 min due to UV inactivation. Assays were conducted in triplicate both in the pres- ence and absence of SOD and the difference calculated as nmollml of superoxide produced. Control reactions were run with and without SOD in the absence of added sensitizer.

Singlet oxygen formation was determined by the bleaching of RNO as originally described by Kraljic and El Moshni (29). Reac- tion mixtures containing 50 pg/mL of AGE compounds to be as- sayed were irradiated in the presence of 2.0 mM carnosine and 50 p l 4 RNO in a 0.2 mL reaction mixture. This assay was standardized with chemically generated singlet oxygen, which showed that a de- crease of 0.084 AUO units was equivalent to 1.0 pmol/mL of singlet oxygen generated (30). All assays were carried out in triplicate both with and without the addition of 10 mM sodium azide and the dif- ference reported both as the decrease in A,,,, and as pmollmL singlet oxygen generated.

Light absorption. Samples (1.0 mglmL) of the purified com- pounds were placed in a cuvette and the transmitted light was mea- sured and compared to the light transmission with water alone. The total light absorbed was compared to the superoxide anion and sin- glet oxygen produced by these compounds to estimate the relative sensitizer efficiency of the AGE compounds.

Spectra. Absorption spectra of the threose-lysine AGE were mea- sured in a Milton Roy Genesys V spectrophotometer. Fluorescence spectra were determined in a 650-40 Perkin Elmer spectrofluoro- meter. Emission spectra were collected with an excitation wave- length at the A,,, of each absorption spectrum.

RESULTS

L-Threose was reacted with N-acetyl-lysine and six AGE peaks were isolated by RP-HPLC. The eluant profile at 328 nm (Fig. 1) showed five distinct peaks, along with a late- eluting mixture of compounds (peak 5). Peak numbering was made with the assumption that peaks 2A and 2B as well as 3A and 3B were isomers. Zero time incubations exhibited no absorbance at 328 nm, and all of the starting materials were unable to bind to the C-18 HPLC column. Each AGE

I 5

I / I

10 20 30 40 50

Retention time (mln)

Figure 1. The separation of AGE products formed by reacting N- acetyl-lysine with L-threose for 1 week at 37°C. The profile shows the elution of A328nm material from a 22 X 250 mm Vydac C-18 RP-HPLC column with a solution of 5% methanol in 0.1% TFA.

peak was pooled and rechromatographed four to five times to purify individual AGE peaks. Absorption spectra for the various peaks are shown in Fig. 2A. Every peak had an absorption maximum of 300 nm or greater and all were fluo- rescent with emission maxima above 400 nm (Fig. 2B), which is consistent with AGE. The emission spectra of peaks 2A and 2B, as well as 3A and 3B were identical.

Each purified peak was assayed for superoxide formation during a 30 min irradiation. The curves in Fig. 3 show an almost linear increase in superoxide formation for the late- eluting peaks (3A to 4) but only a short burst followed by the apparent photodestruction of the sensitizer activity in peak 2B. The total superoxide formed at 30 min was only 4.2 2 0.2 nmoYmL for peak 2A, 4.8 -C 0.3 nmoYmL for peak 2B, 8.1 ? 0.3 nmoVmL for peak 3A, 7.2 2 0.1 nmol/mL for peak 3B, 10.1 2 0.3 nmoVmL for peak 4 and 8.4 2 0.2 nrnoYmL for peak 5. This compares to a value of 2.5 ? 0.1 nmol/mL for control reactions without any AGE compound in the irradiated cuvette (Fig. 4). This control activity was likely due to irradiation of the cytochrome c present in the assay mixture. Therefore, superoxide anion is a minor photolytic product with these AGE compounds. Only a two- to three-fold difference was seen in the rate of superoxide anion synthesis by the various AGE peaks. Peaks 3A, 3B and 4 were essentially pure by thin layer chroma- tography and Nh4R spectrometry. The structures of 3B and 4 are known (28,31). Peak 5 , however, was a complex mix- ture of compounds that had a very high absorbance in the UV region. This AGE peak, however, had no more activity than the other AGE peaks on a weight basis.

