Transition Metal–Catalyzed Oxidation of Ascorbate in Human Cataract Extracts: Possible Role of Advanced Glycation End Products

Abstract

purpose. With age, human lens crystallins become more pigmented, oxidized, modified by ascorbate oxidation and advanced glycation end products (AGEs), and bind copper. The hypothesis was tested that the major AGE and ascorbylation product in the human lens, N^ε-carboxymethyl-l-lysine (CML), has an EDTA-like structure, which may predispose it to bind redox active copper.

methods. Young, old, and cataractous human lens protein fractions were glycated with ascorbic acid and tested for their ability to bind Cu(II) by atomic absorption spectroscopy and oxidize (¹⁴C₁)-ascorbate by radiometric thin-layer chromatography method. AGEs were assayed by high-performance liquid chromatography (HPLC). CML-rich proteins were immunoprecipitated from young, old, and cataractous crystallins using affinity-purified CML antibody and tested for their ability to oxidize ascorbate and generate hydroxyl radicals in the presence of H₂O₂ using 5,5′-dimethyl-1-pyrroline-N-oxide (DMPO) spin-trap and EPR spectroscopy.

results. Ascorbate oxidizing activity at 24 hours of native crystallins was significantly increased in both the water soluble (WS; P < 0.001) and insoluble (WIS; P < 0.05) fractions from cataractous and normal lenses. The chelator DTPA completely prevented oxidation up to 24 hours of incubation but less effectively thereafter. Mean endogenous Cu content in pooled young, old, and cataract fractions increased from 0.016 to 0.026 nmol/mg protein, respectively, in WS (P < 0.05) and WIS (P < 0.001) fractions, and Cu(II) binding was 20% to 30% increased in cataractous versus old and young lenses in WS (P < 0.01) and WIS (P < 0.001) fractions. Mean levels of the AGEs, CML, and pentosidine were markedly elevated in WS and WIS fractions from cataractous versus old or young crystallins (20% to severalfold, P < 0.05 to P < 0.001). In a separate experiment, protein-bound Fe was not elevated. Crystallins ascorbylated in vitro showed an increase in CML as well as Cu(II) binding. CML-rich proteins (immunoprecipitated from cataractous lenses) oxidized ascorbate ∼4 times faster than similar proteins from young and old normal lenses (P < 0.01) and generated hydroxyl radicals in the presence of H₂O₂ and DMPO.

conclusions. The association between CML formation, copper binding, and generation of free radicals by cataractous lens crystallins can be duplicated by ascorbylation in vitro. These effects are only in part attributable to CML itself, and other modifications (AGEs, conformational changes) may participate in the process. A vicious cycle between AGE formation, lipoxidation, and metal binding may exist in the aging lens, suggesting that chelation therapy could be beneficial in delaying cataractogenesis.

The aging of the human lens is associated with progressive changes in the physical–chemical properties of crystallins, which include aggregation, pigmentation, formation of disulfide and nondisulfide cross-links, fragmentation, and ability to form free radicals when exposed to UV light. ^{1 2 3 4} Losses of lysine, arginine, and histidine residues have been described, although not consistently. ^{5 6 7 8} Many of these changes are relatively mild in the aging normal lens but are dramatically accelerated during cataractogenesis. The incubation of lens crystallins with glycating sugars, ascorbate, and oxidizing agents such as transition metals ^{9 10 11} or irradiation of the crystallins with UVA in the presence of endogenous lenticular photosensitizers ^{4 12} can simulate many age-related modifications of lens crystallins. 

Among possible mechanisms, considerable evidence has accumulated in support of the Maillard reaction (i.e., the reaction that leads to the formation of advanced glycation end products [AGEs] and protein cross-links). It is now well recognized that the Maillard reaction does not need to be initiated by glucose but rather can also be initiated by oxoaldehydes such as methylglyoxal. ^{13 14} In addition, there is evidence that suggests ascorbate may be a more important modifier of lens crystallins than glucose, because the modification rate by ascorbate is much faster than that of glucose at equivalent concentrations. ¹⁵ To date, close to 10 AGEs and cross-links have been discovered in the human lens besides glycated lysine itself. They include the following: N^ε-(carboxymethyl)-l-lysine (CML), pentosidine, fluorophore LM-1, pyrraline, N^ε-carboxyethyl-lysine (CEL), methyl glyoxal lysine dimer (MOLD)/imidazolysine, glyoxal lysine dimer (GOLD), argpyrimidine, and oxalic acid monoamide (OMA). ^{16 17 18 19 20 21 22 23 24}  

