Studies on the Mechanism of the UVA Light-Dependent Loss of Glutathione Reductase Activity in Human Lenses

Abstract

purpose. To determine the mechanism that leads to the UVA light-dependent loss of glutathione reductase (GR) activity in human lens (HL).

methods. Both the HL water-soluble (WS) fraction and yeast GR were irradiated with UVA light (200 mW/(cm² · h) for 1 hour at +20°C, and the specific activity (SA) was observed. GR apoenzyme (apo-GR) was prepared from either HL-WS fractions or yeast GR by treatment with a cold solution of acidic ammonium sulfate. Reconstitution of apo-GR was conducted by mixing enzyme with an excess of flavine adenine dinucleotide (FAD) and purification of GR on a size-exclusion separation column.

results. One hour of UVA photolysis of an HL-WS fraction resulted in a 96% decrease in the SA of GR (6.32 ± 0.22 vs. 0.39 ± 0.01 mU/mg lens protein). Action spectra of GR SA in the WS fraction from HL within the range 320 to 500 nm showed that the enzyme was most vulnerable to the wavelengths in the UVA region with the highest decrease in the SA at 320 to 350 nm (∼23%–28% activity loss within 1 hour of irradiation), and lowest with the wavelengths beyond 400 nm (7%–8% SA loss). UVA irradiation of apo-GR in the crude HL-WS fraction, followed by reconstitution with FAD, showed that 90% of the original SA was recovered. The original GR activity either in HL or yeast GR, however, was not recovered by (NH₄)₂SO₄ (pH 2.25) treatment followed by reconstitution with FAD after UVA photolysis. Experiments with UVA-photolyzed yeast GR revealed that UVA photolysis caused the formation of additional SH groups within the enzyme, as shown by the incorporation of an SH-specific fluorescent probe, 4-(aminosulfonyl)-7-fluoro-2,1,3-benzoxadiazole (ABD-F). Similar results were obtained on the photolyzed iodoacetamide-alkylated yeast GR, which was evaluated by matrix-assisted desorption ionization–time of flight (MALDI-TOF) mass spectrometry.

conclusions. The results show that the reduction of HL GR activity by UVA light was directly linked to the presence of FAD within the enzyme. That the irradiated GR showed de novo formed SH groups argues that UVA photolysis of GR leads to the reduction of the redox-active disulfide within the reaction center of the enzyme, making it inactive.

Since the first studies on the properties of the bovine lens glutathione reductase (GR) by van Heyningen and Pirie ¹ a half a century ago, much effort has been devoted to the investigation of the role of this enzyme in lens homeostasis. ^{2 3 4 5} The enzyme catalyzes the reduced nicotinamide adenine dinucleotide phosphate (NADPH)–dependent reduction of the disulfide bond of oxidized glutathione. GR from human lens (HL) is identical with GR from human erythrocytes and represents a dimer with molecular mass of approximately 100 kDa, consisting of two identical subunits with one flavine adenine dinucleotide (FAD) noncovalently bound to each monomer. ⁴ Each subunit contains six cystiene (Cys) residues, two of which are essential for the enzymatic activity of GR and located in the reactive center of the molecule. ⁶ These two Cys residues form a redox-active disulfide in the reactive center of the enzyme, which participates in a two-electron transfer from NADPH-reduced FAD to oxidized glutathione. ⁷ The primary sequence of the reactive center in GR from Escherichia coli, yeast, or humans, is highly conserved and contains the interchange cysteines Cys⁴², Cys⁴⁵, and Cys⁵⁸; charge transfer thiols, Cys⁴⁷, Cys⁵⁰, and Cys⁶³; and histidines His⁴³⁹, His⁴⁵⁶, and His⁴⁶⁷ as the acid catalyst paired with Glu⁴⁴⁴, Glu⁴⁶¹, Glu⁴⁷², respectively. ^{7 8 9} In addition, the mechanism of the reduction of oxidized glutathione is practically identical for GR from all the aforementioned species. ^{7 8 9 10} Numerous studies have shown that the enzyme is essential for the glutathione redox cycle that maintains adequate levels of the glutathione and thiol groups in the mammalian lens. ⁴ The specific activity (SA) gradient of HL GR has been found to decrease in both normal and senile cataractous lenses from epithelium to nucleus. ^{2 3 5} It was also found that the SA of human GR is usually two to three times lower in cataractous lenses than in the aged-matched control. ² The molecular mechanisms that lead to the inactivation of GR are mostly unknown. Some posttranslational modifications of the enzyme (e.g., oxidation by molecular oxygen and its reactive metabolites) that may lead to its inactivation have been described, but their participation in lenticular GR inactivation has never been established. ^{11 12}  

