Tryptophan Metabolites from Young Human Lenses and the Photooxidation of Ascorbic Acid by UVA Light

Abstract

purpose. To determine whether there are UVA light-responsive sensitizers in young human lenses capable of initiating the oxidation of ascorbic acid in the absence of oxygen.

methods. Lens homogenates were fractionated, and low-molecular-weight (LMW) components were separated from the proteins by filtration through a 3000-MWt cutoff filter. Aliquots of each fraction were assayed for sensitizer activity by UVA irradiation (337-nm cutoff filter) with 0.1 mM ascorbic acid, measuring ascorbate oxidation by loss of absorbance at 265 nm. Two major peaks were isolated from a human lens water-soluble (WS)-LMW fraction on a reversed-phase column and were identified by mass spectrometry.

results. All human lens fractions oxidized ascorbate when irradiated by UVA light. Most of the sensitizer activity in young human lenses was in the LMW fractions. An action spectrum for the WS-LMW fraction from human lens showed activity throughout the UVA region. Assays with and without oxygen showed little or no difference in ascorbate oxidized, arguing for a direct transfer of an electron in a so-called type 1 reaction. A human lens WS-LMW fraction contained two major peaks of activity. The greater peak was identified as 3-hydroxykynurenine glucoside (3OHKG) by mass spectrometry and its absorption spectrum, whereas the lesser peak was identified as 4-(2-amino-3-hydroxyphenyl)-4-oxobutanoic acid glucoside (AHBG). The activities were 1.1 and 2.8 nmol of ascorbate oxidized in 30 minutes/nmol 3OHKG and AHBG, respectively.

conclusions. The filter compounds present in human lenses can absorb UVA light and cause the oxidation of ascorbic acid in the presence and absence of oxygen, possibly initiating the glycation of lens proteins.

Ultraviolet radiation from the sun is routinely identified in epidemiologic studies as a risk factor for cortical cataracts. ^{1 2 3 4 5} Similarly, subjecting rodents to UV light leads to the formation of cortical cataracts. ^{6 7 8} These studies show that UVB light (260–320 nm) is considerably more effective than UVA light (320–400 nm) in this regard. ^{8 9 10} Irradiation of rabbit or rat lenses in vitro caused the formation of cataracts in the anterior subcapsular region. ^{11 12} Consistent with this observation are the data of Löfgren and Söderberg, ¹³ who demonstrated that UVB light only penetrates the rat lens to a depth of 0.45 mm. Presumably, this is because UVB light is completely absorbed by the aromatic amino acid residues in the lens proteins. Because most age-related human cataracts exhibit extensive nuclear involvement, it is possible that UVA light may also be a significant contributor to human cataract formation. ^{14 15} UVA light penetrates the human lens into the nucleus, ¹⁶ and the water-insoluble (WI) fraction from human lens has been shown to be capable of absorbing UVA light and generating reactive oxygen species under aerobic conditions. ^{17 18 19 20}  

Although several reactive oxygen species (ROS) are produced by UVA irradiation of human lens proteins in vitro, the major product is singlet oxygen. ^{17 20} Significantly, irradiation of human lens proteins with UVA light in the absence of oxygen in vitro results in the oxidation of ascorbic acid to dehydroascorbic acid, presumably by way of an ascorbyl free radical intermediate. ²¹ This photochemistry likely resulted from the presence of advanced glycation end products (AGEs), ^{22 23 24 25} which can be formed on lens proteins by ascorbate oxidation products. Although it is known that the irradiation of human lenses with UVA light or sunlight causes the oxidation of ascorbate, ^{21 25} it is unclear how ascorbate can be oxidized in young human lenses before AGE formation. In the work presented here, we examine whether UVA sensitizers capable of oxidizing ascorbic acid are present in young human lenses and, if so, try to determine their identity and origin. 

Materials and Methods

Human lenses were obtained from the Heartland Lions Eye Tissue bank after removal of the cornea. Cataract lenses were obtained as a kind gift from the Patney Eye Clinic (Rajkot, India). Procedures adhered to the provisions of the Declaration of Helsinki for research on human tissue. Lenses were processed directly or were stored frozen at −70°C until use. 

