Enhanced C-Terminal Truncation of αA- and αB-Crystallins in Diabetic Lenses

Abstract

purpose. To investigate the influence of diabetes on the cleavage of C-terminal amino acid residues of αA- and αB-crystallins in human and rat lenses.

methods. The human lenses were diabetic or age-matched control lenses from donors 57, 59, 69, and 72 years of age. Lenses were also obtained from streptozotocin-induced diabetic rats. Individual lens crystallins in water-soluble fractions were separated by gel-permeation chromatography. The high (αH)- and low (αL)-molecular-weight fractions were analyzed by electrospray ionization mass spectrometry.

results. A typical mass spectrum of αA-crystallin from human lenses showed intact unmodified αA-crystallin, truncated αA_1-172, and monophosphorylated αA-crystallin. Diabetic lenses showed nearly twofold higher levels of αA_1-172 than did the control lenses. Also, the αH fraction consistently showed significantly higher levels of αA_1-172 than the αL fraction. Human αB-crystallin showed no evidence of C-terminal truncation. Rat αA-crystallin had five C-terminal–truncated components, most of which showed substantial increases in diabetes. Truncated αA_1-162 appeared only in the diabetic rat lenses, suggesting specific activation of m-calpain in diabetes. αB-crystallin had only one C-terminal–truncated component, αB_1-170, which also showed increased levels in diabetes.

conclusions. These data suggest that diabetic stress causes either enzymatic or nonenzymatic cleavage of peptide bonds between specific C-terminal amino acid residues. Such truncated α-crystallins appear to contribute to an increased level of the αH fraction generally present in diabetic lenses. Loss of αA-crystallin chaperone activity seems to be related to truncation of the C-terminal amino acid residues.

Lens crystallins play a crucial role in maintaining the clarity of the lens and its refractive properties. ¹ α-, β-, and γ-Crystallins constitute the three classes of lens crystallins and account for 90% of total lens proteins. ² αA-crystallin and αB-crystallin, the subunits of the oligomeric α-crystallin, are expressed in abundance in the lens and represent approximately 35% of the total lens proteins. In its native state, α-crystallin is water soluble and exists as an oligomer with a molecular mass of 360 to 800 kDa. α-Crystallin functions as a molecular chaperone by suppressing the unfolding of proteins ^{3 4 5 6 7 8} or by mediating in the refolding of unfolded proteins. ^{9 10} αA-crystallin and αB-crystallin, each with a molecular mass of 20 kDa, share 54% sequence homology ^{11 12} and are composed of 173 and 175 amino acid residues, respectively. They are present at molar ratios of 3:1 to 1:1 in the lens, depending on the species and the age of the lens. ^{13 14 15}  

Because the lens crystallins are the longest-lived proteins of the body with little turnover, they are an ideal system for analyzing posttranslational modifications associated with aging and evaluating the effect of these modifications on the protein’s tertiary structure and subsequent cross-linking, aggregation, and insolubilization. ¹⁶ The various posttranslational modifications that have been reported are deamidation, ^{17 18} racemization, ² stereoinversion, ¹⁹ isomerization, ²⁰ site-specific cleavage, ^{2 18 21 22 23 24 25} phosphorylation, ^{2 18 26} oxidation, ^{2 27 28 29} disulfide bonding, ^{2 28 30} photo-oxidation, ³¹ acetylation, ^{32 33} carbamylation, ³³ and glycation. ^{34 35} Some of the listed modifications such as glycation, oxidation, and disulfide bonding, can lead to protein aggregation and subsequent insolubilization. Water-soluble lens proteins, when separated by gel-permeation chromatography, showed the presence of a high-molecular-weight α (αH) fraction and a low-molecular-weight α (αL) fraction. ³⁶ The αH fraction increases with age ³⁷ and diabetes. ³⁸  

