Argpyrimidine, a Blue Fluorophore in Human Lens Proteins: High Levels in Brunescent Cataractous Lenses

Abstract

purpose. To determine whether the human lens contains argpyrimidine, a modification of arginine by methylglyoxal, to establish how argpyrimidine content relates to lens aging and cataract formation.

methods. A monoclonal antibody was used to measure argpyrimidine by a competitive ELISA in water soluble (WS) and insoluble (WI) lens fractions from young, aged, nuclear cataractous, and brunescent cataractous lenses. Brunescent cataractous lens proteins were digested by enzymes, the digest was subjected to HPLC, and the eluate was analyzed for argpyrimidine. Lens proteins from aged lenses (from donors 65 to 80 years of age) were fractionated on a Sephadex G-200 column, and the crystallins were tested for argpyrimidine.

results. The competitive ELISA showed two to three times as much argpyrimidine in water-insoluble proteins as in water-soluble proteins. Although no clear cut increase with the age of the lens donors in either the water-soluble or the insoluble protein fractions was found, the argpyrimidine levels in brunescent cataractous lenses were significantly higher (254.0 ± 155 pmol/mg protein, P < 0.005) than in age-matched, aged (16.1 ± 8 pmol/mg) or nuclear cataractous lenses (49.0 ± 26 pmol/mg). Lenses from diabetic individuals showed a modest increase (50.3 pmol/mg) compared with age-matched normal lenses. HPLC results provided additional evidence that human lenses contain argpyrimidine. Western blotting experiments showed consistently stronger reactions with cataractous lens proteins than those from noncataractous lenses, and argpyrimidine was found in both crystallin monomers and polymers. All crystallins and several cross-linked high-molecular-weight aggregates reacted with the antibody to argpyrimidine, but a protein of∼ 28 kDa in the α-crystallin fraction displayed the greatest immunoreactivity.

conclusions. Methylglyoxal modifies arginine within the human lens, and the changes occur at a much higher rate in brunescent lens proteins than in either nuclear cataractous or normal lenses. All crystallins contained argpyrimidine and covalently cross-linked aggregates. This is the first report of immunologic evidence for an arginine modification in the human lens by a physiologically important α-dicarbonyl compound.

Lens proteins undergo numerous physicochemical changes during aging and cataractogenesis. ^{1 2} The Maillard reaction is one mechanism implicated in such changes. Aldehydes and ketones react nonenzymatically with amino groups on proteins to create irreversible terminal products that are known as advanced glycation end products (AGEs). ^{3 4} Several different AGEs have been detected in the human lens. ^{5 6 7} The lens contains large amounts of ascorbate (vitamin C), which, along with glucose, appears to be a major AGE precursor. ^{6 8 9 10} Various changes of lens aging and cataractogenesis, such as lens protein pigmentation, formation of nontryptophan fluorophores, covalent cross-linking, and formation of reactive oxygen species, are thought to be due to AGEs. 

Recent studies suggest that α-dicarbonyl compounds, methylglyoxal (MG) and glyoxal (GXL), also produce AGEs in lens proteins. ^{11 12} AGEs that are derived from these two compounds include MG-lysine dimer (MOLD or imidazolysine) ¹³ carboxyethyllysine (CEL) ¹⁴ and glyoxal-lysine dimer (GOLD) ¹⁵ ; these products, which derive from the reaction of protein-lysine residues with α-dicarbonyl compounds, are found in human lens proteins. 

α-Dicarbonyls have a high propensity to react with arginine residues on proteins, but as yet, no arginine modification was detected in human lenses. We previously reported a MG-derived blue fluorescent arginine modification, which we designated argpyrimidine (N-δ-(5-hydroxy-4,6-dimethylpyrimidine-2-yl)-l-ornithine; Fig. 1 ). We showed that it formed in lens proteins incubated with MG. ¹⁶ Oya et al. ¹⁷ studied the reaction of MG with arginine and showed that argpyrimidine is one of the major fluorescent products of this reaction. They developed a monoclonal antibody against a MG-modified protein that reacted with argpyrimidine on MG-modified proteins. We used this antibody to detect argpyrimidine in the human lens and to determine how the amounts of argpyrimidine relate to lens aging and cataractogenesis. 