In a prior communication, we reported that both intact human aged and fetal calf lenses irradiated with UVA light for 1 hour show a dose-dependent loss in GR SA. ¹³ Based on our results, we hypothesized that self-sensitization of FAD within the GR may lead to photo-oxidation of its constituents and to a diminution of its activity with aging and during cataract development as a result of prolonged exposure of HL to UVA light. This claim is well justified, because FAD as the enzyme’s prosthetic group can absorb light in the UVA and near-visible regions of the spectrum. FAD, alternatively, is the UVA light-responsive sensitizer, which, during photolysis at neutral pH, undergoes a cleavage reaction to produce lumichrome as the major degradation product. ^{14 15} According to the literature, both light-excited lumichrome and FAD can photo-oxidize numerous compounds, including amino acids, even in an oxygen-free environment. ^{14 15 16 17} The present study was conducted to elucidate a mechanism that can describe an irreversible inactivation of GR by UVA photolysis. 

Materials and Methods

Materials

Human lenses were obtained from the Heartland Lions Eye Tissue Bank of Missouri, and newborn calf lenses were obtained from Pel-Freeze Biologicals (Rogers, AR). All the reagents used in this project were the highest quality available from Sigma-Aldrich (St. Louis, MO). The thiol-reactive probe, 4-(aminosulfonyl)-7-fluoro-2,1,3-benzoxadiazole (ABD-F), was obtained from Fluka (Milwaukee, WI). Glutathione reductase was from Roche Diagnostics, Indianapolis, IN. 

Methods

Irradiation of Lens WS Fraction and Baker’s Yeast GR Preparations.

The HL or fetal calf lens (FCL) WS fractions ¹³ or preparations of baker’s yeast (BY) GR of defined concentration were placed in a 1.0 × 4.0-cm screw-cap quartz cuvette under either an aerobic or an anaerobic (flushed with argon for 15 minutes) atmosphere. The UVA light delivered by a 1000-W Hg/Xe high-pressure lamp was filtered through a 10-cm water-jacketed filter containing 5% CuSO₄ and then through a 338-nm cutoff filter (total irradiance of 0.925 kJ/(h · cm²) for light between 338 and 475 nm). ¹³  

Action Spectra of Human Lens and BY-GR.

Samples containing 1.0 U/mL of BY-GR or an exhaustively dialyzed HL-WS fraction (1.0 mg/mL) in 0.1 M KPO₄ buffer (pH 7.0) were irradiated under aerobic conditions with the incident light from a spectrofluorometer (model F2500; Hitachi, Tokyo, Japan). Irradiations were performed at 10-nm intervals (20 nm bandwidth) from 320 to 500 nm, and spectra were collected after each 2.0-hour irradiation. All the samples were assayed immediately for the specific GR activity in triplicate at the end of the experiment and the data reported as the percentage of GR inactivated per joule of absorbed light. 

Preparation of HL GR and Yeast GR Apoenzymes and Reconstitution with FAD.

Either the HL-WS fraction (5–30 mg/mL) or BY GR (2.0 mg/mL) in 0.2 M KPO₄ buffer (pH 7.0) in 0.5 mL were adjusted to 1.0-mL with a cold 6-M solution of KBr and precipitated with an ice-cold saturated solution of (NH₄)₂SO₄ (pH 2.5). The extensively washed precipitate was resuspended in 0.1 M KPO₄ buffer (pH 7.0), and each protein solution (0.5 mL) was mixed with 0.5 mL of FAD (1.0 mg/mL in water) and incubated for 1 hour at 4°C. The preparations, extensively dialyzed against 0.1 M KPO₄ buffer were purified on a size-exclusion separation column (1 × 5 cm; Sephadex G-25) using the same buffer. 