Cataract develops earlier in patients with diabetes than in those without diabetes. We have shown that in aged human lenses, α-crystallin chaperone activity is significantly decreased. ^{30 39} Recently, we have shown that diabetic human lenses have an even lower α-crystallin chaperone activity. ⁴⁰ Moreover, streptozotocin-diabetic rats with uncontrolled hyperglycemia show substantial loss in the chaperone function of α-crystallin. ^{38 40} Based on in vitro studies, we had concluded that glycation, ³⁹ oxidation, ³⁹ and mixed disulfide formation, ⁴¹ at least in part, are responsible for the loss in chaperone function. Another possibility that has not been considered is the formation of α-crystallin in which C-terminal residue(s) have been cleaved—a known in vivo modification. ^{22 42} Miesbauer et al. ¹⁸ have reported cleavage of the serine-serine (Ser-172–Ser-173) by analyzing the αA-crystallin in water-soluble protein of young human lenses. Working with total lens protein from human lenses, Takemoto ²² has identified and quantified cleavage of the peptide bond between Ser-172 and Ser-173. The cleavage of this peptide bond is an event that occurs predominantly during the first 12 years of life. ⁴² It is believed that, in vivo, this cleavage may result from endogenous protease activity that has not been fully characterized so far. It has been speculated that increased oxidation of α-crystallin, as it may occur in diabetic lenses, could make α-crystallin more susceptible to the action of the proteases. ²¹ In streptozotocin-diabetic rats, which are widely used for studying diabetic cataracts, no systematic investigation of the C-terminal truncation of α-crystallin has been conducted thus far. In the present study, we analyzed by electrospray ionization mass spectrometry (ESIMS) both αL and αH fractions from diabetic and age-matched nondiabetic human and rat lenses. The results show that in diabetic human lenses there is increased cleavage of the bond between Ser-172 and Ser-173 of αA-crystallin and no C-terminal truncation of αB-crystallin. In the rats, there are multiple sites of cleavage at the C terminus of αA-crystallin and at only one site of αB-crystallin. Again, diabetic lenses have higher levels of the truncated proteins. Comparison of the data from αL versus αH fractions shows the presence of more truncated α-crystallin in the αH-fraction. 

Materials and Methods

Human Control and Diabetic Noncataractous Lenses

Human lenses were obtained from the National Disease Research Interchange (Philadelphia, PA). All lenses used in this study were clear lenses from diabetic or nondiabetic donors 57, 59, 69, and 72 years old. Paired lenses of each donor with no diseases known to compromise lens transparency other than diabetes were used. Age-matched lenses from individuals with no pathologic condition that is known to compromise the lens transparency were used as the control. Only the lenses from the 57-and 59-year-old diabetic donors were pooled and analyzed as a single sample. All other samples were from single donors. The studies with human lens tissues were conducted in compliance with the Declaration of Helsinki and approved by the Institutional Review Board. 

Rat Control and Diabetic Noncataractous Lenses

Diabetes was induced in 1-month-old Sprague-Dawley rats, as described before. ⁴³ Rats were intravenously injected once with 60 mg streptozotocin/kg body weight. By 30 days after injection, the mean fasting plasma glucose level was approximately 700 mg/dL and the glycohemoglobin level approximately 16% compared with 110 mg/dL and 5.3%, respectively, in untreated control rats. On the 30th day, the rats were killed and the eye lenses were enucleated. Both lenses from three rats of each group were pooled to form a single sample. Animal experiments were performed at the Medical College of Georgia (Augusta, GA), in conformity with the provisions of the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and of the Institutional Committee for Animal Care. 

Preparation of Crystallins

Lenses were homogenized in phosphate-buffered saline (PBS; pH 7.3) and centrifuged at 10,000g for 20 minutes at 4°C. The supernatants from three washes were collected as the water-soluble fraction. The various crystallin fractions were separated by gel-permeation chromatography (Sephacryl S-300-HR; 1.5 × 120-cm column, Bio-Rad, Richmond, CA). Absorbance of the eluate was monitored at 280 nm. Isolated crystallins were concentrated by ultrafiltration (Amicon, Beverly, MA), with an ultrafiltration membrane (Diaflo; 10,000 molecular weight cutoff; Amicon). Their purity was checked by SDS-PAGE according to Laemmli. ⁴⁴ Protein concentrations were measured by the bicinchoninic acid (BCA) method (BCA Protein Assay Reagent; Pierce, Rockford, IL). The αH and the αL fractions were analyzed by mass spectrometry. 