Materials and Methods

The following were from Sigma Chemical Co. (St. Louis, MO): bovine serum albumin (BSA), dithiothreitol (DTT), pronase E (from Streptomyces griseus), leucine aminopeptidase M (type IV-S from porcine kidney microsomes), carboxypeptidase A (type II, from bovine pancreas), N-α-acetyl lysine (AL), N-α-acetyl arginine (AA), MG 40%, glucose, and ribose. MG was purified by vacuum distillation. Sephadex G-200 was from Pharmacia (Piscataway, NJ). Human lenses were obtained from the Cleveland Eye Bank (Cleveland, Ohio) and from the National Disease Research Interchange (Philadelphia, PA). Bovine lenses were from Pel-Freeze Biologicals (Rogers, AR). The argpyrimidine monoclonal antibody was developed in Koji Uchida’s laboratory at Nagoya University Graduate School of Bioagricultural Sciences, Japan. Polyclonal antibodies against α-crystallins and a monoclonal antibody againstβ _H-crystallin were obtained from StressGen (Sidney, British Columbia, Canada). Anti-γ crystallin antibody was a gift from Samuel Ziegler (National Eye Iinstitute, Bethesda, MD). 

Extraction of Lens Proteins

Each lens was homogenized in 2.0 ml phosphate-buffered saline (PBS) with a motor-driven homogenizer and centrifuged at 20,000g for 30 minutes at 4°C. The supernatant was designated as the water-soluble (WS) and the pellet as the water-insoluble (WI) fraction. The WS fraction was dialyzed for 24 hours against PBS, protein was estimated, and the sample was tested in a competitive ELISA. Five milligrams of WI protein was sequentially digested by proteases as described earlier, ⁸ and the digest was filtered through 10K MWCO filters (Millipore, Bedford, MA). Amino acid content of the filtrate was estimated as previously described. ¹⁸ One micromole of amino acids was used in the competitive ELISA (see below). 

Incubation of AL, AA Sugars, and Ascorbate

All incubations were done in 0.2 M potassium phosphate buffer (pH 7.4) in a total volume of 3.0 ml at 37°C. All samples were filtered through a 0.2-μm filter before incubation. Fifty millimolar AL or AA was incubated with one of the following: glucose (50 mM), ribose (25 mM), ascorbate (25 mM), or MG (10 mM). In some experiments, 10 μM FeCl₃ was added. Aliquots were drawn on day 5 and 30 and tested for argpyrimidine by competitive ELISA. 

Competitive ELISA for Argpyrimidine

Microplate wells were coated by overnight incubation with 0.1μ g BSA-argpyrimidine in 0.05 M sodium carbonate buffer (pH 9.7). After washing three times with PBS containing 0.05% Tween-20, the wells were blocked by incubating with 5% nonfat dry milk for 2 hours. The samples to be tested were preincubated with the antibody (dilution 1:80,000) for 1.5 hours and then added to the wells in quadruplicate. The plates were then incubated for 2 hours at room temperature. The supernatant medium was discarded and the wells were washed three times with PBS-Tween. Then horseradish peroxidase (HRP)-conjugated goat anti-mouse IgG (1:80,000 dilution) was added, and incubation was continued for 2 hours at room temperature. The wells were washed four times with PBS-Tween and incubated with 3,3′,5,5′-tetramethylbenzidine as substrate. We used protein in amounts that corresponded to 0.1 mg in the competitive ELISA. Argpyrimidine was measured by comparing the test sample with an inhibition curve generated from synthetic argpyrimidine. The lower limit of detection in the assay was 0.5 pmol. 