Extraction of FAD from BY GR.

BY GR (50 U) in 0.5 mL 0.1 M KPO₄ buffer (pH 7.0) was mixed with 3.5 mL of an ice-cold saturated solution of (NH₄)₂SO₄ at (pH 2.25), and the protein precipitate was collected by centrifugation at 15,000 rpm for 10 minutes at 4°C. The resultant supernatant was adjusted to pH 7.0 with 5 M NaOH and dried on a rotary evaporator equipped with an oil vacuum pump at the ambient temperature. The salt deposit was extracted three times with 4 mL of freshly distilled pyridine, and the combined extract was evaporated and redried two times with 1 mL of methanol. 

Enzymatic Digestion of BY GR.

BY GR (10 U) in 1.0 mL of 0.1 M KPO₄ buffer (pH 7.0) was placed in two 1.0-mL quartz cuvettes with septum screw caps and bubbled with argon for 15 minutes. At the end of a 1-hour irradiation, the UVA and dark control preparations were adjusted to pH 8.0–8.2 with 2 M NaOH. The solutions were digested for 12 hours at 37°C by adding 100 μL of 1.0 mg/mL protease solution (Pronase; Calbiochem-Novabiochem, San Diego, CA) in 0.1 M KPO₄ buffer containing 0.1 mM EDTA and 0.9% NaCl (pH 8.0). The samples were dried in an evaporating centrifuge (Savant Instruments, Inc., Farmingdale, NY) at room temperature. 

Incorporation of a Sulfhydryl-Specific Label into the Non- and UVA-Photolyzed Preparations of BY GR.

BY preparation of GR (2.2 mg/mL; 1-mL aliquots) in 0.2 M KPO₄ buffer (pH 7.0) was placed in an 1.5-mL quartz cuvette with a screw septum cap and deaerated with argon for 15 minutes under ambient conditions. This solution was UVA photolyzed for 1 hour, and both the dark control and UVA-irradiated samples were transferred to separate 7-mL amber vials and precipitated with 6 mL cold acetone (−70°C). The precipitates were washed extensively with ice-cold water, redissolved in 8.0 M GnHCl in 0.2 M KPO₄ buffer (pH 8.5) and derivatized with ABD-F ¹⁸ for 60 minutes at 60°C. The modified GR was extensively dialyzed against 0.1 M KPO₄ buffer (pH 7.0) and protein content and fluorescence (λ_ex=375 nm, λ_em=525 nm) were determined. 

Alkylation of Native and UVA-Irradiated BY GR.

A solution of BY GR (1.0 mg/mL) and BSA (1.0 mg/mL) as a stabilizer in 0.1 KPO₄ buffer containing 0.15 M NaCl (pH 7.2) was placed in 1.5-mL quartz cuvettes with screw caps containing septa and deaerated with argon for 15 minutes. At the end of a 1-hour irradiation period 1 mL of both the dark control and UVA-irradiated sample was transferred to separate 7-mL amber vials and deaerated again. To these solutions 3.0 mL of oxygen-free 8.0 M GnHCl in 0.1 KPO₄ buffer (pH 8.5) was added, followed by the addition of a 100-μL dimethylsulfoxide solution of iodoacetamide (IAM; 15 mg/mL). After a 1-hour incubation period in the dark, the samples were extensively dialyzed against 5% formic acid (1:1 × 10⁶) with 30-kDa cutoff spin filters to remove GnHCl and the reagent. 

Enzyme Assays.

GR was assayed as described by Carlberg and Mannervik. ¹⁹ One unit of enzyme was defined ¹⁹ as 1 micromole NADPH oxidized per minute at 25°C. 

Physicochemical Analyses.