ESIMS Analysis of Human and Rat α-Crystallins

The molecular weights of the α-crystallins were determined with an online reversed-phase microbore column (C-4; 30 nm, 5-cm × 1.0 mm inner diameter; Vydac, Hesperia, CA) attached to an electrospray ionization mass spectrometer (Micromass Platform II; Quadrupole, Manchester, UK) calibrated with myoglobin. Mass accuracy was approximately ±2 Da for a 20-kDa protein. A 20-mL sample was injected into the column at a flow-rate of 50 μL/min. A postcolumn splitter directed 5 μL/min to the mass spectrometer and 45 μL/min to a UV monitor and a fraction collector. Solvents used were A-5% formic acid in water and B-5% formic acid in acetonitrile. The gradient ranged from 15% to 50% B in A over 45 minutes. A computer was used (MassLynx software software, Micromass, Manchester, UK) to calculate the molecular weights of the proteins from the multiply charged peaks of the mass spectrum. 

Results

Human and rat lenses, both diabetic, and age-matched control samples were analyzed for posttranstlational modifications in αA- and αB-crystallins. The α-crystallins used for the study were from the αH and αL fractions obtained by gel-permeation chromatography of the water-soluble lens proteins. The water-soluble crystallins separated into six peaks, as seen in Fig. 1 , a representative chromatogram of a 70-year-old human diabetic lens. As expected, this senile diabetic lens had high levels of the αH fraction. 

The molecular weights of the unmodified αA- and αB-crystallins and their modified counterparts were determined by ESIMS. Figure 2 shows ESI mass spectra of αA-crystallins from the αL-crystallin fraction (Figs. 2A 2B) and the αH-crystallin fraction (Figs. 2C 2D) from the 72-year-old human control and diabetic lenses. The first peak at a relative molecular mass (M_r) of 19,861 was attributed to αA-crystallin minus the C-terminal serine (αA_1-172). The peak at M_r 19,950 was αA-crystallin, unmodified except for the N-terminal acetylation common to all α-crystallins, and the peak at M_r 20,030 was monophosphorylated αA-crystallin (αA+1PO₄; Fig. 2A ). For both αL and αH, the relative intensity of the peak due to truncated αA_1-172 was nearly two times higher in the diabetic lenses than in the age-matched control (Fig. 2) . Also, in the αH fractions of both the diabetic and the control lenses, levels of αA_1-172 were higher than in the αL fractions (Fig. 2C 2D) . Results from the 69- and 57/59-year-old diabetic lenses and the age-matched control lenses were essentially the same as those from the 72-year-old lens. However, the peak due to the truncated αA_1-172 was only marginally higher in the αL fraction from diabetic 69- and 57/59-year-old lenses (mass spectra not shown). In addition, a small peak (<5%) at M_r 20,111 was present only in the 57/59-year-old diabetic lenses. This mass increase of 161 Da is close to the 162-Da increase expected for glycated αA-crystallin (results not shown). Varying amounts of acetylated αA-crystallin (peak at M_r 19,993) were seen in all the samples. 

The ratios of the ESIMS responses for αA_1-172 to intact αA-crystallin from the lenses of three different ages were statistically analyzed (Table 1) . Paired t-tests show that the ratio of the signals for truncated αA_1-172 to intact αA-crystallin in the αH fraction from the diabetic lenses is significantly higher (nearly two times higher) than that from the control lenses (P < 0.05); however, this ratio from the αL fraction in the diabetic lens is not statistically significant compared with that in the control. Also in the diabetic lenses, the ratios of the signals of the truncated-to-intact αA-crystallins were higher in the αH than the αL fraction (P < 0.05); however, the same ratio in the control was not significantly different. 

The mass spectrum of αB-crystallin from the 72-year-old lens (Fig. 3) showed a principal peak corresponding to unmodified αB-crystallin (M_r 20,201) and two lesser components corresponding to monophosphorylated (M_r 20,281) and diphosphorylated αB-crystallin (M_r 20,361). No truncated forms of αB-crystallin were observed. Similarly, no truncated αB-crystallin was seen in the 69-year-old and 57/59-year-old lenses. 