Ascorbic acid (99% pure) was purchased from Acros Organics (Geel, Belgium) as the sodium salt. All solutions were made fresh for every experiment. Tryptophan metabolites (anthranilic acid, xanthurenic acid, kynurenine, 3-OH-kynurenine, quinolinic acid, and 5-OH tryptophan) were purchased from Sigma Chemical (St. Louis, MO). All other chemicals were of analytical grade. 

Preparation of Lens Fractions

Decapsulated human lenses were homogenized in deionized (DI) water (1 mL/lens) with a handheld Dounce homogenizer and were centrifuged at 27,000g for 20 minutes. The supernatant was decanted, and the protein pellet was resuspended with 1 mL DI water and centrifuged. Combined supernatants were placed in a centrifugal filter device (Centriplus YM3; Millipore, Billerica, MA) with a 3000-MWt cutoff and were centrifuged at 3000g until the filtration was almost complete. The filtrate was collected, 1 mL water was added to the small amount of unfiltered solution, and the protein on the filter was resuspended and centrifuged again. The time of centrifugation varied with the volume and protein concentration of each sample. This procedure was repeated once more to remove all the low-molecular-weight (LMW) components. The filtrates were pooled, lyophilized, and dissolved in 1 mL water, and the final protein on the filter was dissolved by sequential extraction with aliquots of 1 mL water, which were combined. The washed WI pellet was resuspended and sonicated for 9 minutes in the pulsed mode to prepare the water-insoluble sonication supernatant (WISS), as described previously. ²⁶ This sonication completely dissolved the protein pellet. The solubilized pellet was centrifuged through a filter (Centriplus YM3; Millipore) to separate the LMW and protein fractions, as described. The protein content of each fraction was measured by the BCA method, as described by the manufacturer (Pierce, Rockford, IL), and the total protein represented the sum of the water-soluble (WS) and WISS values. 

Ascorbate Oxidation Assays

Assay mixtures were prepared containing 0.1 mM ascorbic acid and 1 mM DTPA in 50 mM CHELEX-treated phosphate buffer (pH 7). Aliquots of each lens fraction, ranging from 10 to 100 μL, were added to the assay solution in a final volume of 1 mL and were irradiated with UVA light for 30 minutes. At 10-minute intervals, the cuvette was removed, and absorbance at 265 nm was measured to follow the loss of ascorbic acid. Irradiation was carried out with the light from a 1000-W Hg/Xe lamp (Oriel Corp., Stratford, CT) filtered through a 5% CuSO₄ solution and a 337-nm cutoff filter. The light measured was 170 mW/cm² at the cuvette surface. Irradiation was routinely carried out in the presence of air; however, anaerobic assays were also conducted after oxygen removal by blowing argon into the screw-capped cuvette for 15 minutes. Commercially obtained compounds were assayed as described at 10-μM concentrations in the cuvette or levels of 0.02 to 0.1 μM in the case of riboflavin. Reactions with purified compounds were also assayed by measuring the ascorbate peak at the λ_max of 243 nm (in m-phosphoric acid) during an isocratic elution of a carbohydrate column (in the sodium form; Resex-RNM; Phenomenex, Torrance, CA) with a 1% m-phosphoric acid and 0.1 mM DTPA solution. ²¹  

Action Spectrum

Standard assay solutions for ascorbate oxidation measurements, with constant amounts of a WS-LMW fraction from human lens, were irradiated in a spectrofluorometer (model 2500; Hitachi, Yokohama, Japan) set at specific wavelengths in the UVA region ± 5 nm. Assay mixtures of 1 mL were irradiated over 2 hours, and at various intervals the cuvette was removed and the absorbance was measured at 265 nm. The activity at each wavelength was expressed as nanomoles of ascorbate oxidized per 2 hours of irradiation. 

Isolation, Identification, and Assay of the Components of the WS-LMW Fraction from Human Lens