All GR UV-visible (UV-VIS) spectra were recorded on a spectrophotometer (Cary 1E; Varian Analytical Instruments, Sunnyvale, CA). Fluorescence measurements were recorded on a spectrofluorometer (model F2500; Hitachi). BY GR mass determinations were performed by matrix-assisted desorption ionization–time of flight (MALDI-TOF) mass spectrometry (Voyager DePro; Applied Biosystems, Foster City, CA). Spectra were acquired in the positive ion linear delayed extraction mode under optimized conditions. GR samples of 0.5 μL were applied to the stainless-steel target and topped with 0.5 μL sinapinic acid solution (20 mg/mL in acetonitrile, water, and 10% aqueous trifluoroacetic acid solution [5:4.7:0.3 vol/vol/vol]) to initiate crystallization. 

HPLC Separation of FAD and Lumichrome.

The HPLC system (Waters model 600; Milford, MA) was equipped with a C₁₈ column (300 × 4.6 mm; μ-Bondapack; Phenomenex, Torrance, CA), an absorbance detector (Dynamax UV-1; Rainin Instrument Co., Woburn, MA) set at 254 nm, and a detector (model RF-501; Shimadzu Scientific Instruments, Inc., Columbia, MD) set at λ_ex = 334 nm and λ_em = 478 nm (FAD determination) or 450/528 nm (lumichrome determination). HPLC separations were performed at a flow rate of 0.5 mL/min using 1% acetic acid (vol/vol) as buffer A and 1% acetic acid in 50% methanol as buffer B. The following gradient was used: 0 to 5 minutes, 0% buffer B; 5 to 30 minutes, 0% to 100%; 30 to 35 minutes, 100%; and 35 to 40 minutes, 0%. Data were acquired by either an integrator (SC-3R; Shimadzu) or with the acquisition software (Axis; LabAlliance, State College, PA). 

Results

Our results (Fig. 1) show that the SA GR in the HL-WS fraction was more susceptible to UVA photolysis than that of whole lenses. ¹³ UVA photolysis caused a 95.8% loss of HL-WS GR (6.32 ± 0.22 vs. 0.39 ± 0.01 mU/mg), 40.1% in FCL-WS GR (0.31 ± 0.01 vs. 0.19 ± 0.03 mU/mg), and 83.7% in BY GR (9.03 ± 2.55 vs. 1.46 ± 0.04 U/mg). A very similar reduction in GR SA in the above fractions after 1 hour of irradiation was observed under aerobic conditions (data not shown). 

Therefore, oxygen plays little or no role in the photo-inactivation of the enzyme. There was no difference in the rate of GR SA loss from WS fractions derived from the lens nucleus and cortex after 1 hour of photolysis (74% ± 12% decrease in cortical extracts vs. 80% ± 4% loss in the nucleus extracts). 

FAD, which is the only UVA-absorbing molecule in FCL and BY GR, could be responsible for GR inactivation in the studied preparations. The data in Figure 2 show that an extensively dialyzed preparation of HL-WS fraction and BY GR in 0.1 M KPO₄ and 1 mM EDTA buffer (pH 7.0) lost their SA in two spectral ranges (320–360 and 460–480 nm) when irradiated with monochromatic light under anaerobic conditions. The action spectra of the GR SA in the WS fraction from HL also showed that the enzyme was most vulnerable to the wavelengths in the UVA region, with the highest decrease in the SA at 320 to 350 nm (∼23%–28% activity loss), and lowest with the wavelengths beyond 400 nm (∼7%–8% SA loss). The data in Fig. 2A show that most of the enzyme inactivation occurred at the wavelengths closest to the λ_max of FAD (370 and 450 nm). ⁷ At the same time, the data in Figure 2B show that under identical anaerobic conditions of UVA-VIS photolysis, the absorption spectrum of the FAD molecule also showed bleaching, with a loss of the intrinsic absorbance at its λ_max. 