In an attempt to compare the modifications of αA- and αB-crystallin in the human lenses with the modifications of α-crystallins of lenses of streptozotocin-diabetic rats and age-matched control lenses, the αL and αH fractions from rat lenses were analyzed in the same manner as the human lenses. Table 2 shows the M_r of the rat αA- and αB-crystallins with their probable identities. The major peak in the ESIMS analysis of αA-crystallin from the rat lenses was at M_r 19,834, the mass of unmodified αA-crystallin (αA_1-173; Fig. 4 ). The truncated forms of αA-crystallin had C-terminal truncations of five (M_r 19,405; αA_1-168), 10 (M_r 18,834; αA_1-163), 11 (M_r 18,678; αA_1-162), 16 (M_r 18,211; αA_1-157), and 22 (M_r 17,573; αA_1-151) amino acids. The peak at M_r 19,913 was due to αA-crystallin with one phosphorylation (αA+1PO₄). In the diabetic lens there were increased levels of all the truncated components. Compared with the αL fractions, the αH fractions from both the control and the diabetic lenses had increased amounts of all the truncated forms of αA-crystallin. The truncated αA at M_r 19,405 (αA_1-168) is the most abundant component in the αH fraction from the diabetic lens. αA_1-162 at M_r 18,678 is present in noteworthy amounts only in the diabetic lens. 

Unmodified αB-crystallin from the rat diabetic and control lenses has an M_r of 20,131 (αB_1-175; Fig. 5 ). The major modifications to αB-crystallin were C-terminal truncation of five amino acids (M_r 19,636; αB_1-170), truncation of five C-terminal amino acids with one phosphorylation (M_r 19,716; αB_1-170+1PO₄), monophosphorylated αB (M_r 20,211; αB+1PO₄) and diphosphorylated αB (M_r 20,291; αB+2 PO₄). The intensities of αB_1-175 and αB+2PO₄ were nearly the same in the control and diabetic lenses. There was an increase in the peak for αB_1-170 in the αL fraction of diabetic lenses. The intensities of αB_1-170 were also increased in the αH fraction of the control and diabetic lenses, and αB_1-170 was the major component of αH from the diabetic lenses. The peak for αB_1-170+1PO₄ (M_r 19,716) also was higher in the diabetic lenses. 

Discussion

Because of its abundance in the lens and its functional significance as a molecular chaperone, α-crystallin has been singled out for extensive characterization of its posttranslational modifications as an index to aging and cataractogenesis. Of the many posttranslational modifications reported for clear lenses, phosphorylation and C-terminal truncation of the α-crystallin are the most prominent. ^{18 21} Truncation was reported as a posttranslational modification by van Kleef et al. ⁴⁵ as early as 1974. Much of their work was based on ion-exchange chromatography and isoelectric focusing that showed cleavage of various amino acids from the C-terminal ends of α-crystallin of bovine lenses. Subsequent studies by these investigators and others have shown that truncation of the α-crystallins increases with age. ^{46 47}  

The ESIMS spectra of αA-crystallin from human diabetic lenses indicate that the major modifications were phosphorylation and C-terminal truncation (Figs. 2B 2D) . In the αH fraction of the diabetic and control lenses, there was more αA_1-172 than in the αL fraction. An increase in αA_1-172 was observed in both the αH and αL fractions of the diabetic lenses; however, this increase was not statistically significant in the αL fraction (Table 1) . Thus, it appears that the truncated αA_1-172 contributes significantly to the increased formation of the αH fractions in the diabetic lenses. ³⁸ The results of a study by Takemoto ⁴² in which the age of the human lenses varied from 1.3 months to 60 years, showed that although the αA-crystallin in the very young lenses (1.3 and 2 months) had no significant cleavage, cleavage of the peptide bond between Ser-172 and Ser-173 increased with age until approximately 12 years, after which the amount of αA_1-172 remained nearly the same. This led the author to conclude that the Ser-172–Ser-173 truncation of the αA-crystallin is related to development of the lens. One conclusion from our study is that increased truncation of Ser-173 from the αA-crystallin is associated with exposure of the lens to stress—in this case, diabetes. Because endogenous proteases responsible for the increased truncation of Ser-173 of the αA-crystallin in human lenses have not been found, it has been suggested that these truncations may be nonenzymatic, possibly due to a long-term exposure to oxidative species and the susceptibility of the hydroxyl-containing serine (Ser-172 and Ser-173) to spontaneous cleavage. ^{46 21}  