The ability of 50 pg/mL solutions of threose-lysine AGE to produce superoxide anion was compared to that produced by tryptophan, tryptophan oxidation products and riboflavin. Figure 4 shows that the activity of peak 3B was equivalent to NATA and greater than kynurenine but markedly less active than kynurenic acid, which was assayed at 1.0 pgl mL, and riboflavin, which was assayed at 0.5 pg/mL. Based on initial rates after 5 min, riboflavin was a 400-fold more efficient sensitizer. The assay of 3-OH kynurenine at 1.0 pg/mL was no different from the control assay. This com- pound absorbed a considerable amount of light and caused

Page 3: The Relative UV Sensitizer Activity of Purified Advanced Glycation Endproducts

668 B. J. Ortwerth eta/.

1 A) 1 .o

8 2

9

E

0.5

200 250 300 350 400 1

Wavelength (nm)

I I I

350 450 550

Wavelength (nm)

Figure 2. The absorption spectra (A) and the fluorescence spectra (B) of the major peaks shown in Fig. 1. The fluorescence emission spectra were obtained during excitation at the A,,, for absorption of each compound.

a rapid oxidation of cytochrome c, but very little of this oxidation was SOD dependent, as was also seen in the con- trol reaction.

Singlet oxygen generation was measured by the bleaching of RNO. The decrease in absorbance was measured over 30 min and converted to pmoYmL of singlet oxygen accumu- lated. The formation of singlet oxygen (Fig. 5) varied from 1.2 pmol to 2.7 pmoYmL over a 30 min irradiation, and all AGE were capable of generating singlet oxygen in response to UV light. The 30 min values were 1.17 t 0.05 pmol/mL for peak 2A, 1.95 ? 0.07 pmoYmL for peak 2B, 1.63 ? 0.06/pmoYmL for peak 3A, 2.50 ? 0.10 pmoYmL for peak 3B, 2.95 ? 0.11 FmoYmL for peak 4 and 2.25 2 0.01 pmoYmL for peak 5. These values were obtained with so- lutions containing 50 pg/mL of the AGE compounds. Peaks 2A and 2B showed a sharper increase at 5 min after which singlet oxygen synthesis continued linearly for the remainder of the time. This was in marked contrast to the production of superoxide anion with these peaks. The other AGE pro- duced a rapid, linear increase in singlet oxygen throughout

Time of irradiation (min)

Figure 3. The UVA-dependent generation of superoxide anion by the AGE peaks isolated from a threose and N-Ac-Lys incubation. All assays were carried out in triplicate as described in the Materials and Methods with 50 pg/mL solutions of each AGE peak. Standard deviations for all values were 5% or less. The left axis shows the A,,, measured and the right axis shows the calculated superoxide produced. Control values are shown in Fig. 4.

the 30 min irradiation, generating 500-1000-fold more sin- glet oxygen than superoxide anion.

The rate of singlet oxygen formation for other purified AGE compounds and known singlet oxygen sensitizers were determined and are presented in Fig. 6. The structures of

Riboflavin

10 20 30 T h e of irradiation (min)

Figure 4. The UVA-dependent generation of superoxide anion by peak 3B (50 pg/mL) compared to the superoxide anion generated by tryptophan oxidation products (1 .O pg/mL) and riboflavin (0.5 pg/mL). The data are plotted as described in Fig. 3. KA, kynurenic acid; Kyn, kynurenine; 3-OH-Kyn, 3-hydroxy-o,~-kynurenine.

Page 4: The Relative UV Sensitizer Activity of Purified Advanced Glycation Endproducts

Photochemistry and Photobiology, 1997, 65(4) 669

0.25

0.2

0.15

5 2 8

2A

O i I I I I

5 10 15 20 25 30 Time of irradiation (mln)

Figure 5. The UVA-dependent generation of singlet oxygen by the AGE peaks isolated from a threose and N-Ac-Lys incubation. All assays were carried out in triplicate with and without sodium azide under the conditions described in the Materials and Methods with 50 pg/mL of each AGE peak. Standard deviations for all values were 5% or less. The left axis shows the decrease in A,, and the right axis shows the calculated nmoles of singlet oxygen generated (30). Control values are shown in Fig. 6.