To assess whether FAD was directly involved in the GR inactivation, FAD was extracted from the HL-WS GR and BY preparations, and the apoenzyme preparations ²⁰ were subjected to 1 hour of UVA irradiation under anaerobic conditions, followed by the reconstitution with FAD (Fig. 3) . Triple extraction of either the HL-WS fraction or BY GR with (NH₄)₂SO₄ (pH 2.25) led to a reduction in SA of 88% and 84%, respectively (Fig. 3 , bars 1 and 1′ vs. 3 and 3′), which is consistent with the removal of FAD from the enzyme preparations. ²⁰ Although UVA irradiation of native enzyme in both BY-GR and HL-WS led to almost 90% to 95% loss of SA in both fractions (Fig. 3 , bars 1 and 1′ vs. 2 and 2′), the irradiation of the apoenzyme preparations followed by FAD reconstitution leads to 83% recovery of the original BY GR and 77% of the original GR SA in the HL-WS fraction (Fig. 3 , bars 1 and 1′ vs. 6 and 6′). At the same time, reconstitution of the dark control apoenzyme preparations from both the HL-WS fraction and BY GR led to the almost complete recovery of the original SA in both preparations (Fig. 3 , bars 1 and 1′ vs. 5 and 5′). These findings imply that the photolysis of FAD as a prosthetic group is directly involved in the inactivation of GR. 

Consistent with these data is the finding that the original GR activity from both HL and BY GR after 1 hour of UVA photolysis was not recovered by acidic (NH₄)₂SO₄ extraction after reconstitution with FAD (Fig. 4) . Similar to the data in Figure 1 , the data from the experiments with UVA-photolyzed HL-WS fraction and BY GR revealed a decrease in SA of 76% and 91%, respectively (Fig. 4 , bars 1 and 1′ vs. 2 and 2′). Repeated extraction of FAD from both native BY GR and HL-WS fraction with (NH₄)₂SO₄ (pH 2.25) lowered the SA of these preparations to 12.4% and 3% of the original SA, respectively, in the irradiated fractions and 12.4% and 5.6%, respectively, of the original SA in the control preparations. Although the original GR SA in the dark control apoenzyme preparations from the HL-WS fraction and BY GR was almost fully recovered by FAD reconstitution (87% and 84%, respectively; Fig. 4 , bars 1 and 1′ vs. 5 and 5′), these activities did not recover to the original levels, even after FAD was extracted from the photolyzed enzyme preparations followed by reconstitution with fresh FAD (16% and 2.7%, respectively; Fig. 4 , bars 1 and 1′ vs. 6 and 6′). These data point out that UVA-mediated photolysis of FAD within the structure of the enzyme may cause irreversible changes to the reactive center constituents. 

Despite the dramatic loss of the SA in BY GR (Figs. 1A 3A 4A) , the spectrum of FAD within the enzyme was unchanged either before or after photolysis (Fig. 5) . Adjustment to 4 M GnHCl released FAD from the molecule. A comparison of the spectrum of FAD released from the irradiated enzyme in Figure 4A with pure FAD in Figure 2B shows that no spectral change occurred in this prosthetic group after 1 hour of irradiation. 

We observed only a minor change in FAD from the enzyme photolyzed under aerobic conditions within the 300- to 350-nm range. The slight increase in absorbance within this region may be indicative of accumulation of lumichrome. ¹⁵ Consistent with these results were data from two series of experiments designed to trace the fate of FAD within the structure of the UVA-photolyzed enzyme. In the first series, FAD and its possible UVA-degradation products were extracted from the enzyme by acidic (NH₄)₂SO₄, dried and re-extracted with pyridine (see the Methods section). HPLC separation of these extracts showed that approximately 80% to 85% of the FAD was extracted from both dark control and UVA-irradiated enzyme preparations (Fig. 6A) . No additional peaks, which could indicate the formation of new molecules from UVA-irradiated enzyme, were detected in this system. Because the FAD peaks before and after UVA irradiation are practically identical (Fig. 6A) , it argues that this portion of FAD within the photolyzed enzyme is intact and did not undergo the UVA-mediated degradation. 