Although there is strong evidence that amino-carbonyl reactions of the Maillard type are involved in the aging process of lens crystallins, recent evidence also suggests that purely oxidative modifications of lens crystallins occur in aging. This has been demonstrated by the accumulation of hydroxylated amino acids. ²⁵ Although some of these oxidation reactions could be of photooxidative nature and involve singlet oxygen (¹O₂) or superoxide anion, the formation of most hydroxylated amino acids cannot occur without a metal catalyst such as copper. Presence of the latter was indeed found to be increased in old and cataractous lenses in most, although not all, studies. ^{26 27 28 29 30}  

In a preliminary study, we found that the rate of spontaneous ascorbate oxidation increases in the presence of old and pigmented cataractous versus young human lens crystallins, ³¹ thus raising the question of the mechanism by which this oxidation process is catalyzed in the aging lens. Here we hypothesized, based on our recent finding of a structural analogy between the major AGE CML and EDTA, ³² that AGE formation in lens crystallins leads to binding of redox active copper. We present below evidence that strongly suggests a close relationship between CML formation and copper-catalyzed ascorbate oxidation by lens crystallins modified by AGEs in vitro and lens crystallins from old and cataractous human lenses. 

Methods

Human lenses were obtained from the Division of Surgical Pathology at University Hospitals of Cleveland, the Cleveland Eye Bank, and the National Disease Research Interchange (Philadelphia, PA). Chemicals were of the highest analytical grade available and were obtained from Sigma Chemical (St. Louis, MO) or Aldrich Chemical (Milwaukee, WI). l-Ascorbic acid (ASA) and DTPA were obtained from Sigma, and metaphosphoric acid was purchased from Aldrich. (¹⁴C₁)-ASA was purchased from New England Nuclear Research Products (Boston, MA) and dissolved fresh in water that had been treated with Chelex 100 resin (Bio-Rad Laboratories, Hercules, CA) to remove metal ions. Silica gel reverse-phase thin-layer chromatography (TLC) plates (0.2-mm-thick) were from Alltech (Deerfield, IL). Reverse-phase C18 column (0.4 × 25 cm, 5 μm, catalog No. 218TP54) was from Vydac (Hesperia, CA). Dialysis membranes (MW cutoff 5000) were obtained from Spectra/Por (Houston, TX). Protein A–Sepharose CL-4B was obtained from Pharmacia Biotech (Uppsala, Sweden). 

Lens Crystallin Preparations

All solvents, solutions, buffers, and dialysis buffers used throughout this work were treated with Chelex 100 according to manufacturer’s instructions to prevent artificial enrichment of protein solutions with transition metals. Human lenses of various ages and pigmentation were divided into three groups (young, 10–25 years; old, 65–80 years; and cataractous, 75–90 years) and classified according to degree and pattern of pigmentation into type IV and brunescent according to the system of Pirie. ¹ Each group comprised pools of 12 lenses each. Lenses were homogenized with metal-free 100 mM sodium phosphate buffer (pH, 7.4). The homogenate was centrifuged at 30,000g for 30 minutes, and the resultant supernatant and pellet were separated. The pellet was suspended in 1 ml homogenization buffer and centrifuged as above. The two supernatant fractions were pooled and dialyzed twice against 10 mM phosphate buffer for 24 hours. The dialysates were centrifuged, and the clear supernatants were designated as water soluble (WS) fraction. The water insoluble (WIS) pellet was resuspended in buffer and sonicated in ice for 5 minutes as previously described. ³³ The solubilized protein was recovered after centrifugation at 30,000g, resuspended, and sonicated again. The second supernatant was combined with the first and designated as the water insoluble sonicated supernatant (WISS). 

Binding of Copper to Crystallins

Copper binding experiments were performed by reverse dialysis. The protein solutions (∼5 mg/ml) were packed in dialysis membranes and incubated in a beaker with freshly prepared 500 μM CuCl₂ in 50 mM Tris–HCl (pH, 7.4). After 6 hours at room temperature, excess free metal was removed by dialysis at 4°C against a 3-l buffer for 3 days with 2 to 3 buffer changes at 4°C. Copper content in the crystallin solutions was quantified by atomic absorption spectrophotometry (atomic absorption spectrophotometer model 2280; Perkin–Elmer, Foster City, CA) using homemade standards in buffer. 