The WS-LMW fraction was isolated from pooled human lenses and subjected to high-performance liquid chromatography (HPLC) on a 10 × 250-mm column (Prodigy; Phenomenex, Torrance, CA) using a HPLC system (Shimadzu, Kyoto, Japan) with a diode array detector. Elution was carried out using the following conditions: solvent A (0.1% vol/vol trifluoroacetic acid [TFA] in water), solvent B (0.1% vol/vol TFA in acetonitrile), at a flow rate 0.8 mL/min; 0% to 10% solvent B over 20 minutes, 10% to 100% solvent B over the next 25 minutes, 100% solvent B for 5 minutes, and the re-equilibration of the column in solvent A for 15 minutes. Spectra were continuously acquired, and the elution profile at 370 nm was displayed to show the elution profile of the UVA-absorbing species. The absorption spectrum of each of the major peaks was determined, and the pooled peaks were concentrated under reduced pressure at room temperature, lyophilized, and redissolved in water. The isolated sensitizers S1 (retention time, 17.5 minutes) and S2 (retention time, 30 minutes) were further analyzed by electrospray ionization (ESI) and ESI-MS/MS mass spectrometry with a 0 triple-quadrupole mass spectrometer (TSQ700; Thermo Finnigan, Waltham, MA). Concentrations of S1 and S2 were determined based on the extinction coefficients at their absorption maximum. ^{27 28 29} Both S1 and S2 were assayed for their sensitizer activity at increasing levels by the UVA-dependent decrease in ascorbate absorbance at 265 nm. 

Effect of Reduced Glutathione (GSH) Treatment on the Loss of Ascorbate-Oxidizing Sensitizers Bound to Lens Proteins

The WISS fraction from various human lenses was concentrated to 100 μL using a centrifuge filter (Ultrafree-MC5000; Millipore). This solution was brought to 50 mM phosphate buffer, 1.6 M guanidine HCl, and 5 mg/mL GSH in a total volume of 1 mL at a final pH of 7.2 to remove any protein-bound filter compounds, as described by Korlimbinus et al.. ³⁰ After incubation for 7 hours at 37°C, each reaction mixture was dialyzed against 50 mM phosphate buffer, pH 7, for 24 hours with three buffer changes. DTPA was added to 1 mM, and each WISS fraction was assayed for UVA-mediated ascorbate oxidation. Control extractions were carried out with aliquots of the same WISS fraction, as described, except that the addition of GSH was omitted. 

Results

Ascorbate Oxidation by UVA Irradiation

An assay system was established to measure the oxidation of ascorbic acid by UVA light sensitizers in various lens fractions and by purified compounds known to absorb UVA light. The loss of ascorbate was measured by the decrease in absorbance at 265 nm with time of irradiation. The rate of ascorbate oxidation over 30 minutes is shown in Figure 1Ain the presence of increasing levels of the known sensitizer, riboflavin. A linear increase in ascorbate loss was observed up to 0.1 μM riboflavin, indicating the range of activity that can be validly measured in this assy. Riboflavin at 0.1 μM oxidized 60 nmol ascorbate in 30 minutes or 600 nmol ascorbate oxidized/30 minutes/nmol riboflavin in the assay cuvette. The linear response with time showed that riboflavin was not photolyzed during the assay. Ascorbate oxidation was similarly measured for 50 to 150 μL undiluted WS-LMW fraction from human lens and is shown in Figure 1B . This fraction displayed a roughly linear increase in ascorbate oxidized with increasing WS-LMW fraction; however, 10 μL WS-LMW fraction oxidized only 6 nmol ascorbate in 30 minutes or a total of 600 nmol ascorbate oxidized/lens. The sensitizer activity in this fraction/lens, therefore, was equivalent to only 1 nmol riboflavin. 

The linear response also argues that there was no photolysis of the lens sensitizers during UVA irradiation. To confirm this, we assayed a WS-LMW fraction before and after 30 minutes of irradiation in the presence of ascorbate. The sensitizer activity was not decreased by the earlier irradiation; however, sensitizer activity was decreased if the LMW fraction was irradiated in the absence of ascorbic acid (data not shown). To show that the decrease in A₂₆₅ was a valid measure of ascorbate oxidation, a difference spectrum was determined comparing spectra before and after irradiation. The difference spectrum was identical with the absorption spectrum of pure ascorbic acid, showing that ascorbate was the only compound oxidized by the UVA irradiation (data not shown). To confirm that ascorbate was oxidized, the ascorbate peak was measured by HPLC chromatography on a carbohydrate column (in the sodium form; Resex-RNM; Phenomenex). ²¹ The loss of the ascorbate peak corresponded to the loss of ascorbate measured spectrophotometrically (data not shown). Irradiation in the absence of added sensitizer produced only 0.2 nmol ascorbate oxidation in 30 minutes, and no ascorbate oxidation was seen in a dark control with or without sensitizer (Fig. 1B) . The possible presence of GSH in the WS-LMW fraction was a concern because it would reduce any dehydroascorbate (DHA) formed back to ascorbate. This would have resulted in a lag in apparent ascorbate oxidation at early times but was not observed, possibly because of the oxidation of any GSH in the LMW fraction to GSSG during the sample preparation in air. 

