Protein-Bound Kynurenine Decreases with the Progression of Age-Related Nuclear Cataract

Abstract

purpose. Posttranslational modification by UV filters is a key event in human lenses that appears to be largely responsible for normal age-dependent yellowing. It has been proposed that subsequent reactions of these covalently bound UV filters may also be involved in the genesis of age-related nuclear cataract. To examine this hypothesis, the levels of kynurenine-lysine and kynurenine-histidine were measured in both normal and cataractous human lenses.

methods. Proteins isolated from the nuclei of normal lenses and lenses with and types I to IV nuclear cataract were hydrolyzed in 6 M HCl, and the levels of kynurenine-lysine and kynurenine-histidine were determined by HPLC.

results. The content of kynurenine-lysine and kynurenine-histidine decreased substantially with the progression of age-related nuclear cataract. On average, levels of both kynurenine adducts were four times lower in advanced cataract (type IV) than in normal lenses. Simple autoxidation of the derivatives did not appear to be responsible for this decrease, because incubation in the presence of oxygen or H₂O₂ did not affect adduct stability.

conclusions. Although protein-bound kynurenine accumulates over time in normal lenses, the levels attached to the proteins decrease significantly with the progression of age-related nuclear cataract. This finding suggests that in cataract there is a breakdown of the protein-bound adducts. Such further reactions of bound UV filters may contribute to the etiology of age-related nuclear cataract.

Old human lenses are yellow; the coloration being more pronounced in the center. This primate-specific coloration may be due principally to the covalent modification of crystallins by tryptophan-derived UV filters. This major posttranslational modification is especially notable after middle age and is caused by the formation of the barrier ¹ that allows greater time within the lens nucleus for the intrinsically unstable molecules to deaminate. ² Consequently, protein-bound 3-hydroxykynurenine glucoside (3OHKG) ³ and kynurenine (Kyn) ⁴ both increase significantly after middle age. 

The process of UV filter attachment is a dynamic one, in that not all adducts that are formed initially are stable at pH 7 and 37°C. Thus, in the older normal human lens, Kyn-histidine (Kyn-His) is predominant, with lower amounts of Kyn-lysine (Kyn-Lys), and little Kyn-cysteine (Kyn-Cys). Cysteine residues of the crystallins appear to be a preferred site for binding; however, Kyn-Cys is unstable under physiological conditions. ⁵  

Although normal lenses become increasingly yellow with age, age-related nuclear (ARN) cataract is characterized by more pronounced coloration of lenses, ranging from orange-brown to black. Proteins isolated from such lenses are colored, oxidized, and insoluble. ^{6 7 8 9 10} Because UV filters such as 3-hydroxykynurenine (3OHKyn) oxidize readily to form a range of colored products, ^{11 12 13} it has been postulated that the characteristic coloration associated with ARN cataract may be explained as a result of the further reaction (oxidation) of the bound UV filters. ^{14 15 16 17} If this hypothesis were true, it is possible that the levels of Kyn-amino acids actually decrease with the onset of cataract, despite the fact that the content of free reduced glutathione (GSH), which acts to inhibit binding of UV filters, ¹⁸ in these cataractous lenses is lower than in normal eyes. ¹⁹  

In this study, proteins in normal lenses and in lenses with types I to IV ARN cataract were examined for the content of Kyn-amino acids. 

Materials and Methods

Materials

All organic solvents and acids were HPLC grade (Ajax, Auburn, New South Wales [NSW], Australia). Ultrapure water (purified to 18.2 MΩ/cm²; Milli-Q; Millipore, Bedford, MA) was used in the preparation of all solutions. Amino acids (N-α-t-Boc-l-histidine, N-α-t-Boc-l-lysine, and l-cysteine) and Kyn-sulfate were all obtained from Sigma-Aldrich (St. Louis, MO). Sequencing grade HCl (6 M) was purchased from Pierce (Rockford, IL). Hydrogen peroxide (3% wt/wt) was purchased from Sigma-Aldrich. 

Lenses

Normal human lenses were obtained from the Sydney Eye Bank (NSW, Australia) or from the National Disease Research Interchange (Philadelphia, PA) with ethics approval from the University of Wollongong Human Ethics Committee (HE99/001). Normal lenses were 70 to 80 years old. The cataractous lenses used in this study were removed from patients attending eye camps in India and were scored, labeled, and stored frozen. ⁶ The ages of the patients were not recorded. Lenses were classified as types I to IV based on the classification system of Pirie. ²⁰ Lenses were stored at −20°C until processed. 