pentosidine, LM-I and Ac-FTP are shown in Fig. 7 along with peak 3B (threosidine 3B). These compounds, as well as LM-1 were all active in singlet oxygen synthesis, but at 1/3 to 112 the rate of peak 3B. Peak 4 was shown to be identical to Ac-FTP by TLC, spectral properties and proton NMR spectroscopy (unpublished results). Peak 4, which was assayed immediately after isolation, was several-fold more active than the Ac-FTP sample, likely due to the observed instability of this molecule on storage (28). Peak 3B was equivalent to kynurenine as a sensitizer and only slightly less activ.: than 3-OH kynurenine. When compared to kynurenic acid and riboflavin; however, all the purified AGE com- pounds were at least two orders of magnitude less active, as only 0.5 pg/mL of kynurenic acid and riboflavin were ca- pable of producing 3.5-4.0 pmoYmL, of singlet oxygen dur- ing the 30 min irradiation. The AGE, therefore, are markedly less active than classic singlet oxygen sensitizers. No singlet oxygen was generated in the absence of sensitizer.

The relative efficiency of several sensitizers for singlet oxygen synthesis was calculated. The actual light absorbed by a 1.0 mg1mL solution of 3B (285475 nm) was deter- mined using the irradiation rig. This value was compared to similar measurements with 5-OH-kynurenine, kynurenic acid and riboflavin. Based upon these data, peak 3B pro- duced 0.67 pmol of singlet oxygen/J, 3-OH kynurenine pro- duced 0.33 pmoYJ, kynurenic acid produced 27 pmoYJ and riboflavin produced 80 pmol/J of absorbed light. This is in comparison to 0.03 pmol/J of absorbed light produced by

Time of irradiation (mln)

Figure 6. The UVA-dependent generation of singlet oxygen by peak 3B (50 pg/mL) compared to the singlet oxygen generated by known AGE (50 pg/mL), tryptophan oxidation products (50 pg/ mL), kynurenic acid (0.5 pg/mL) and riboflavin (0.5 pg/mL). All assays were carried out in duplicate and plotted as described in Fig. 5. KA, kynurenic acid; Kyn, kynurenine; 3-OH-Kyn. 3-hydroxy-~ ,~- kynurenine.

the irradiation of either ascorbate-glycated lens proteins (24) or aged human lens water-insoluble proteins (30).

DISCUSSION The incubation of lens proteins with sugars for several weeks produces modified proteins that closely resemble the proteins isolated from brunescent cataracts (13,1520). This has been demonstrated with a variety of sugars (13,32-34) and by the

Pentosidine LM-1

N-AC-Lp I

N-Ac-Lys

AC-FTP Threosidine 38 Figure 7. Structures of purified AGE compounds tested for singlet oxygen generation in Fig. 6. For details of structure analysis see data for pentosidine (26). LM-I (44). Ac-FTP (28) and threosidine 3B (31).

Page 5: The Relative UV Sensitizer Activity of Purified Advanced Glycation Endproducts

670 B. J. Ortwerth eta/.

oxidation products of ascorbic acid (19,35,36). The initial formation of the Amadori compound may have little effect on protein structure, but with time, complex aromatic struc- tures are produced, which are responsible for the browning, fluorescence and protein crosslinks seen in these proteins. In the lens this chemical damage can be amplified by the pro- duction of ROS in response to UVA irradiation of the gly- cated proteins (23,24,30).

Due in large part to the work from the laboratory of Vin- cent Monnier, several of the AGE present in cataract lenses have been isolated and their structures determined (26-28). These compounds, however, account for only a small frac- tion of the total AGE present. It was assumed that these AGE were responsible for the absorption of UVA light and may be the sensitizers responsible for the generation of the re- active oxygen species.