A second series of experiments was conducted to evaluate whether FAD and its photo-degradation products became covalently attached to GR during photolysis. The native and UVA-photolyzed BY GR were digested with protease solution (Pronase; Calbiochem-Novabiochem), and the final digest was submitted to HPLC analysis (Fig. 6B) . The data show that no change in the FAD peak’s shape occurred before and after irradiation, and no additional change was seen in the peaks area. This supports an absence of FAD photolysis within the enzyme during irradiation (Fig. 6B) . In addition, no new UV-absorbing peaks emerged in the chromatographed digests from native and photolyzed GR recorded with the UV-detector set at λ = 254 nm (data not shown), which indicates that UVA-irradiation of GR does not facilitate photolysis of the enzyme’s prosthetic group. 

Alkylation of both the dark control and UVA-irradiated preparations of BY GR with the SH-specific ABD-F probe under denaturing conditions led to the incorporation of this probe into the enzyme. This probe is highly specific toward thiols and becomes fluorescent only after the probe is covalently bound to mercaptans. ¹⁸ Our data show (Fig. 7) that the incorporation of this probe into dark control samples of GR yielded an ABD-F–related fluorescence of 13.8 ± 0.4 relative fluorescence units (RFU). In the UVA-photolyzed samples we observed a statistically significant increase in ABD-F–mediated fluorescence by approximately 20.0% to 17.2% ± 0.7% RFU. The fact that, under the denaturing conditions of the ABD-F reaction with GR preparations (6 M GnHCl; 60°C) the homodimeric molecule of GR exists as a solution of monomers with each of the units containing four Cys and one disulfide, ²¹ indicates that at least one additional equivalent of the SH group was generated as the result of UVA photolysis of GR. 

The de novo formation of additional SH groups within the enzyme was independently verified, with the data from the experiments designed to determine the molecular mass of the iodoacetamide-alkylated preparations of the native and UVA-irradiated enzyme under anaerobic conditions (Fig. 8) . In these experiments the dark control and UVA-photolyzed GR was alkylated with IAM in 8.0 M GnHCl at pH 8.5 (Fig. 8) . After extensive dialysis against 5% formic acid, the molecular mass of GR was determined with MALDI-TOF mass spectrometry. Our data on the nonalkylated enzyme show that dark control and UVA species exhibited a single-charge species with molecular mass of 51,483.25 Da and 51,499.99 Da, respectively (data not shown), which indicates that no UVA-mediated binding between FAD or its UVA-degradation products occurred in the course of photolysis of the enzyme under anaerobic conditions. An alkylation of these preparations with IAM increased the molecular mass of the enzyme in the dark control preparation to 51,670.1 Da (Fig. 8A) . IAM treatment of the UVA-irradiated enzyme increased its mass to 51,791.77 Da (Fig. 8B) . The molecular mass data on the alkylated GR preparations indicate that two additional equivalents of the SH group were generated as a result of UVA-photolysis of GR, because the adduct of the attachment of IAM to cysteine would increase the molecular mass by 58 Da, and the difference between the mass of the control (Fig. 8A) and irradiated GR is 121 Da (Fig. 8B) . 

Discussion

The mechanisms of GR diminution in the HL are not well understood. Some posttranslational modifications of the enzyme (e.g., oxidation by reactive oxygen derivatives) that may lead to its inactivation are described, but their participation in lenticular GR inactivation has never been established. ^{11 12} There is only one report that that bovine GR becomes heat labile with age, purporting to show that some entropy-driven changes occur within the enzyme. ²² Conversely, the report of Latta and Augusteyn ⁴ showed that there is no difference in the kinetic parameters (V _max and K _m) for oxidized glutathione and NADPH₂ and for the heat lability of human GR isolated from the cortex and nucleus of the human type I cataractous lens. ⁴ According to Zhang and Augusteyn ²³ the accumulation of the metabolically inert crystallins in the maturing lens fiber cell in the absence of the GR biosynthesis leads to a decrease in the SA of GR in the inner layers of mammalian lens. Although this hypothesis may explain the drop in GR SA in the aged lenses, its validity is questionable when applied to the enzyme in cataractous lens. Srivastava et al. ² showed that the SA of the enzyme in the cataractous lenses is lower by 25% and 60% in the cortex and nucleus, respectively, when compared with the aged-matched control. 