Isolation of Human Lens Proteins

The methods used for isolation of lens proteins, acid hydrolysis, and subsequent analysis by HPLC were performed as previously described. ⁴ The methods are summarized briefly below. Individual lenses were placed on a glass plate on dry ice, and the nucleus was cored (using a 6-mm cork borer) and separated from the cortex. The poles of the nucleus were removed by dissection. The nucleus was then homogenized in absolute ethanol. After cooling for 1 hour at −20°C the homogenate was centrifuged for 20 minutes at 14,000 rpm. The supernatant liquid was removed, and the pellet was reextracted in 80% ethanol and again centrifuged. The supernatant was discarded, and the pellet lyophilized and weighed. 

Acid Hydrolysis of Lens Protein and Kyn-Amino Acid Adducts

Lens protein samples (∼10 mg) were hydrolyzed with 6 M HCl (1 mL) for 24 hours at 110°C in an evacuated hydrolysis tube. After hydrolysis, the samples were lyophilized overnight and then each dissolved in 400 μL of 0.1 M NaH₂PO₄ and 200 μL of 1 M Na₂HPO₄ (pH ∼5). The solutions were then examined by reversed-phase HPLC (RP-HPLC). The results have been corrected for the recovery of Kyn-Lys (96%) and Kyn-His (99%), as determined by acid hydrolysis of the authentic standards prepared as described, ⁴ and then quantified using HPLC by reference to the relevant standard curve. 

Kyn-Amino Acid Adduct Stability Studies

The t-Boc protected Kyn-amino acid adducts, Kyn-t-Boc-His and Kyn-t-Boc-Lys, were prepared as previously described. ⁴ Each standard was dissolved in 0.2 M phosphate buffer (pH 7.2), to obtain a concentration of 1.0 to 2.5 mM. Solutions were bubbled with argon or oxygen, and each incubation vial was sealed, wrapped in foil, and incubated at 37°C for up to 100 hours. Triplicate samples were taken every 20 hours. The incubation of adducts in the presence of hydrogen peroxide was performed as just stated with the addition of hydrogen peroxide to achieve a final concentration of 2.0 mM. 

High-Performance Liquid Chromatography

RP-HPLC of acid hydrolyzed samples was performed on an HPLC system (Gold; Beckman-Coulter, Fullerton, CA) equipped with a solvent module (127S) and a UV-Vis detector (model 166; both from Beckman-Coulter). For analytical separations a column (Microsorb-MV C-18 [1 nm, 5 μm, 4.6 × 250 mm]; Varian, Sunnyvale, CA) column was used with the following mobile phase conditions: solvent A (aqueous 4 mM ammonium acetate, pH 6.5) for 5 minutes followed by a linear gradient of 0% to 50% solvent B (80% acetonitrile/H₂O, 4 mM ammonium acetate) over 20 minutes, followed by a linear gradient of 50% to 100% solvent B over 15 minutes. The flow rate was 1 mL/min. The identities of the Kyn-Lys and Kyn-His in the samples were confirmed using liquid chromatography-mass spectrometry (LC/MS). ⁴ RP-HPLC of samples from Kyn-amino acid adduct stability studies was performed on either of two systems (a Knauer [Berlin, Germany], consisting of two Knauer pumps, a model 7125 sample injector [Rheodyne, Rohnert Park, CA] and a Knauer UV-vis detector, monitoring at 360 nm; or model LC-10A; Shimadzu, [Columbia, MD] HPLC system, consisting of two pumps, a high-pressure mixer, an autosampler, and a UV/Vis photodiode array detector). System operation and peak integration were performed using Class VP software (Shimadzu). In both cases, separations were performed with an acetonitrile/H₂O gradient in 0.05% (vol/vol) trifluoroacetic acid (TFA) at 1 mL/min. The percentage of acetonitrile used in the gradient was as follows: 0% (5 minutes), 0% to 40% (50 minutes), and 40% to 0% (5 minutes). 

Results

Proteins isolated from the nuclei of normal and cataractous human lenses were analyzed by HPLC after acid hydrolysis, ⁴ to determine the degree to which the crystallins were modified by Kyn. The analysis of human lens protein digests from individuals over a wide age range has shown that the pattern of Kyn modification of the amino acid residues is fairly consistent: the amount of Kyn-His was always higher than that of Kyn-Lys. ⁴ Kyn-Cys typically represented less than 5% of the total amount of Kyn-modified amino acids. There was, however, a significant increase in the total amount of Kyn bound to the lens proteins with age. 