Acid Hydrolysis of Proteins

Crystallin protein fractions were hydrolyzed with 6N HCl for 18 hours at 110°C. Hydrolysates were dried and taken up in an aqueous solution of 0.01 M heptafluorobutyric acid for pentosidine and CML assay by HPLC. Total amino acid concentration estimation in the hydrolysate of insoluble crystallins was determined using l-leucine as standard as described previously. ¹⁸  

HPLC Assay for N^ε-(Carboxymethyl)-l-Lysine

CML was assayed in the protein acid hydrolysate using the o-phthaldialdehyde (OPA) method and post-column derivatization as previously described, ³⁴ with some modifications. Briefly, CML-containing fractions were isolated using system 1 consisting of water (eluent A) and 60% acetonitrile in water (eluent B) both with 0.01 M heptafluorobutyric acid (Aldrich), reverse-phase C18 column (0.4 × 25 cm, 5 μM), Vydac 218TP54, flow rate 1 ml/min. Gradient was 2% B for 20 minutes, followed by 2 minutes to 100% B. The collected CML-rich fractions were analyzed by OPA method using system 2. Column and flow were as for system 1, with 5% propanol (eluent A) and 60% propanol in H₂O (eluent B), both with 3 g of sodium dodecyl sulfate (SDS) and 1 g of monobasic sodium phosphate monohydrate/l adjusted to pH 2.8 with phosphoric acid. The gradient was 15% B to 22% B in 30 minutes, and 22% to 100% B in 5 minutes. 

HPLC Assay for Pentosidine

Pentosidine in the protein hydrolysate was quantified by combined reverse-phase and cation-exchange HPLC as described by Odetti et al. ³⁵ Approximately 10 μmol (leucine equivalent) of crystallins was injected into a C18 reverse-phase column (Vydac 218TP, 0.46 × 25 cm, 10 μm) and eluted from 0 to 35 minutes with a linear gradient of 10% to 16.8% acetonitrile from 0 to 35 minutes in water and 0.01 M heptafluorobutyric acid (HFBA). Eluate containing pentosidine was collected, dried in a Speed Vac concentrator (Savant Instruments, Hicksville, NY), and reconstituted in 200 μl of 0.02 M sodium acetate buffer, and 125 μl was injected into a cation-exchange column (Water Protein Pak SP 5PW, 7.5 × 75 mm). Elution of pentosidine was achieved by using a gradient of 0 to 0.06 M NaCl over a period of 40 minutes. Pentosidine was quantified with a fluorescence detector by comparing peak areas with those of the standards. 

Preparation of Ascorbylated Crystallins

Ascorbylated crystallins were prepared by incubating normal crystallins (15 mg/ml) with various concentrations of ascorbic acid (0.5, 5, and 20 mM) in 100 mM phosphate buffered saline (PBS) (pH, 7.4) for a period of 1, 2, 4, and 30 days. All the incubations were carried out in the dark and under sterile conditions. At the end of the incubation period, all reaction mixtures were extensively dialyzed (twice overnight) against either non–Chelex-treated phosphate buffer to remove the reactants, but not the transition metals that would bind to the AGEs, or the native protein. The solutions were stored in polypropylene tubes at −80°C. 

Ascorbate Oxidation Assay

Ascorbate oxidation assay was performed as previously described. ³¹ The reaction volume of 300 μl consisted of protein (3 mg/ml) with 2.0 mM of ¹⁴C₁-ascorbate (1.25 mCi/mmol) in Chelex-treated 100 mM Na/PO buffer (pH 7.4) with and without 1 mM DTPA. The incubation was performed at 37°C in the dark, with mild rotary shaking. Aliquots were withdrawn at various time intervals, mixed immediately with cold 4% metaphosphoric acid containing 1 mM DTPA, and centrifuged at 10,000g for 1 minute, and the protein-free supernatants were spotted immediately under N₂ and analyzed for ASA degradation products by TLC. The peaks were identified by comparison to known standards and quantified by determining the distribution of radioactivity using a Berthold (Wildbad, Germany) radioactivity scanner (type LB 2852). TLC was performed on silica gel reverse-phase plates (catalog No. 4809–820; Alltech) with a solvent system comprised of acetonitrile–water–acetone–glacial acetic acid (80:15:5:2). 