In this work we have, for the first time, quantified the formation of superoxide anion and singlet oxygen by puri- fied AGE molecules and a late-eluting mixture of AGE in response to UVA light. Every compound tested had sensi- tizer activity, displaying similar activity. The synthesis of superoxide anion was measurable but not significant. The irradiation with UVA light, which was 380 times greater than human lens exposure on a sunny day at noon (37), produced only 5-10 nmol of superoxide anion after 30 min. This was an order of magnitude less than that produced by irradiating ascorbate-glycated lens proteins (22) and only three-fold greater than the superoxide anion produced chem- ically by 10 mg of glycated RNase (6). Also, the rate of synthesis was no greater than the superoxide produced by the chemical degradation of synthetic Amadori compounds (7).

The rapid photodestruction of peaks 2A and 2B suggests that the superoxide anion was produced by a small amount of impurity in these preparations. The slow, linear generation of superoxide anion by the other AGE compounds argue for an unstable triplet state for these molecules. The generation of superoxide anion relative to light absorbed was 130 nmoYJ for riboflavin, 6.6 nmoYJ for kynurenic acid and 1.4 nmoYJ of light absorbed for peak 3B. This compares to 0.06 nmoYJ of light absorbed for either a human lens water-in- soluble fraction or ascorbate-glycated lens proteins (22,23). The UVA-absorbing molecules in these proteins are, there- fore, only 3 4 % as active as purified AGE, suggesting that much of the UV light absorbed by these proteins is absorbed by compounds that do not produce superoxide anions.

The level of superoxide anion formation produced by UVA illumination of AGE compounds must be judged to be insignificant for cataract formation because (1) the levels synthesized under ambiant solar exposure would be less than 10 pmolh, (2) most of the superoxide anion would dismutate to H202 at physiological pH (22). (3) even traces of super- oxide dismutase could convert all the superoxide to the less reactive H202 and (4) there is little, if any, damage to lens proteins by superoxide anion directly (22).

Due to their higher reactivity, kynurenic acid and N-for- my1 kynurenine, if present in lens proteins, could be more significant for superoxide anion synthesis. The major UV filter compound in human lens is 3-hydroxykynurenine gly- coside (38-40). 3-Hydroxylkynurenine had no sensitizer ac- tivity for superoxide anion generation, which is consistent

with the picosecond lifetime of the triplet state of this sen- sitizer (41,42).

The UVA irradiation of 2.0 mg/mL solutions of either a sonicated water-insoluble fraction from human lens or ascor- bate-glycated proteins produced greater than 2.0 pmoYmL of singlet oxygen over a 60 min irradiation (21,30). The purified AGE, tested here at 50 pmoYmL, generated 1.0-2.5 pmoYmL of singlet oxygen in comparable assays. Based upon light absorption, riboflavin was at least 20-fold and kynurenic acid 50-fold more active than AGE. Peak 3B gen- erated 0.67 pmol of singlet oxygen/J of absorbed light, which compares to 0.03 pmoYJ of absorbed light for either human lens water-insoluble proteins or ascorbate-glycated lens proteins. Because this AGE is a 20-fold more efficient sensitizer, it would have to represent only 5% of the light- absorbing compounds in these protein preparations to ac- count for the singlet oxygen and superoxide anion generated. Solutions containing 50 pg/mL of peak 3B (80 nmoYmL based on a molecular weight of 628) produced singlet oxy- gen equivalent to that produced (3 mg/mL) by human lens water-insoluble proteins (24 pmoYmL amino acid residues). Therefore, on a weight basis the human lens water-insoluble proteins need only contain 0.3% of its total amino acids or 6% of the total lysine residues as AGE with the activity of peak 3B to account for the singlet oxygen generated by UVA light. The ascorbate-glycated proteins exhibited a 20% loss of lysine residues (23), more than sufficient to accommodate 6% AGE formation. While it is likely that some tryptophan was oxidized during this 4 week glycation period, thereby producing sensitizers, ascorbate-glycated RNase A, which contains no tryptophan residues, exhibited more singlet ox- ygen-generating activity than ascorbate-glycated lens pro- teins (23).