Assays of UVA-Dependent Ascorbate-Oxidizing Activity in Human Lens Fractions

The various fractions isolated from human lens were assayed for UVA sensitizer activity capable of oxidizing ascorbic acid. Aliquot volumes, equivalent to 2% to 20% of the total, were assayed for each fraction. Several extracts from young and aged human lenses were each assayed in duplicate, and the results were calculated as nanomole ascorbate oxidized per lens. Results of these assays are presented in Figure 2and show that there was significant activity in the WS-LMW fractions from young human lenses and in the WS protein fraction. LMW fractions were unchanged in the aged lenses, but the WISS protein fraction increased roughly 3- to 6-fold, which was greater than the increase in protein in this fraction. Indian cataract lenses (type 1 or 2, as described by Pirie ³¹ ) had a distribution of ascorbate-oxidizing activity similar to the activities in aged normal lens fractions (data not shown). Standard deviations were great in the aged human lens. Figure 3shows that this resulted from a roughly linear increase in the UVA sensitizer activity in the WS and WISS protein fractions with age. 

Action Spectrum of the WS-LMW Fraction

To confirm that UVA light was indeed capable of causing the oxidation of ascorbate, an action spectrum was determined for ascorbate oxidation by the WS-LMW fraction. This spectrum (Fig. 4)shows definite sensitizer activity from 320 to 380 nm, which represents UVA light wavelengths. The data suggest that the sensitizer activity was considerably higher in the UVB region, but light below 300 nm does not penetrate to the lens. ^{13 32}  

Effect of Oxygen on UVA Light-Dependent Ascorbate Oxidation

The assays shown in Figures 2 and 3were carried out in the presence of air, which provides approximately 200 nmol oxygen/mL assay mixture. It was possible that the activity measured in these assays in vitro was not relevant to conditions in vivo; several reports argue that lens tissue contains low oxygen levels. ^{33 34 35} Therefore, we repeated the measurements with the various human lens fractions in the presence and in the absence of oxygen. These comparisons are shown in Figure 5 . The activity measured without oxygen was almost identical with that seen with oxygen for every lens fraction. Where there was a slight difference, the activity in the absence of oxygen was greater. These data are consistent with earlier observations that showed ascorbic acid can compete favorably with oxygen for the triplet state of the lens sensitizers ²¹ and that oxygen was not required for the oxidation of ascorbate with the LMW or the protein-bound sensitizers from human lens. 

Fractionation and Identification of the WS-LMW Sensitizers in Human Lens

To identify the nature of the UVA sensitizers in the WS-LMW fraction from human lens, this fraction was separated on an HPLC column (Prodigy; Phenomenex). Elution was monitored by a diode array detector with the elution profile at 370 nm shown in Figure 6A . Two major compounds in this fraction (labeled S1 and S2) were observed. Absorption spectra of each of these compounds (Figs. 6B 6C)suggest they are both derivatives of kynurenine. Both compounds were isolated and subjected to ESI-MS (Figs. 7A 7B)and to ESI-MS/MS (Figs. 7C 7D) . The mass spectrum of S1 indicated a mass of 387 for the parent molecule with fragments of mass 225 and 208 by ESI-MS/MS. These spectra were identical with those reported for 3-hydroxykynurenine glucoside, ²⁸ with the 225 fragment representing the loss of glucose (162 amu). The mass spectrum of S2 showed a mass of 372 with a major fragment of 210. These spectra correspond to 4-(2-amino-3-hydroxyphenyl)-4-oxobutanoic acid (AHBG), ²⁷ with the fragment of 210 representing the loss of glucose (162 amu). 