In an earlier communication, we showed that a 1-hour UVA light exposure of the whole HL leads to a 70% reduction in GR SA. ¹³ A similar loss of the enzyme’s SA in different species under the anaerobic conditions of the experiment suggests that the same mechanism is involved in the UVA-mediated inactivation of the enzyme (Fig. 1) . That the GR cofactor FAD is the only molecule in BY GR that absorbs UVA light suggests that direct photolysis of FAD is responsible for the inactivation of the enzyme (Fig. 1) . This idea is supported by the action spectra of these preparations (Fig. 2A) , which are very similar to the absorption spectrum of the FAD in the UV-VIS region (Fig. 2B) . Our data show that most of the enzyme inactivation occurred at the absorption maxima of FAD (370 and 450 nm). The validity of this claim is independently verified by the results obtained by irradiation of the apoenzyme from both HL and BY followed by its reconstitution with FAD (Fig. 3) . Full recovery of GR SA in the FAD-reconstituted enzyme preparations from both lenticular and BY enzymes after 1 hour of UVA photolysis once again showed that FAD is necessary for photo-inactivation. It was found in this study that the extraction of FAD and the reconstitution of the UVA-photo-inactivated enzyme from both BY and HL did not recover the original SA. Therefore, FAD appears to have caused some irreparable changes in the enzyme reaction center as the result of its UVA-mediated self-sensitization (Fig. 4) . 

The photo-induced inactivation of various enzymes in the presence of FAD, flavine mononucleotide (FMN), and riboflavin has been demonstrated in several studies. ^{24 25 26 27} Based on these studies, we conclude that there are few possibilities on how enzyme-bound and UVA-excited FAD may cause irreversible changes to GR. (1) It could form degradation products that can no longer function as the prosthetic group, (2) FAD and/or its photo-degradation products could bind to the enzyme covalently, or (3) FAD and/or its photo-degradation products could reduce the enzyme’s reaction center disulfide bond. The results of FAD re-extraction with pyridine followed by HPLC determination show that almost 90% of FAD remains intact, which rules out the first possibility. This finding is independently corroborated by the absorption spectra of native and UVA inactivated under aerobic conditions BY GR, which show little or no difference (Fig. 5A) . We also found that lumichrome, which is a possible FAD photo-degradation product, ^{14 15 16 17} does not bind to the reaction center of the enzyme, as judged from the enzyme’s absorption spectrum taken after the reconstitution of the BY apoenzyme with this compound (data not shown). Taken together, these data rule out the possibility that UVA-mediated GR inactivation occurs through the photo-induced splitting of FAD as the prosthetic group in GR. The absence of any additional fluorescent or UV-absorbing peaks in the HPLC profile of the protease digests of the control and UVA-inactivated BY enzyme (Fig. 6B) also indicates that no covalent binding of FAD and/or its photo-degradation products to the reaction center of the enzyme occurred as the result of photolysis. 

Within the GR reaction center, FAD is located in very close proximity to the redox active disulfide, ^{6 7 8 9 10} to facilitate the flow of electrons and hydrides from NADPH₂ to its cystine and then to oxidized glutathione. Our data with ABD-F incorporation in the control and the UVA-photolyzed BY GR preparations show that UVA-excited FAD as the prosthetic group of GR appears to have reduced the redox-active cystine in the GR molecule (Fig. 7) . Although at this point it is not clear why only a 20% increase in ABD-F–induced fluorescence in the UVA–photo-inactivated enzyme was observed, rather than a 40% increase, the data show that UVA irradiation causes the reduction of the disulfide in the reaction center of the molecule by FAD. These data, however, are in good agreement with the results of molecular mass determination of IAM-alkylated GR. We have determined that the difference between the molecular mass of the control (Fig. 8A) and irradiated GR is 121 Da (Fig. 8B) , which roughly corresponds to the incorporation of two additional equivalents of IAM per one monomer of GR. Taken together, our results support a UVA-mediated inactivation of the enzyme that probably proceeds through a one- or two-electron transfer from FAD to the redox-active disulfide within the reactive center of the molecule. 