Cataractous lenses, graded on the basis of nuclear color as described by Pirie, ²⁰ were examined in the same way. The quantities of Kyn-His and Kyn-Lys in the nuclear cataract protein samples, plotted according to lens type are shown in Figures 1A and 1B . Because of the low levels present, Kyn-Cys was not examined. Four lenses were analyzed for each type of cataract, and analyses were performed in duplicate. In addition, four normal lenses from donors of an age similar to that of typical patients with cataract (70–80 years) were examined. Only the nuclear protein from the lenses was analyzed, owing to limited availability of cortical tissue from the cataractous lens samples. 

As expected, some variation in the data was observed; however, both Kyn-His and Kyn-Lys displayed very similar trends across the lens cataract types analyzed. The mean Kyn-His level fell by approximately fourfold, from an initial mean level of approximately 2 nmol/mg of protein in type I cataractous lenses to 0.5 nmol/mg of protein in type IV cataractous lenses (Fig. 1A) . Type I lenses were removed from patients as a result of cortical cataract and were expected to have Kyn adduct values similar to those of aged normal lenses. This was found to be the case. 

The Kyn-Lys level fell from an initial mean value of 0.11 nmol/mg of protein in type I cataract to 0.03 nmol/mg of protein in type IV cataract—again, a fourfold loss (Fig. 1B) . As was found to be the case in the nuclei of normal lenses, ⁴ the level of Kyn-His observed was approximately 20 times greater than that of Kyn-Lys. 

Statistical analysis of data sets using the Kruskal-Wallis (Wilcoxon) test indicated that there was a significant difference in the median value between lens types (P = 0.0056 and 0.0021 for Kyn-His and Kyn-Lys, respectively). 

It was considered possible that the decrease in the levels of Kyn-Lys and Kyn-His with the worsening of nuclear cataract may simply reflect exposure within the lens to a more oxidizing environment. It is clear that this is the case based on numerous indicators such as the loss of GSH, ^{7 19} and oxidation of protein methionine ²¹ and cysteine, ¹⁹ as well as aromatic and aliphatic residues. ²² To test this, authentic t-Boc protected Kyn-Lys and Kyn-His standards were exposed to both oxygen and H₂O₂. Figures 2A and 2B show that, whereas Kyn-His was stable to extended incubation at 37°C, approximately 46% of Kyn-Lys remained after 80 hours. Further, the presence of oxygen made no observable difference in the time course for the loss of either amino acid adduct. Addition of H₂O₂ to the incubation also did not appear to affect the stabilities of either Kyn-His or Kyn-Lys (Fig. 2C) . 

Discussion

This study was conducted to determine whether the levels of Kyn attached to proteins from lenses with nuclear cataract are significantly different from those in normal lenses. Before undertaking the experiments, it was recognized that two competing factors might influence the outcome. 

It was considered that the levels of protein-bound Kyn may be higher in lenses with ARN cataract because the levels of GSH in the nuclei of such lenses are very low, ^{23 24} and it has been demonstrated that GSH is the primary defense agent of the lens, in that it competes for the reactive unsaturated ketone intermediate produced by Kyn decomposition. If GSH levels are low, this ketone binds to proteins. ¹⁸ Alternatively, if the Kyn-amino acid adducts that are known to be present in older normal human lenses were involved in further reactions or were unstable in the more oxidative environment associated with ARN cataract, then it may be that the levels found could be significantly lower. 

The results clearly demonstrate that there is a significant progressive decrease in the content of both Kyn-His and Kyn-Lys with the progression of ARN cataract. We do not as yet know the reason for this, but, on the basis of adduct stabilities to oxygen and H₂O₂, it seems that further reactions of the Kyn adducts that cannot readily be reproduced by incubation in vitro may be a plausible explanation. 

Several factors must be considered, however, in interpreting these results. It may be possible that UV filter metabolism is different in cataract than in normal lenses and that formation of Kyn and its deamination does not occur to the same extent. This seems unlikely because the activity of the UV filter synthetic pathway is relatively independent of the age of the lens. ²⁵ There is evidence, however, that some UV filters may be present in cataractous lenses but not in normal lenses. For example, xanthurenic acid glucoside has been isolated from lenses with advanced nuclear cataract. This UV filter is predicted to be formed by deamination of 3-hydroxykynurenine glucoside followed by intramolecular condensation and oxidation. ²⁶ Such processes can take place in the absence of GSH, as occurs in advanced nuclear cataract cell nuclei. ¹⁹ It is also possible that components specific to the cataractous lens may intercept the reactive intermediates. Unless lenses from cataractous eyes have a UV filter metabolic pathway that is substantially different from that in normal eyes, one would expect the levels of Kyn-His present in these lenses before cataract to be quite high, because cataractous lenses are usually extracted from older people. Because intact cataractous lenses are now almost impossible to obtain because of changes in surgical methodology, it is difficult to test these hypotheses. 