S1 and S2 were each isolated from several pooled human lenses and assayed for ascorbate-oxidizing activity over a 2-hour period. These data are presented in Figure 8 . Both compounds gave a linear response over time. Quantification gave values of 18 nmol ascorbate oxidized per hour for an 8-μM solution of S1 and 52 nmol ascorbate oxidized per hour for an 8-μM solution of S2 during the 30-minute photolysis. This corresponded to 1.1 nmol ascorbate oxidized per 30 minutes per nanomole S1 and 2.8 nmol ascorbate oxidized per 30 minutes per nanomole S2. Each human lens sensitizer, therefore, was several hundred times less active than riboflavin on a molar basis. Several minor peaks were observed in the human lens WS-LMW fraction. These peaks were not assayed because, even if their activity were equivalent to that of riboflavin, they would not have made a significant contribution to the total activity. A mixture of S1 and S2 equivalent to that in the WS-LMW fraction from human lens showed the same activity as the crude WS-LMW fraction, suggesting the absence of other sensitizers. In addition, the absorption spectra of these compounds were roughly similar to the action spectrum determined for the WS-LMW fraction, as shown in Figure 4 . 

The effect of increasing levels of S1 and S2 on ascorbate oxidation are shown in Figure 9Acompared with the increasing riboflavin shown in Figure 9B . Molar concentrations of S1 and S2 were determined using the molar extinction coefficient at their respective absorption maxima. ^{27 28}  

Comparison to the Activity of Similar Compounds

Several compounds similar to S1 and S2 have been reported in lens tissue. Purified preparations of compounds similar to S1 and S2 were purchased and assayed for ascorbate oxidizing activity in response to UVA light. The activity of these compounds is shown in Figure 10 , assayed in the presence and absence of air. Anthranylic acid and 3-OH kynurenine had significant activity at concentrations of 10 μM, but their activity was reduced in the absence of air, suggesting that oxygen reacts preferentially with these sensitizers compared with ascorbate. The presence of oxygen had little effect on the sensitizer activity of S1 and S2, similar to the data obtained with the crude WS-LMW fraction. The other compounds tested had higher activity in the absence of oxygen. 

Effect of GSH Treatment on the Activity of the WISS Protein Fraction from Various Human Lenses

The increased activity in the protein fractions from older human lenses could be attributed in part to advanced glycation end products ^{22 36 37} and in part to the incorporation of S1 and S2 into protein. ^{38 39 40} It has recently been shown that protein-bound filter compounds can be released by treatment with high levels of GSH. ³⁰ Table 1shows the effects of such GSH treatments on several protein fractions. Calf lens proteins, which were glycated by ascorbic acid for 4 weeks, showed no loss of ascorbate-oxidizing activity by treatment with GSH. Aged human lens proteins showed variable results. Three WISS protein fractions showed little or no loss of activity, whereas two others showed almost a 50% loss of activity. Both reactions, therefore, likely contribute to the sensitizer activity of the human lens WISS protein. The activity remaining after the GSH treatment, however, was significant and increased with age, consistent with AGE formation. 

Discussion

UVB light has been implicated in human cortical cataract formation. In addition, UVB light is the only light significantly able to cause cortical cataracts in a rat model, ^{6 11} possibly because while UVB enters rat lenses, it penetrates only a short distance into the lens, ¹³ resulting in anterior cortical cataract. ^{7 11 12} Nuclear involvement, however, is observed in most human cataracts. It is possible, therefore, that UVA light could contribute to nuclear cataract because UVA light has been shown to penetrate the nucleus of the lens. ¹⁶ Rodent lenses do not absorb significant amounts of UVA light. Human lenses, however, contain proteins with UVA-absorbing chromophores and free LMW filter compounds. The latter absorb UVA light but are thought to be incapable of producing reactive oxygen species. ²⁹ They are thought to act mainly to protect the lens and retina from UVA light in primates ³² but are not needed in rodents, which are nocturnal. 

Proteins isolated from young human lenses absorb little or no UVA light, but UVA absorbance increases with age in a process commonly referred to as browning. This is thought to be the result of protein modification and is prominent in the WI protein fraction. ³¹ The modification of proteins by sugars or ascorbate oxidation products give rise to AGEs, many of which absorb in the UVA portion of the spectrum. ^{22 23} The modification of calf lens proteins by ascorbic acid in vitro under air for 4 weeks produced AGEs that have been shown to be largely the same as the browning compounds in aged human lenses by thin-layer chromatography, ³⁶ HPLC, ³⁶ and mass spectrometry. ³⁷ In addition, ascorbate-glycated proteins exhibit sensitizer activity, producing the same distribution of ROS in air as seen with proteins isolated from aged human lenses and cataracts. ²² A concern from these observations was that ascorbate must be oxidized to produce glycating species, ⁴¹ but this seemed unlikely in vivo because of the limiting oxygen levels in the human lens. ^{33 34 35} Previous work from this and other laboratories has shown, however, that ascorbate can be oxidized by UVA light in the presence of aged human lens WI proteins but in the absence of oxygen, presumably through the formation of ascorbyl free radicals. ^{21 24} Two ascorbyl free radicals can then dismutate into one molecule of ascorbate and one molecule of dehydroascorbate (Fig. 11) . In the absence of AGE compounds in the young lens, however, it was unclear how the oxidation of ascorbic acid could be initiated. 