The AGE tested here must be considered inefficient sen- sitizers. The quantum yield for peak 3B, whose purity has been confirmed, is <0.01 for singlet oxygen production when compared to the quantum yield for kynurenic acid (37). These AGE, however, are likely sufficient to cause the protein damage seen in brunescent cataract. This is especial- ly true because the singlet oxygen would be generated within the protein aggregates that make up the water-insoluble frac- tion. This would place the sensitizers in close proximity to the target amino acid side chains (His and Trp) in these proteins. The UV filter compound, 3-hydroxykynurenine, was half as active as peak 3B in singlet oxygen formation/J of absorbed light. This demonstration of singlet oxygen for- mation is consistent with the observations of Dillon e? al. (41), who showed a transient state formed by UV light ab- sorption (triplet state), which was rapidly quenched by ox- ygen. The 3-OH kynurenine glucoside, while proposed as a “filter molecule” to protect the retina against UVA irradi- ation, may be capable of UVA-dependent singlet oxygen generation even in young human lenses (40). One cannot predict the activity of the glycoside. however, because it has a 30-fold higher quantum yield for fluorescence than 3-hy- droxykynurenine (41). it may be considerably less active than 3-OH kynurenine at generating singlet oxygen from the triplet state.

Singlet oxygen is quantitatively the major ROS produced by UVA irradiation of AGE and human lens proteins. Sin- glet oxygen can directly cause the oxidation of His and Trp

Page 6: The Relative UV Sensitizer Activity of Purified Advanced Glycation Endproducts

Photochemistry and Photobiology, 1997, 65(4) 671

residues in protein aggregates, and both of these amino acids are dramatically decreased in the proteins isolated from ad- vanced brunescent lenses (43). These data represent the first attempts to quantify the ROS produced by purified AGE and support the proposal that AGE can play a significant role in protein oxidative damage during brunescent cataract forma- tion. Proof, however, will require the demonstration of com- mon AGE molecules in glycated and cataract proteins. Tryp- tophan oxidation products, and possibly riboflavin, could also make a significant contribution to the lens sensitizer activity, but the content o f these compounds in aged human lens proteins is not known.

Acknowledgrments-The authors express their appreciation to Dr. Vincent Monnier and coworkers for their pioneering efforts in the isolation and identification of AGE and to Mr. Paul Olescn for car- rying out the initial work to establish the UVA irradiation system. This work was supported in part by NIH grants EY 02035, EY 07070, EY 09912 and in part by Research to Prevent Blindness Inc.

REFERENCES I .

2.

3.

4.

5.

6.

7.

8.

9

10.

1 1 .

12.

13.

14.

Brownlee, M. (1995) Advanced protein glycosylation in diabe- tes and aging. Annu. Rev. Med. 46, 223-234. Sensi, M., F. Pricci, G. Pugliese, M. De Rossi, A. F. G. Petrucci, A. Cristina, S. Morano, G. Pozzessere, E. Valle, D. Andreani and U. Di Mario (1995) Role of advanced glycation end-prod- ucts (AGE) in late diabetic complications. Diabetes Res. Clin. P racr. 28, 9- 17. McCance, D. R., D. G. Dyer, J. A. Dunn, K. E. Ballie, S. R. Thorpe, J. W. Baynes and T. J. Lyons (1993) Maillard reaction products and their relation to complications in insulin-dependent diabetes mellitus. J . Clin. Invest. 91. 2470-2478. Arai, K., S. Iizuka, Y. Tada, K. Oikawa and N. Taniguchi (1987) Increase in the glucosylated form of erythrocyte Cu-Zn- superoxide dismutase in diabetes and close association of the nonenzymatic glucosylation with the enzyme activity. Biochim.