HPLC Detection of Argpyrimidine in Brunescent Cataractous Lens Proteins

WI protein from brunescent cataractous lenses (8.3 mg protein) was hydrolyzed by proteolytic enzymes ⁸ and injected onto a Vydac RP C₁₈ column (The Separations Group, Hesperia, CA) with MeOH/HFBA solvent system (solvent A, 100% water; solvent B, 70% MeOH, both with 1.2 ml HFBA/l, gradient 80/20 to 65/35 in 35 minutes). Argpyrimidine-containing fractions (eluting between 24 and 27 minutes) were collected, dried, suspended in water, and reinjected onto a Knauer RP C₁₈ column using a propanol/SDS solvent system (solvent A, 100% water with 1.13 g NaH₂PO₄0.2H₂O/l; solvent B, 60% propanol with 1.0 g NaH₂PO₄0.2H₂O/l, both with 3 g SDS/l, adjusted to pH 6.5, B adjusted to 7.0, gradient 98/2 isocratic for 15 minutes). The retention time of argpyrimidine was determined by injecting 1.17 pmol of argpyrimidine standard. One-milliliter fractions were collected, lyophilized, and dissolved in 0.5 ml of 0.2 N ammonium formate (pH 6.75; buffer A) and passed through a Dowex 50W-X4 (Bio-Rad, Hercules, CA) cation exchange column using buffer A as the eluant. Fractions of 1.0 ml were collected, lyophilized, and dissolved in 100 μl of PBS. A 30-μl sample of each fraction was tested for argpyrimidine in a competitive ELISA. 

Column Chromatography of Human Lens Crystallins

Human lens crystallins were isolated by column chromatography as described by Ziegler et al., ¹⁹ with slight modifications. Four normal lenses from donors of 65 to 80 years of age were homogenized in 2 ml PBS. After centrifugation at 20,000g for 30 minutes, the supernatant was collected and dialyzed against 2 L of Tris-HCL buffer (pH 7.4) containing 0.5 M NaCl, 0.001 M EDTA, 0.02% NaN₃. The dialysate was concentrated by lyophilization, suspended in 4.0 ml buffer, and loaded onto a Sephadex G-200 column (100 × 1.5 cm) equilibrated in same buffer. Fractions corresponding to α-, β₁-,β ₂-_,β ₃-, and γ-crystallins were pooled, concentrated by ultrafiltration, and dialyzed against 4.0 L of PBS for 48 hours and lyophilized. 

SDS-PAGE and Western Blotting

The proteins were separated by one-dimensional SDS-PAGE by the method of Laemmli et al. using a 12% gel. The proteins in the WI material were solubilized by sonication as described by Ortwerth et al. ²⁰ Proteins were electrophoretically transferred onto a nitrocellulose membrane and blocked in 5% nonfat dry milk for 2 hours. The membrane was then exposed to the primary antibody at 1:5000 dilution. Argpyrimidine immunoreactivity was detected using an HRP-labeled secondary antibody (anti-mouse Fab specific; Sigma) with an ECL detection system (Pierce, Rockford, IL). 

Statistics

Differences in means were evaluated by one-way ANOVA using Stat View Software (SAS Institute, Inc., Cary, NC). We applied Fisher’s protected least significant difference test in these calculations and considered P < 0.05 to be statistically significant. 

Results

We measured the formation of argpyrimidine from ascorbate and sugars to determine whether they served as precursors. After incubation of samples for up to 30 days with AL or AA, we detected argpyrimidine in those incubations that had AA. Glucose alone failed to generate argpyrimidine, even after 30 days of incubation (Table 1) , but in samples containing both glucose and 10 μM FeCl_3, we detected up to 66 pmol/ml of reaction mixture, suggesting MG production from autoxidation of glucose. In contrast, ribose alone produced argpyrimidine. Ascorbate generated nearly 10 times as much argpyrimidine as did equimolar concentration (25 mM) of ribose. However, these levels were far lower than those obtained with 10 mM MG. As expected, incubations with AL did not produce argpyrimidine, because argpyrimidine is synthesized from arginine. Together these results indicate that ascorbate may be one of the precursors for MG in the human lens and that it may contribute to MG-mediated modification of lens proteins. 