In this study, we investigated whether LMW compounds in young human lenses could act as UVA sensitizers and initiate the oxidation of ascorbate independent of AGE formation. LMW activity was seen in the WS and the WISS fractions. This activity did not change in older human lenses, but there was a threefold to sixfold increase in the activity in the protein fractions, consistent with AGE formation by ascorbate oxidation products. Significantly, all the fractions from human lenses demonstrated equivalent activities in the presence and the absence of oxygen. These data support the lack of ROS formation but argue that the sensitizers would be functional for ascorbate oxidation in the human lens in vivo. We have previously shown that irradiation of human lenses did not produce singlet oxygen damage but did cause ascorbate oxidation, ²¹ which is consistent with the lack of an oxygen effect reported here. 

A sum of the sensitizer activity in young human lens shows the ability to oxidize more than 300 nmol ascorbate in 30 minutes; however, the light source produced 100-fold more UVA light absorbed than the lens would be subjected to in vivo. In the young human lens, the DHA produced could be reduced back to ascorbate by GSH. Still, DHA levels measured in the human lens with age were significant and constant at 1.9 μmol/100 g lens tissue. ⁴² As the lens ages further, the protein-bound sensitizers increase, which produces increased ascorbate oxidation, more AGE formation, and therefore more sensitizers. This leads to a marked increase in protein-bound sensitizers in lenses of persons older than 40 to 50 years. ^{43 44} Ascorbate-modified proteins ultimately lead to the formation of high molecular–weight protein aggregates, ⁴⁵ which compose the WI fraction in human lens homogenates. It is this fraction that has the highest UVA sensitizer activity in aged human lenses. 

The LMW sensitizer activity in the WS fraction from young human lenses was independent of AGE formation but was attributed to two products of tryptophan metabolism, both of which were glucosylated. 3OHKG was first isolated and identified by van Heyningen, ⁴⁶ and S1 and S2 were confirmed as major components of human lens by Bova et al. ⁴⁷ and by Inoue and Satoh ²⁸ for AHBG, which they named deamino-3-OH-kynurenine-O-β-D-glucoside. Both isolated compounds had sensitizer activity for ascorbate oxidation, which was surprising considering that these compounds have been identified as absorbing UV light but are unable to undergo photochemistry. ⁴⁸ In this way they act to protect the retina from UV damage and have been described as filter compounds. ²⁹ The oxidation of ascorbate likely occurs by the transfer of an electron from the triplet state anion of these compounds forming ascorbyl radical, which dismutates to ascorbate and DHA. ²¹ The increase in activity in the protein fractions may be attributed, in part, to the incorporation of 3OHKG into the Cys residues in lens proteins. ^{38 39 40} The WISS-LMW fraction may represent compounds bound to the WI proteins, which are released when the proteins are disaggregated by sonication. HPLC separation of this fraction showed only trace amounts of 3OHKG and AHBG (data not shown), so other LMW compounds must be responsible for the sensitizer activity in this fraction. 

Although aged human lens proteins contain significant sensitizer activity for singlet oxygen formation and singlet oxygen readily oxidizes ascorbic acid, ⁴⁹ the presence of oxygen did not increase the UVA-mediated oxidation of ascorbate with any of the human lens fractions. This argues that ascorbate reacts with the lens sensitizers at a much greater rate than oxygen and agrees with earlier studies showing that the irradiation of whole human lenses caused a UVA-dependent loss of ascorbic acid though no loss of His or Trp because of singlet oxygen was observed. ⁵⁰ The activity in the protein fraction could have been attributed to 3OHKG bound to Cys residues in the protein, ^{30 38} or to AGEs formed by glycation. ^{22 24} The amount of bound 3OHKG has been measured, ³² and aged human lens has enough to account for approximately 100 nmol ascorbate oxidized per lens provided the bound 3OHKG has the same ascorbate-oxidizing activity as the free compound. The data in Table 1suggest that bound filter compounds account for only half, or less, of the total ascorbate-oxidizing activity in the WI protein fraction. 