Preparation of Affinity-Purified CML Antibody

A polyclonal CML antiserum from a 6-week-old New Zealand white rabbit immunized with carboxymethylated Keyhole Limpet Hemocyanin prepared as described by Reddy et al. ³⁶ For purification by affinity chromatography the polyclonal antiserum (15 ml) was mixed with saturated (NH₄)₂SO₄ (1:1 vol/vol) and incubated at 4°C for 6 hours. The mixture was centrifuged at 3000g for 15 minutes, and the pellet was reconstituted in 30 ml of 50 mM Tris–HCl (pH, 7.4). 

A CML–Sepharose affinity column was prepared by reacting ∼5 g l-lysine Sepharose (catalog No. L-6132; Sigma) swollen in 50 ml of PBS with 25 mM glyoxylic acid and 25 mM of NaCNBH₃ in PBS so as to make a total volume of 75 ml. The mixture was incubated at 37°C overnight with mild rotary shaking. Thereafter, the gel was extensively washed with PBS, packed in a column (Bio-Rad, 2.5 cm × 7 cm), and equilibrated in 50 mM Tris–HCl (pH, 7.4). 

Reconstituted (NH₄)₂SO₄ precipitated serum sample (30 ml) was passed over the CML–Sepharose column and allowed to sit for ∼30 minutes. The column was washed until the absorbance at 280 nm reached zero (approximately 40 ml). The bound antibody fraction was eluted with 100 mM glycine/HCl buffer (pH, 2.5) containing 150 mM NaCl. The eluent was collected in tubes containing 0.5 ml of 1 M Tris–HCl (pH, 8.0) to immediately neutralize the pH. The collected eluent was adjusted to a pH of 7.4, dialyzed against 50 mM Tris–HCl (pH, 7.4) for 24 hours with one change of buffer, and concentrated to ∼0.8 ml by Centriprep (10,000 cutoff) from Amicon (Beverly, MA). The purity and activity of the affinity-purified antibody was analyzed by SDS–polyacrylamide gel electrophoresis and by direct enzyme-linked immunosorbent assay (ELISA), respectively. The ELISA showed antibody binding was completely inhibited by 3 different CML-rich proteins, by bovine serum albumin modified with glyoxylic acid but not glycolaldehyde under reductive alkylation with NaCNBH₃ (data not shown). Thus, the antibody was highly specific for CML. 

Immunoprecipitation of CML-Rich Lens Proteins

To water-soluble crystallins (0.4 mg/200 μl) from normal young, normal old, and cataractous lenses were added 50 μl of affinity-purified anti-CML antibody, and the mixture was allowed to incubate for 3 hours at room temperature. Thereafter, 175 μl of the wet slurry of Protein A–Sepharose CL-4B was added, followed by incubation at 4°C for another 3 to 4 hours. The mixture was centrifuged, and the pellet was washed once with dilution buffer and three times with 20 mM Tris–HCl (pH, 7.4) and 150 mM NaCl. Glycine buffer (pH, 2.5) with 150 mM NaCl was used to elute the antigen–antibody complex that was collected by centrifugation after removal of Protein A–Sepharose by centrifugation. The immunoprecipitated protein was analyzed for protein concentration, CML content, and ascorbate oxidizability. For the latter experiment, 75μ l of 0.1 M glycine buffer (pH, 2.5)–containing Chelex-treated NaCl (150 mM) was added to the precipitate to release the crystallins. 

Detection of Hydroxyl Radical Formation by CML-Rich Cataractous Human Lens Crystallins

The ability of CML-rich crystallins from cataractous lenses to generate free radicals in the presence of H₂O₂ was evaluated using the modified protocol of Buettner et al. ³⁷ The final incubation mixture consisted of Chelex-treated 50 mM sodium phosphate buffer (pH, 7.4) with, or without, either CuCl₂ (0.5 mM) or the CML-rich immunoprecipitated crystallins (0.5 mg/ml) prepared as described above, 80 mM 5,5′-dimethyl-1-pyrroline-N-oxide (DMPO; with or without 1 mM DTPA), and 0.5 mM H₂O₂ in a total volume of 0.490 ml. The reaction was started by the addition of 10 μl of 25 mM H₂O₂ and the EPR spectrum was recorded after 8 minutes at room temperature with a Varian Century model Series E instrument. The settings were as follows: scan range 2 × 10² G, time constant 2.5 seconds, modulation amplitude 4.0 G, gain 2 × 10⁴, microwave power 12.5 mW, and microwave frequency 9.44 GHz. 