We showed that our monoclonal antibody is specific for argpyrimidine and that it requires 4,6 dimethyl 5-OH moiety for reaction (unpublished observations). We initially tested the enzyme-digested WI fraction of brunescent cataractous lenses in a competitive ELISA to establish that human lens proteins contain argpyrimidine. As can be seen in Figure 2 , the antibody reacted with increasing intensity as the protein concentration (amino acids) increased; maximal binding occurred with 2.8 μmol of amino acids. The immunoreactivity with lens proteins could be completely abolished by preincubating the antibody with 4000 pmol of synthetic argpyrimidine (not shown). These results confirm that the lens protein immunoreactivity is due to argpyrimidine. 

To further confirm this, we attempted HPLC separation of argpyrimidine from brunescent lens WI proteins. Proteins were digested by proteolytic enzymes and subjected to HPLC (Fig. 3 , trace 2). The column eluate was collected and reinjected. One-milliliter fractions were collected, and detergent removed by chromatography on Dowex cation exchange resin and tested for argpyrimidine by ELISA. Figure 3 shows a well-resolved peak from the lens protein digest that corresponded to the argpyrimidine standard (trace 1). The coincidence of immunoreactivity in the lens fractions and the HPLC peak shows that the substance in human lens proteins is argpyrimidine (Fig. 3 , trace 3). 

Once we established that our lens preparations contained argpyrimidine, we wanted to determine the effect of donor age and disease processes, such as cataract formation and diabetes. We selected lenses from young donors (13–17 years), older donors (65–80 years), and from individuals with diabetes (65–80 years) as well as those with nuclear cataracts (55–75 years) and brunescent cataracts (65–80 years). We separated the lens proteins into WS and WI fractions. The WS fractions were used directly in the competitive ELISA, whereas the WI fractions were assayed only after digestion with proteolytic enzymes. The WS fractions had considerably lower levels of argpyrimidine than the WI fractions, which contained <40 pmol/mg protein (Fig. 4) . Aged lenses (from older donors) had the highest levels (39.0 ± 20.0 pmol). However, the WI fractions of nuclear cataractous and brunescent cataractous lenses contained relatively high levels. Nuclear cataractous lenses had approximately three times as much argpyrimidine as aged lenses, but brunescent lenses had, by far, the most, with 254.0 ± 155.0 pmol/mg protein. The argpyrimidine levels in diabetic lenses (50.3 ± 23.6 pmol/mg) approximated those of nuclear cataractous lenses (49.0 ± 26.0 pmol/mg). 

The Western blotting results showed immunoreactivity in several proteins of nuclear and brunescent cataractous lenses, with the immunoreactivity considerably higher in the brunescent lens than in the nuclear cataractous lens. (Fig. 5) . Most of the immunoreactivity was associated with high-molecular-weight protein aggregates. However, two discrete protein bands (∼30 and ∼50 kDa) could be seen in extracts from nuclear cataractous lenses. Pretreatment (1 hour at 37°C) of the antibody with 4000 pmol of argpyrimidine abolished the immunoreactivity suggesting that the reaction of antibody with lens proteins was due to argpyrimidine. 

To determine which crystallin is more susceptible to argpyrimidine modification, we fractionated proteins from aged lenses on a Sephadex G-200 column. The fractions were divided into α-,β ₁-, β₂-,β ₃-, and γ-crystallins, based on UV chromatography, and were used in Western blotting. All crystallins were immunoreactive (Fig. 6) . The α-crystallin fraction displayed an intense immunoreactivity associated with a minor protein that had a molecular weight higher than the 20-kDa α-crystallin (∼28 kDa). Unlike α-crystallins, monomers of β-crystallin subspecies were immunoreactive as well as the cross-linked high-molecular-weight aggregates. In general, proteins around 43 kDa reacted strongly with the antibody. 

To determine the nature of 28-kDa protein, we performed Western blotting with anti–α-crystallin antibodies. Our results showed weak immunoreactivity of 28-kDa protein with anti–α-crystallin antibodies (Fig. 7) . Incubation of purified α-crystallins with 25 mM MG did not generate this protein. This suggests that the 28-kDa protein in the human lens has other modifications in addition to argpyrimidine. 