The data presented here argue that LMW sensitizers in the lens can act early in life to initiate ascorbate oxidation in response to UVA light causing the formation of products that can glycate lens proteins. ²¹ As the formation of AGEs increases, and free 3OHKG binds to lens proteins, the protein fraction displays increasing UVA sensitizer activity, stimulating more ascorbate oxidation and AGE formation, leading to aggregate formation and possibly to nuclear cataract formation. Because of the decline of GSH levels with age, the ability to reduce DHA back to ascorbate decreases, thereby increasing the levels of glycation-active ascorbate degradation products. 

 Supported in part by National Eye Institute Grant EY02035 and a departmental grant from Research to Prevent Blindness Inc.

 Submitted for publication September 22, 2008; revised December 29, 2008; accepted April 29, 2008.

 Disclosure: B.J. Ortwerth, None; J. Bhattacharyya, None; E. Shipova, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Effect of increasing sensitizer on the extent of ascorbate oxidation by UVA light. (A) Oxidation of ascorbate over 30 minutes in 1-mL reactions containing 0.02 to 0.10 μM riboflavin, a known UVA sensitizer. (B) Oxidation of ascorbate over 30 minutes in 1-mL reactions containing 50 to 150 μL of a WS-LMW fraction from human lens and reactions without lens fractions and the dark control with 50 μL WS-LMW fraction added.

Figure 2.

 

Oxidation of ascorbate by the LMW and protein fractions from the WS and WISS fractions from young (10–23 years) and old (55–74 years) human lenses. Four lenses at each of three ages were processed separately and assayed in duplicate. The activity represents total nanomole ascorbate oxidized per lens fraction after 30 minutes of irradiation with UVA light.

Figure 3.

 

Effect of age on the UVA-dependent ascorbate-oxidizing activity of human lens WS and WISS protein fractions after 30 minutes of UVA irradiation.

Figure 4.

 

Action spectrum of the human lens WS-LMW fraction for ascorbate oxidation.

Figure 5.

 

UVA-dependent ascorbate oxidation of human lens fractions assayed in the presence and absence of oxygen after 30 minutes of UVA irradiation.

Figure 6.

 

Isolation of the LMW sensitizers in the WS fraction from human lens. (A) HPLC profile of 370 nm absorbing chromophores in the WS-LMW fraction and the structure of the major peaks. (B) Absorption spectrum of S1, the major compound. (C) Absorption spectrum of S2, the minor compound in the HPLC profile in (A).

Figure 7.

 

Mass spectral identification of compounds S1 and S2. (A, B) show the electrospray mass spectra of S1 and S2, respectively. (C, D) ESI-MS/MS spectra of the major species in (A) and (B), respectively.

Figure 8.

 

Oxidation of ascorbate by the UVA irradiation of an 8-μM solution of S1 and an 8-μM solution of S2 isolated from pooled 60-year-old human lenses over 60 minutes of irradiation.

Figure 9.

 

Effect of increasing S1 and S2 on the UVA-dependent oxidation of ascorbate (A) compared with riboflavin (B) after 20 minutes of UVA irradiation.

Figure 10.

 

Sensitizer activity of purified compounds of similar structure in the presence and absence of air after 30 minutes of UVA irradiation.

Table 1.

 

Effect of GSH Treatment on the Loss of Ascorbate Oxidizing Activity of Human Lens WI Protein and 4-Week Ascorbate-Glycated CLP

Sample Loss	Activity*			% Activity Loss
		Original	After GSH
4-Week glycated CLP (3 mM ascorbate)	13.3	13.5	0
4-Week glycated CLP (20 mM ascorbate)	23.8	23.4	1.7
WI protein (55 years)	15.3	8.0	47.7
WI protein (55 years)	9.0	7.7	14.4
WI protein (59 years)	12.2	6.2	49.2
WI protein (65 years)	15.5	17.3	0
WI protein (81 years)	27.0	24.0	12.5

 CLP, calf lens protein.

*  Activity is expressed as nanomoles of ascorbate oxidized after 30-minute UVA/mg protein

Figure 11.

 

Oxidation of ascorbate by a UVA sensitizer through an ascorbyl free radical intermediate.