Baldwin, J. S., L. Lee, T. K. Leung, A. Muruganandam and B. Mutus (1995) Identification of the site of non-enzymatic gly- cation of glutathione peroxidase: rationalization of the glyca- tion-related catalytic alterations on the basis of three-dimen- sional protein structure. Biochim. Biophys. Acra 1247, 60-64. Mullarkey, C. I., D. Edelstein and M. Brownlee (1990) Free radical generation by early glycation products: a mechanism for accelerated atherogenesis in diabetes. Biochem. Biophys. Res. Commun. 173, 932-939. Zu, J., J. Morita, S. Nishikawa and N. Kashimura (1996) Gen- eration of active oxygen species by some low-molecular weight Amadori rearrangement products under physiological condi- tions. Carbohydr. Len. 1, 457464. Kristal, B. S. and B. P. Yu (1992) An emerging hypothesis: synergistic induction of aging by free radicals and Maillard re- actions. J. Cerontol. 47, B 107-B 1 14. Hunt, J. V. and S. P. Wolf (1991) Oxidative glycation and free radical production: a causal mechanism of diabetic complica- tions. Free Radical Res. Commun. 12-13, 115-123. Lyons, T. J. (1993) Glycation and oxidation: a role in the patho- genesis of atherosclerosis. Am. J. Cardiol. 71, 268-31B. Low, P. A. and K. K. Nickander (1991) Oxygen free radical effects in sciatic nerve in experimental diabetes. Diabetes 40, 873-877. Smith, M. A,, P. G. Richey, S. Taneda, R. K. Kutty, L. M. Sayre, V. M. Monnier and G. Perry (1994) Advanced Maillard reaction end products, free radicals, and protein oxidation in Alzheimer’s discase. Ann. N. Y. Acad. Sci. 738, 447-454. Stevens, V. J.. C. A. Rouzer, V. M. Monnier and A. Cerami (1978) Diabetic cataract formation: potential role of glycosyla- tion of lens crystallins. Proc. Narl. Acad. Sci. USA 75, 2918- 2922. Oimomi, M., Y. Maeda, F. Hata, Y. Kitamura, S. Matsumoto,

Biophys. A C I ~ 924, 292-296.

15

16

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27

28

29.

30.

31.

32.

33.

34.

35.

S. Baba, T. Iga and M. Yamamoto (1988) Glycation of catar- actous lens in non-diabetic senile subjects and in diabetic pa- tients. Exp. E,ye Res. 46, 415-420. Ortwerth, B. J. and P. R. Olesen (1988) Ascorbic acid-induced crosslinking of lens proteins: evidence supporting a Maillard reaction. Biochim. Biophys. Acra 956, 10-22. Ortwerth, B. J. and P. R. Olesen (1988) Glutathione inhibits the glycation and crosslinking of lens proteins by ascorbic acid. Exp. Eye Res. 47, 737-750. Wells-Knecht, K. J., D. V. Zyzak, J. E. Litchfield, S. R. Thorpe and J. W. Baynes (1995) Mechanism of autoxidative glycosy- lation: identification of glyoxal and arabinose as intermediates in the autoxidative modification of proteins by glucose. Bio- chemistry 34, 3702-3709. Zyzak, D. V., J. M. Richardon, S. R. Thorpe and J. W. Baynes (1995) Formation of reactive intermediates from Amadori com- pounds under physiological conditions. Arch. Biochem. Biophys.