Results

Pigmented Cataractous Human Lens Crystallins Are Redox Active and Bind Copper

Our previous observation of increased ascorbate oxidation by old and cataractous lens crystallins was made using intact lens homogenate. Levels of soluble antioxidants may have therefore affected the oxidation rate. To investigate whether highly pigmented crystallins, per se, have oxidizing activity, we repeated these experiments using dialyzed protein fractions from young, old normal, and cataractous lenses consisting of grade IV and brunescent lenses according to Pirie. ¹ As shown in Figure 1 , the extent of ascorbate oxidation after 1 and 2 days of incubation with (¹⁴C)-ASA was similar in young and old normal dialyzed lens crystallins but was significantly enhanced by cataractous lens crystallins. The oxidizing activity was more pronounced in WS than WISS fractions. The ascorbate oxidizing activity was significantly suppressed in most specimens by chelation with DTPA, suggesting involvement of transition metals. At day 2 of incubation with radiolabeled ASC (Figs. 1C 1D) , ASC oxidation was further enhanced, and less effectively prevented by DTPA, in cataractous crystallins compared with those of normals. 

Based on preliminary studies implicating copper binding to AGE-modified proteins, ³² we have determined endogenous crystallin-bound copper content as well as ability of lens proteins to bind this transition metal (Fig. 2) . A significantly enhanced Cu content was detected in the protein fraction of cataractous lenses extracted with, and dialyzed against, metal-free solvents. This suggests the detected copper is of endogenous origin, and our values are in agreement with those from the literature, which varied from 0.008 to 0.028 nmol/mg dry weight. ^{26 27 28 29 30} When the crystallins were loaded with 500 μM CuCl₂ by reverse dialysis and then dialyzed, the total amount of Cu bound to crystallins increased 1,000-fold, most likely because of peptide backbone binding (Fig. 2B) . Yet, the relationship with old age and the cataractous process remained significant. 

In a separate experiment, Fe levels were determined in the dialyzed WS fractions of 4 young, 4 old, and 4 cataractous lenses. Mean ± SD levels were 4.1 ± 2.1, 6.7 ± 5.0, and 4.2 ± 1.5 pmol/mg protein, respectively. One outlier was responsible for the high mean values in the old group, which on removal became 4.4 ± 1.9 pmol/mg protein. Differences between groups were not statistically significant, with or without considering the outlier (P > 0.05). Fe values from the literature vary from 0.2 to 12 μg/g lens wet weight and are thus are quite comparable. ^{38 39}  

AGE Content of Pigmented Cataractous Lens Crystallins

Taken together, the data shown in Figures 1 and 2 suggested a possible relationship between the degree of crystallin pigmentation, the extent of Cu binding, and the ability of crystallins to oxidize ascorbate. 

To investigate the role of AGEs in this process, a series of experiments were carried out. First, we determined the levels of the two advanced glycation and glycoxidation products pentosidine and CML. CML levels in the young, old normal, and cataractous protein fractions were 1.4 ± 0.2, 2.3 ± 0.5, and 4.4 ± 0.6 nmol/mg protein in WS fractions and 0.8 ± 0.2, 2.9 ± 0.3, and 6.2 ± 0.5 nmol/mg protein in WISS fractions, respectively. Pentosidine values were 0.2 ± 0.05, 0.4 ± 0.1, and 3.3 ± 0.6 pmol/mg in WS fractions, and 1.3 ± 0.4, 1.5 ± 0.4, and 3.2 ± 0.5 pmol/mg protein in WISS fractions, respectively. Thus, compared with young lens, the cataractous lens proteins showed∼ 15-fold (P < 0.001) and ∼threefold (P < 0.05) increases in pentosidine levels, and CML levels were similarly (∼3.5-fold, P < 0.01; and 6-fold, P < 0.001) increased in the WS and WISS fractions, respectively. We and others previously reported similar data. ^{17 40 41}  