Discussion

We wanted to determine whether the human lens contained argpyrimidine, a fluorescent product derived from the reaction of MG with arginyl residues on proteins, and if so, whether the argpyrimidine content of lenses had any relation to lens aging and cataract formation. Our study was prompted by two major observations. First, human lenses contain MG in concentrations at least 20 times higher than those in plasma. ²¹ Second, our own studies and those in other laboratories showed that MG modifies lysine residues on lens proteins and that cataractous lenses accumulate more modifications than age-matched normal lenses. ^{11 15 22} Because α-dicarbonyl compounds are highly reactive with guanidino group of arginine, we suspected that MG modifies arginyl residues in lens proteins as well. 

Argpyrimidine is only one of the three known arginine modifications of MG. The other two modifications, N-δ-(4-carboxy-4,6-dimethyl-5,6-dihydroxy-1,4,5,6-tetrahydropyri-midine-2-yl)-l-ornithine and N-δ-(5-hydro-5-methyl-4-imidazolon-2-yl)-l-ornithine, ¹⁷ have not yet been detected in tissues. A monoclonal antibody was raised against MG-modified KLH ¹⁷ and shown to react with argpyrimidine. We established that this antibody was specific for argpyrimidine and that it could react with argpyrimidine-modified proteins, specifically with the 4,6 dimethyl-5-OH structure of argpyrimidine. We then used this highly specific antibody to probe argpyrimidine in the human lens and to establish its relation to lens aging and cataractogenesis. 

In addition to the immunochemical evidence, we used HPLC to provide independent chemical evidence for argpyrimidine in the human lens. Our studies showed that the WS proteins had very little argpyrimidine in normal and cataractous lenses, whereas the WI proteins had considerably higher levels. This is not surprising, because the WI proteins are derived mostly from the older nuclear region of the lens, whereas the WS fraction is derived mostly from the relatively younger cortical region. AGEs and other chemical modifications accumulate with age in lens proteins. Some of these modifications presumably cause cross-linking and aggregation of lens proteins. As the lens ages, these modified proteins are pushed to the central nuclear region. Argpyrimidine modifications probably occur concurrently with other modifications and thus contribute to high levels in water-insoluble fraction in brunescent cataractous lenses. 

Our observation that ascorbate could form argpyrimidine suggests that it could be a precursor for MG in the lens, and it confirmed our previous chromatographic data showing formation of argpyrimidine from ascorbate. ¹⁶ The human lens contains relatively large concentrations of ascorbate (up to 2 mM). ²³ In the inner nuclear region of the lens, which has the highest concentrations of argpyrimidine, glycolytic intermediates are probably not the source of MG. The fiber cells of the inner nuclear region lack organelles and exhibit low oxygen tension with little or no metabolism. This may be sufficient for ascorbate oxidation. In fact, ascorbate oxidation products have been detected in the lens. ^{23 24} Thus, it is conceivable that argpyrimidine is produced from ascorbate, at least in the nuclear region. However, we cannot rule out argpyrimidine formation from MG that is directly derived from glycolytic intermediates, because new cortical proteins would be exposed to products of glycolysis in the epithelial layer. These proteins are gradually pushed inward to the center of the lens during aging. 

An obvious question is whether argpyrimidine modification contributes significantly to chemical modification of lens proteins in aging and cataractogenesis. We find that normal aged lenses have much less argpyrimidine than brunescent cataractous lenses in WI fractions of lens protein (16 vs. 253 pmol/mg). Human lens α-crystallin has 13 arginine residues per molecule. If the molecular weight ofα -crystallin monomer is 20 kDa, our values correspond to only 0.002% and 0.04% modification of arginine residues in aged and brunescent lenses, respectively. However, argpyrimidine levels are considerably higher than pentosidine, a lysine–arginine cross-linking AGE in lens proteins. ^{6 25} Argpyrimidine, itself, may not cause major structural changes in lens proteins, and it is only one of many products generated by the reaction of MG with arginine, but it is a useful bio-marker for MG-mediated arginine modifications in the lens and other tissues. 

 Supported by National Eye Institute Grants EY-09912 (RHN) and P-30 EY11373 to Case Western Reserve University Visual Science Center and grants from Research to Prevent Blindness, New York, and Ohio Lions Eye Research Foundation.