Ortwerth, B. J., J. A. Speaker, M. Prabhakaram, M. Lopez, E. Y. Li and M. S. Feather (1994) Ascorbic acid glycation: the reaction of L-threose in lens tissue. Exp. Eye Res. 58. 665-674. Kamei, A. and M. Kato (1991) Contribution of glycation to human lens coloration. Chem. Pharm. Bull. 39, 1272-1276. Linetsky, M. and B. J. Ortwerth (1995) The generation of hy- drogen peroxide by the UVA irradiation of human lens proteins. Photochem. Photobiol. 62. 87-93. Linetsky, M., H.-L. James and B. J. Ortwerth (1996) The gen- eration of superoxide anion by the UVA irradiation of human lens proteins. Exp. Eye Res. 63, 67-74. Ortwerth, B. J., M. Linetsky and P. R. Olesen (1995) Ascorbic acid glycation of lens proteins produces UVA sensitizers similar to those in human lens. Phorochem. Photobiol. 62, 454-462. Linetksy. M. and B. J. Ortwerth (1996) Quantitation of the re- active oxygen species generated by the UVA irradiation of as- corbic acid-glycated lens proteins. Photochem. Phorobiol. 63, 649-655. Perlin, A. S. (1962) D- and L-threose: glycol cleavage of 1,3-0- benzylidine-D-(or L-)arabinitol. In Methods in Carbohydraie Chemistry, Vol. 1 (Edited by R. L. Wistler and M. L. Wolfrom), pp. 68-70. Academic Press, New York. Sell, D. R. and V. M. Monnier (1989) Structure elucidation of a senescence cross-link from human extracellular matrix. J. B i d . Chem. 264, 2 1597-2 1602. Nagaraj, R. H. and V. M. Monnier (1992) Isolation and char- acterization of a blue fluorophore from human eye lens crystal- lins: in vitro formation from Maillard reaction with ascorbate and ribose. Biochim. Biophys. Acta 1116, 34-42. Nagaraj, R. H. and V. M. Monnier (1995) Protein modification by the degradation products of ascorbate: formation of a novel pyrrole from the Maillard reaction of L-threose with proteins. Biochim. Biophys. Acra 1253, 75-84. Kraljic, I. and S. El Moshni (1978) A new method for the de- tection of singlet oxygen in aqueous solution. Phorochern. Pho- robiol. 28, 577-581. Linetsky, M. and B. J. Ortwerth (1997) Quantitation of the sin- glet oxygen produced by the UVA irradiation of human lens proteins. Phofochem. Phorobiol. (In press) Prabhakaram, M., Q. Cheng, M. S. Feather and B. J. Ortwerth (1997) Structural elucidation of a novel lysine-lysine crosslink generated in a glycation reaction with L-threose. Amino Acids. (In press) Swamy, M. S., C. Tsai, A. Abraham and E. C. Abraham (1993) Glycation mediated lens crystallin aggregation and cross-linking by various sugars and sugar phosphates in virro. Exp. Eye Res. 56, 177-185. Lal, S., B. S. Szwergold, A. H. Taylor, W. C. Randall, R. Kap- pler, K. Wells-Knecht, J. W. Baynes and T. R. Brown (1995) Metabolism of fructose-3-phosphate in the diabetic rat lens. Arch. Biochem. Biophys. 318, 191-199. Kamei, A. (1992) Glycation and insolubility of human lens pro- tein. Chem. Pharm. Bull. 40, 2787-279 1. Ortwerth, B. J., M. S. Feather and P. R. Olesen (1988) The precipitation and cross-linking of lens crystallins by ascorbic acid. Exp. Eye Res. 47. 155-168.

316, 547-554.

Page 7: The Relative UV Sensitizer Activity of Purified Advanced Glycation Endproducts

672 B. J. Ortwerth eta/.

36. Slight, S. H., M. S. Feather and B. J. Ortwerth (1990) Glycation of lens proteins by the oxidation products of ascorbic acid. Biochim. Biophys. Acfa 1038, 367-374.

37. Sliney, D. H. (1987) Estimating the solar ultraviolet radiation exposure to an intraocular lens implant. J . Curaract Refrucr. Surg. 13, 296-301.

38. van Heyningen, R. (1971) Fluorescent glucoside in the human lens. Narure 230, 393-394.

39. Dillon, J. (1984) Photolytic changes in lens proteins. Curr. Eye Res. 3, 145-150.

40. Wood, A. M. and R. J. W. Truscott (1994) Ultraviolet filter compounds in human lenses: 3-hydroxykynurenine glucoside formation. Vision Res. 34, 1369-1374.

41. Dillon, J., R.-H. Wang and S. J. Atherton (1990) Photochemical and photophysical studies on human lens constituents. Phoro- chem. Phorobiol. 52, 849-854.

42. Dillon, J. and S. J. Atherton (1990) Time resolved spectroscopic studies on the intact human lens. Phorochem. Phorobiol. 51. 465-468.

43. Kamei, A. (1993) Glutathione levels of the human crystalline lens in aging and its antioxidant effect against the oxidation of lens proteins. Biol. Pharm. Bull. 16, 87G875.

44. Graham, L.. V. M. Monnier and R. H. Nagaraj (1996) Structure elucidation of an ascorbatelribose derived lys-lys crosslink pres- ent in the human lens. Invest. Ophrhalmol. & Visual Sci. 37, S886.