Effects of Ascorbylation of Human Crystallins on Copper Binding

The results noted above suggested that the increased oxidizing effect of pigmented crystallins could be related to the presence of AGEs. To investigate this hypothesis, we evaluated the ability of lens crystallins to oxidize ascorbate on glycation up to 30 days under aerobic conditions, with increasing concentrations of ascorbate (0, 0.5, 5, and 20 mM). Figure 3 shows that within 1 to 2 days ascorbate-modified protein oxidized more ascorbate than at time 0, suggesting that incubation alone may be exposing redox active sites. Importantly, however, not only was the extent of oxidation related to the concentration of ascorbate and partially suppressible with DTPA, but the ascorbate concentration required to induce these effects was near physiological concentration (i.e., 1–2 mM in the human lens). Interestingly, on prolonged incubation (30 days) the protective effect of DTPA on ascorbate oxidation by the AGE proteins waned, suggesting thereby that the heavily modified proteins behave similarly to the cataractous lens crystallins in which ascorbate oxidation was only partially suppressible by DTPA (see Fig. 1 ). The inability of DTPA to completely prevent oxidation at day 30 (Fig. 3) may be explained either by the inability of DTPA to remove the CML-bound metal (Cu or Fe) because of excessive cross-linking associated with long-term glycated crystallins, or because ascorbate oxidation was not transition metal dependent. 

To further clarify this phenomenon, CML formation and copper binding by the ascorbylated lens proteins were determined. Both the amount of CML formed and the amount of copper bound on incubation of the crystallins with 500 μM CuCl₂ increased as a function of glycation duration (Fig. 4) . CML formation with 20 mM ascorbate was not significantly higher than with 5 mM, suggesting that competing reactions may have prevented further CML increase. The close relationship between CML formation and copper binding suggests a possible direct relationship between the two phenomena. However, the relationship with redox activity may be more complex, because the data in Figure 3 indicate a marked effect of DTPA-suppressible ascorbate oxidation after 24 hours of incubation only. 

Binding of Copper by CML-Rich Human Cataractous Crystallins

Seeking in vivo evidence for the role of CML residues in the binding of redox active metal and catalysis of ascorbate oxidation, we immunoprecipitated CML-rich proteins from normal young, old normal, and cataractous human lenses using affinity-purified polyclonal anti-CML antibodies. With this approach we were able to immunoprecipitate 2 and 3 times more protein from pooled old normal and cataractous lenses, respectively, than young human lenses (Fig. 5 , top). This was associated with a 1.5- and 2.5-fold enrichment in CML content (in nanomoles per milligram of protein) in old normal and cataractous proteins, respectively (Fig. 5 , middle). After 6 hours of incubation with ascorbate, CML-rich proteins from cataractous lenses oxidized 3 times more ascorbate than the precipitate from normal lenses (P < 0.01; Fig. 5 , bottom). Moreover, the oxidizing activity was linked to the CML-rich fraction in contrast to the supernatant, which had comparatively little activity (Fig. 6) . DTPA was highly effective at preventing ASA oxidation, thereby implicating transition metal(s) in the process (Fig. 5 , bottom). 

Finally, we sought to clarify the nature of the free radical formed during exposure of the cataractous proteins to H₂O₂ by trapping the radical formed with DMPO. Figure 7 shows the typical hydroxyl radical formed when Cu(II) is reacted with 500 μM H₂O₂. No such signal was detected in the homogenate after 15 minutes of reaction time. However, a signal was detected in the CML-rich immunoprecipitated crystallins, which was quenched by the addition of DTPA (Fig. 7) but not superoxide dismutase (not shown). 

Discussion

This study provides the first evidence of a close relationship between AGE formation and binding of copper by in vitro ascorbylated lens crystallins. It also demonstrates that highly pigmented crystallins from human cataracts are more redox active than crystallins from noncataractous and younger lenses and that the activity is linked to a CML-rich crystallin fraction and in part suppressible by a chelating agent. 

Carboxymethyl-Lysine as a Ligand for Redox Active Copper in the Lens

While this article was under editorial review, a related study by Ortwerth and James ⁴² showed that the addition of calf lens proteins to a mixture of copper and ascorbate completely suppresses its oxidation and free radical generation, leading to the conclusion that“ copper, even though it is increased in cataracts, would be strongly bound by lens proteins, where it cannot function in the oxidation of sugars.” Paradoxically, however, the two studies are not in contradiction! We too made the identical observation (not shown) that the addition of human lens proteins to a protein-free solution consisting of Cu²⁺ and ascorbate suppresses oxidative processes. Similar data were reported previously for albumin ^{43 44} and are in principle applicable to any protein whereby Cu²⁺ likely binds to the glyceryl structure of the peptide backbone. 