 Submitted for publication September 27, 2000; revised January 10, 2001; accepted February 7, 2001.

 Commercial relationships policy: N.

 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.

 

Two molecules of methylglyoxal react with an arginyl residue on proteins to form argpyrimidine.

Table 1.

 

Formation of Argpyrimidine from Sugars and Ascorbate

Incubation	5 Day	30 Day
AL+ glucose (50 mM)	0*	0
AL+ glucose+ FeCl3 (10 μM)	0	0
AA+ glucose	0	0
AA + glucose+ FeCl3	54.1	66.6
AL+ ribose (25 mM)	0	0
AA+ ribose (25 mM)	70.8	120.8
AL+ ascorbate (25 mM)	0	0
AA+ ascorbate (25 mM)	591.6	787.5
AL+ methylglyoxal (10 mM)	0	0
AA+ methylglyoxal (10 mM)	8112.3	11463.3
AL (50 mM)	0	0
AA (50 mM)	0	0

*  Argpyrimidine (pmol/ml reaction mixture). N-α-acetyllysine (AL) or N-α-acetylarginine (AA; 50 mM each) was incubated with the indicated carbohydrate at pH 7.4 and 37°C and assayed for argpyrimidine by a competitive ELISA.

Figure 2.

 

Immunoreactivity of brunescent cataractous lens proteins with anti-argpyrimidine antibody. WI proteins were digested with proteolytic enzymes and tested by competitive ELISA. Values are means ± SD of four determinations.

Figure 3.

 

HPLC detection of argpyrimidine. Brunescent cataractous lens WI proteins were digested by proteolytic enzymes and separated by HPLC. The HPLC peak that corresponded to the argpyrimidine standard (arrow) reacted with the anti-argpyrimidine antibody. Trace 1, chromatogram for argpyrimidine standard; trace 2, chromatogram for brunescent cataractous lens digest; trace 3, argpyrimidine immunoreactivity in fractions.

Figure 4.

 

Quantification of argpyrimidine in WS (A) and WI (B) lens fractions. One hundred micrograms protein was used in competitive ELISA. Data are mean ± SD and are obtained from four independent assays. Numbers in parentheses indicate the number of samples processed. N. cataractous, nuclear cataractous; B. cataractous, brunescent cataractous. Mean values that do not share a common superscript letter are statistically significant at P < 0.05.

Figure 5.

 

Western blotting of cataractous lens proteins. WI proteins from a nuclear cataractous (lane 1) and a brunescent cataractous lens (lane 2) were solubilized by sonication and subjected to SDS-PAGE on a 15% gel. Proteins were electrophoretically transferred to a nitrocellulose membrane and reacted with anti-argpyrimidine antibody, followed by reaction with HRP-conjugated goat anti-mouse IgG. Chemiluminescence detected the immunoreactivity. (A) Coomassie-stained gel; (B) Western blotting.

Figure 6.

 

Detection of argpyrimidine in crystallins. WS proteins from aged human lenses (65–80 years of age) were subjected to column chromatography on a Sephadex G-200 column. The purified crystallin fractions were subjected to SDS-PAGE on a 12% gel under denaturing conditions. The proteins were transferred to a nitrocellulose membrane and probed with anti-argpyrimidine antibody. (A) Western blotting; (B) Coomassie staining. Lanes 1, 2, 3, 4, and 5: α-, β₁-,β ₂-, β₃-, andγ -crystallins, respectively.

Figure 7.

 

Detection of argpyrimidine in the α-crystallin fraction from aged human lenses (see Fig. 6 for details). α-Crystallin fraction was subjected to SDS-PAGE on a 12% gel, and the proteins were transferred to a nitrocellulose membrane and probed with either anti-argpyrimidine antibody or anti-α_A+B antibodies. Lane 1, molecular weight markers; lane 2, α-crystallin fraction stained with Coomassie blue; lane 3,α -crystallin from aged lens probed with anti-α_A+B antibody; lane 4, α-crystallin from aged lens probed with anti-argpyrimidine antibody.