 Supported by National Eye Institute Grants EY13244 and EY07070 and by a departmental grant from Research to Prevent Blindness, Inc.

 Submitted for publication April 16, 2003; revised May 20, 2003; accepted May 23, 2003.

 Disclosure M.D. Linetsky, None; J.M.W. Hill, None; V.G. Chemoganskiy, None; F. Hu, None; B.J. Ortwerth, 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.

 

Decrease in the GR SA of different lens preparations (55–66 years old) and BY GR (milliunits × 10⁴ per mg protein), due to 1 hour of UVA photolysis under anaerobic conditions. The data represent the average of results in three independent experiments performed in duplicate.

Figure 2.

 

Normalized action spectra of the GR SA in the dialyzed HL-WS fraction (6.95 mU/mL; 1 mg protein per milliliter) and BY GR (10 U/mL) in 0.1 M KPO₄ buffer (pH 7.0), volume = 1.0 mL under anaerobic conditions (A). All the assays were performed in duplicate. Each data point showing reduction of SA was calculated as the decrease per joule of light absorbed. (B) A nonirradiated solution of FAD (50 μg/mL) and FAD solution after 60-minute of UVA photolysis under anaerobic conditions.

Figure 3.

 

One-hour UVA anaerobic photolysis of both BY GR (50U/mL) and the HL-WS fraction (20 mg/mL) from aged lenses (ages 55–66 years) and their corresponding apoezyme preparations. Bars 1and 1′ are the dark control, and bars 2 and 2′ are the correspondent UVA-irradiated preparations. Bars 3 and 3′ are the apoenzyme preparations from 1 and 1′, respectively. Bars 4 and 4′ are the 1-hour UVA-photolyzed BY GR and HL-WS apo-GR preparations. Bars 5 and 5′ are native apoenzyme preparations of 3 and 3′ reconstituted with FAD (1 mg/mL), respectively (dark control). Bars 6 and 6′ are UVA-irradiated apoenzyme preparations of 4 and 4′ reconstituted with FAD (1 mg/mL), respectively. The data represent the average of two independent experiments assayed in triplicate.

Figure 4.

 

Reconstitution experiments of 1-hour UVA photolyzed BY GR (50 U/mL) or the HL-WS fraction (20 mg/mL) from aged lenses (ages, 55–66 years) under anaerobic conditions at 20°C. Bars 1 and 1′ are the dark control, and bars 2 and 2′ are the correspondent UVA-irradiated preparations. Bars 3 and 3′ are apoenzyme preparations from 1 and 1′, respectively. Bars 4 and 4′ are apoenzyme preparations made from 1-hour UVA-photolyzed BY GR and HL-WS GR fractions. Bars 5 and 5′ are native apoenzyme preparations of 3 and 3′ reconstituted with FAD, respectively. Bars 6 and 6′ are UVA-irradiated apoenzyme preparations of 4 and 4′ reconstituted with FAD, respectively. The data represent an average of two independent experiments assayed in triplicate.

Figure 5.

 

Spectra of 1-hour UVA-photolyzed BY GR and its corresponding dark control under anaerobic (A) and aerobic conditions (B). Each sample was diluted to 0.5 mg/mL with 8 M GnHCL, and spectra were recorded in the 200- to 600-nm range.

Figure 6.

 

HPLC separation of FAD from acidic ammonium sulfate extracts of native (bottom trace) and UVA-irradiated (top trace) BY GR (A) and FAD liberated from enzymatic digests of native (bottom trace) and UVA-irradiated enzyme (top trace; B). The chromatograms were recorded with the fluorescence detector set at λ_ex = 334 nm, λ_em = 478 nm.

Figure 7.

 

Incorporation of the ABD-F probe into non- and UVA-irradiated preparations of BY GR (2.2 mg/mL) in 8.0 M GnHCl in 0.2 KPO₄ buffer (pH 7.2) at 60°C for 1 hour. The data represent the average of results in three independent experiments.

Figure 8.

 

Determination of the molecular masses of the IAM-alkylated native and BY GR UVA photolyzed for 1 hour under anaerobic conditions.