What, thus, is the relevance of our observations to the oxidative processes in the aging lens? From the data in Figure 4 , it is apparent that CML content at the highest ascorbate concentration used for in vitro ascorbylation was approximately 2.4 nmol/mg protein. However, because the increase in Cu bound was 30 nmol/mg protein, at most 10% of binding can be attributed to CML itself, and, thus, 90% of the copper was obviously bound to other sites, such as the peptide backbone. By contrast, in cataracts there was 1000-fold less Cu²⁺. However, there was an inverse stoichiometric ratio of 20 to 40 mol CML per mole of Cu, which suggests that some of these sites may indeed serve to bind Cu²⁺. 

Compared with a protein-free system, in which ascorbate oxidation occurs within minutes, the oxidizing activity of the human lens crystallins was slow and extended over several hours. It could thus be argued that it is not a relevant process. However, more ascorbate was oxidized than could be accounted for based on the stoichiometric amount of bound copper. Thus, either the oxidizing activity was due to other transition metals, such as iron, or whatever copper is bound must in part be capable of redox cycling, as recently demonstrated by us. ³² Such mechanism(s) could contribute to lipid peroxidation in cataracts, whereby a vicious cycle could exist ³² in which CML–protein–metal complexes oxidize lipids and release glycolaldehyde, which itself is a CML precursor. ⁴⁵ In that sense CML is a “molecular sink” for catabolites and processes that are insufficiently detoxified by the old and cataractous lenses. 

Other AGEs that May Bind Metals in the Lens

Other ligands, besides CML, may conceivably form coordination sites for redox active metals. For example, methylglyoxal is the precursor of CEL, a molecule which has structural homology with CML. It is the second most abundant modification in the aging human lens next to CML. ²⁰ Thus, this EDTA-like modification should also have metal-binding properties. Furthermore, whereas CML has been found to form in the lens and from ascorbate by several groups, ^{16 46 47} Pischetsrieder et al. discovered a novel modification resulting from protein ascorbylation (i.e., OMA). ⁴⁸ Immunoreactive OMA was recently detected in aging and cataractous crystallins. ²⁴ Of interest is the fact that this modification also has a carboxylic group, which could conceivably serve as a coordination site for transition metals. 

There are several sites in human lens crystallins where suitable coordination sites for metal binding to CML-modified lysines might exist. Such sites are rich in histidine and glutamic acid residues. For example, potential copper-binding sequences in human βA3-crystallins include 121HKE123, 129EKE131, and 192YKHW195. In the latter sequence, the proximity of tyrosine and tryptophan may lead to “in situ” metal-catalyzed oxidation of these residues. In humanβ B3-crystallins, lysine is found in a histidine-rich sequence: 113HHKLH117. 

Nature of the Redox Activity of Cataractous Crystallins

The data presented above implicate transition metals in ascorbate oxidation, whereby the precise extent to which Cu(II) or Fe(III) are involved, both in the in vitro ascorbylated samples (Fig. 4) and in the cataractous crystallins (Figs. 5 6 7) , remains to be determined. Indeed, the fact that mean Fe levels were identical in each group suggests, but does not prove, that Fe is not the redox active element, because Fe can be bound in redox active as well as redox inactive states. However, the inability of DTPA to completely prevent the oxidation of ascorbate suggests that other mechanisms of ascorbate oxidation are also operational. Because the degradation of ascorbate occurred in the incubator, in the absence of light, photooxidative destruction, as proposed by Linetsky et al., ⁴⁹ is unlikely. However, a recent study by Lee et al. points to the existence of non–metal-catalyzed oxidation of ascorbate by methylglyoxal-modified protein, which mimics metal-catalyzed oxidation. ⁵⁰ Such mechanism is very attractive in light of the recent work implicating methylglyoxal in lenticular aging. ¹⁴  