The same considerations do not apply for the other Kyn adduct Kyn-Lys, because it is intrinsically less stable under conditions of physiological temperature and pH (Fig. 2B) . Measured levels in normal lenses represent a static image of competing Kyn-Lys synthesis and breakdown pathways. Therefore, the levels of Kyn-Lys could decrease in cataractous lenses, not because of chemical breakdown, but simply because of a reduced synthesis of UV filters in cataractous lenses or, alternatively, if the reactive unsaturated ketone intermediate were somehow prevented from binding to the proteins (e.g., by a less-permeable barrier in cataractous lenses). This consideration is not likely in the case of Kyn-His, because it is stable, and the decrease in its content, together with the fact that it is progressive and mirrors that of Kyn-Lys, tends to support the hypothesis that there is breakdown of preexisting bound UV filters in these lenses. 

Unfortunately, one cannot be definitive, because the cataractous lenses used in this study were sourced from India, and the ages of the patients were not recorded. It is now established that the levels of protein-bound Kyn increase as a function of age. ⁴ It is possible, therefore, that the phenomenon observed in Figures 1A and 1B is merely a reflection of a chance phenomenon—that the average age of patients with type IV cataract was younger than that of those with type III, which in turn was younger than that of those with type II. That this occurs by chance seems unlikely, but the possibility cannot be eliminated at this stage. It is also possible that Kyn-modified proteins are selectively degraded in cataractous lenses, although protease activity in the nuclear region is low. Selective protein degradation also seems an improbable explanation for the results obtained. 

If the significant, progressive decrease in protein-bound Kyn that we observed with advancing stages of cataract reflects degradation of preexisting Kyn, we can say little about the details of the mechanism, except that it is unlikely that simple autoxidation of the Kyn adducts is responsible. Development of ARN cataract is associated with a change to a more oxidative environment in the lens nucleus; however, incubation of the Kyn-His and Kyn-Lys adducts in the presence of oxygen did not lead to an accelerated breakdown of these molecules (Figs. 2A 2B) . Treatment of the adducts with H₂O₂ also did not affect stability. Other more potent agents that have been implicated in ARN cataract—for example, hydroxyl radical, ²² metals, ^{27 28 29 30} or photo-oxidation ^{31 32} —may be involved. 

In agreement with the findings in the current study (i.e., that there is a progressive decrease in the content of covalently bound Kyn as a function of nuclear cataract), a recent study using total enzyme digestion of cataractous lens proteins has also demonstrated a progressive decrease in the content of free Kyn in the digests of types I to IV lenses by HPLC (Ortwerth BJ, personal communication, May 2003). It is assumed that the free Kyn is derived from the breakdown of covalently bound Kyn (e.g., Kyn-Lys), during the course of extended proteolysis. 

There are undoubtedly a number of posttranslational modifications that occur in normal lenses with age. Several of these have been documented and include pentosidine protein cross-links and glucosepane adducts. ^{33 34} If such modifications play a role in subsequent cataract formation, one may expect the levels to alter progressively as a function of cataract progression, as has been clearly demonstrated for the content of methionine sulfoxide, ²¹ protein sulfhydryls, ¹⁹ and hydroxylated amino acid derivatives. ²² The fact that the levels of both Kyn-His and Kyn-Lys decrease in the proteins from types I to IV lenses suggest that indeed Kyn modification may indeed play a role in cataractogenesis. The task now is to determine the reasons for the decline and the roles of any decomposition products in, for example, lens coloration, cross-linking, and insolubilization—features that are the hallmark of ARN cataract. 

 Supported in part by Grant 162707 from the National Health and Medical Research Council (NHMRC) and by funds from the Australian Research Council, Ramaciotti Foundation, and the University of Wollongong, enabling the purchase of the mass spectrometer.

 Submitted for publication June 4, 2003; revised September 23, 2003; accepted October 20, 2003.

 Disclosure: S. Vazquez, None; N.R. Parker, None; M. Sheil, None; R.J.W. Truscott, 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.

 

Nuclear levels of Kyn-modified His (A) and Lys (B) residues in aged normal lenses and in cataractous lenses. Error bars: standard deviation (n = 8). Insets: results of Kruskal-Wallis analysis of the same data sets.

Figure 2.

 

Concentration of Kyn-t-Boc-His (A) and Kyn-t-Boc-Lys (B) remaining after incubation under physiological conditions (pH 7.2, 37°C), in the presence of oxygen or argon. Effect of an equimolar concentration of hydrogen peroxide on the stability of Kyn-t-Boc-His and Kyn-t-Boc-Lys under physiological conditions (C).