The data presented above and the data existing in the literature show that there are several modifications and mechanisms, in addition to CML, that could explain the redox activity of cataractous crystallins. Evidently, the single major element needed for oxidation to occur is O₂. Eaton ⁵¹ has pointed out that the lens should be considered “canned” (i.e., it protects itself from oxidation by having a very low O₂ tension, especially toward the nucleus). Indeed, exposure of whole animals to hyperbaric oxygen can induce cataracts. ⁵² However, even the concept that O₂ is necessary for oxidation to occur is challenged by Lee et al., who were able to reduce ferricytochrome c in absence of O₂ with methylglyoxal Schiff base protein adduct. ⁵⁰  

Based on the many observations showing that lens crystallin modifications by oxidative processes and carbonyl agents are dramatically increased in age-matched cataractous versus non-cataractous lenses, it is not difficult to speculate that a breakdown of the permeability barrier of the fiber cell must be a key event in cataractogenesis. Indeed, in previous experiments in which we exposed rat lens crystallins to high galactose stress in vitro or in vivo, we found the permeability barrier to 2,3-diketogulonate completely lost. ³⁶ This resulted in accelerated ascorbylation of the crystallins on exposure of the lens to ¹⁴C-ascorbate. 

Conclusions

One conclusion from this study is that chelation therapy may be helpful for assessing the role of protein-bound copper in senile cataractogenesis, and may even help delay the process. This suggestion has already grown out of previous work by Ou and colleagues who demonstrated aldose reductase inhibitors that delay sugar cataracts, such as sorbinil, have chelating properties. ⁵³  

 Supported in part by Grant EY07099 (VMM) and a P-30 EY11373 Core Grant to the Case Western Reserve University from the National Eye Institute. Poonam and Amit K. Saxena were recipients of a Fellowship from the Juvenile Diabetes Foundation International.

 Submitted for publication April 26, 1999; revised September 3 and December 1, 1999; accepted December 9, 1999.

 Commercial relationships policy: N.

 

Figure 1.

 

Effects of incubation of pooled dialyzed human lens homogenate (0.27 mg/300 μl) from young, old, and cataractous lens crystallins extracted in metal-free buffer on spontaneous oxidation of ¹⁴C-ascorbate (2 mM) after 1 and 2 days of incubation. Addition of the metal chelator DTPA (1 mM) markedly diminished the oxidation rate. The baseline (100%) corresponds to the amount of radioactivity in the ascorbate peak from nonincubated, fresh ¹⁴C₁-ascorbic acid.

Figure 2.

 

(A) Endogenous copper content of the pooled young, old normal, and cataractous crystallins that have been prepared as described in Figure 1 . (B) Copper binding by the same proteins on reverse dialysis against 500 μm CuCl₂ and removal of excess unbound copper by dialysis against metal-free buffer.

Figure 3.

 

Effects of crystallin ascorbylation on their ability to oxidize ascorbate. Lens crystallins from young lenses were incubated with 0.5, 5, and 20 mM ascorbate in 100 mM sodium phosphate for the times indicated and dialyzed extensively. They were incubated at a concentration of 3 mg/ml with 2 mM ¹⁴C-labeled ascorbate, and the extent of ascorbate oxidation was measured after 8 hours at 37°C (also see Fig. 1 ).

Figure 4.

 

Relationship between CML formation and copper binding in crystallins that were ascorbylated as described in Figure 3 . ASA, ascorbic acid.

Figure 5.

 

Protein (top) and carboxymethyl-lysine content (middle) and ascorbate oxidizing activity (bottom) of CML-rich crystallins from the WS fractions of young, old normal, and cataractous lenses after immunoprecipitation with a polyclonal CML antibody (also see Fig. 1 ).

Figure 6.

 

Ascorbate-oxidizing activity from WS cataractous lens crystallins before (homogenate) and after immunoprecipitation with CML antibody (precipitate) compared with activity in the supernatant (also see Fig. 1 ).

Figure 7.

 

Formation of DMPO spin-trap adduct from cataractous lens crystallins (from Fig. 6 ) after incubation for 15 minutes with H₂O₂ and DMPO. The signal was not suppressed by superoxide dismutase, thereby implicating hydroxyl radicals in the DMPO formation.

Human lenses were obtained from the National Disease Research Interchange (Philadelphia, PA). The authors thank Donita Garland and Ramanakoppa H. Nagaraj for supplying the brunescent lenses, Ram Subramaniam for assistance with the EPR experiments, and the University Medical Laboratory Foundation for assistance with atomic absorption spectroscopy measurements.