Water-soluble proteins from 2-month-old diabetic and nondiabetic rats, studied for the presence of C-terminal–truncated αA-crystallin showed five different truncated products for αA-crystallin with cleavage of 5 (αA_1-168), 10 (αA_1-163), 11 (αA_1-162), 16 (αA_1-157), and 22 (αA_1-151) amino acid residues from the C terminus (Fig. 4 ; Table-2 ). These findings agree with a recent report by Ueda et al. ⁴⁸ who studied in vitro proteolysis of rat αA-crystallin from the water-soluble lens fraction by Lp82 and m-calpain and in vivo proteolysis of αA-crystallin from the water-insoluble lens fraction during normal maturation of rat lens and during selenite cataract formation. In addition to all the truncated αA-crystallin products mentioned herein, they reported the presence of two additional truncated products: αA_1-165 and αA _1-156 during in vitro and in vivo proteolysis. These differences may be explained by the fact that we have analyzed the whole lens water-soluble fraction instead of the insoluble fraction from rat lens nucleus studied by Ueda et al. ⁴⁸ Moreover, two-dimensional gel electrophoresis is considered a superior method compared with size-exclusion chromatography for separating the various proteins. 

In this study, we used 2-month-old streptozotocin-induced diabetic rats (1 month after streptozotocin injection) and normal rats, and none had cataract; diabetic rats show development of cataract only 3 to 4 months after the initiation of diabetes. ⁴³ The goal of this study, and a major difference from the study of Ueda et al., ⁴⁸ was to show the influence of diabetes on αA-crystallin without any interference from the effect of cataract. Lp82 is more active than m-calpain in very young rats, and there is rapid loss of Lp82 activity during maturation of rat lenses, whereas m-calpain activity decreased at a slower rate during the same time span. ⁴⁹ A marginal increase in the levels of Lp82-specific αA_1-168 was observed in the 2-month-old diabetic rat lenses compared with age-matched control rats (Fig. 4) . However, the m-calpain cleavage product αA_1-162 appeared only in the αH and αL of the diabetic rat lenses (Fig. 4) , suggesting that m-calpain is specifically activated by diabetes-induced stress. Lp82, although present in species ranging from mouse to cow, is not present in human lenses. Yet, an Lp82 cleavage product, αA_1-168, has been reported in human lenses by Lund et al., ⁵⁰ who analyzed water-insoluble proteins. Thus, it appears that the water-soluble protein fraction of human lenses, which we used in this study, is not a source of αA_1-168. It is likely that in human lenses, an as yet unknown protease must be responsible for this cleavage. The truncated human αA-crystallin reported in this study also involves a serine–serine bond (Ser-172–Ser-173). In another study, Sharma et al. ⁵¹ incubated bovine and human lens extracts containing proteases with peptide substrates that contained the αA crystallin sequences that showed age-dependent in vivo cleavage sites in human and bovine lenses. Their results show that the peptide substrates cleaved at their expected age-related cleavage sites, which were between Ser-168–Ser-169, Ser-169–Ala-170, and Ser-172–Ser-173 for the peptide substrates incubated with the bovine lens extract and Ser-169–Ala-170 and Ser-172–Ser-173 for those incubated with human lens extract. An age-dependent protease unique to the human lens cannot therefore be eliminated. Characterization of young donor lenses aged less than 12 years would probably help in the search for the evasive protease(s). 

The posttranslational modifications evident in the mass spectra of human αB-crystallin were predominantly phosphorylation, both mono- and diphosphorylation. Truncated αB-crystallin was not seen in these lenses, even though truncation of the C-terminal lysine of the αB-crystallin has been reported. ^{24 25 52 53} The rat αB-crystallin showed αB_1-170, αB_1-170+1PO₄, and αB+1PO₄ and αB +2PO₄ components (Fig. 5) . In contrast, Ueda et al., ⁴⁸ reported the presence of two truncated αB-crystallins, αB_1-170 and αB_1-163. Here again, the only explanation for these disparities is that they used water-insoluble protein of rat lens nucleus as opposed to water-soluble protein of the whole lens used in our study. The C-terminal extensions in both αA- and the αB-crystallin are flexible and solvent accessible; hence, they are easy targets for cleavage. Perhaps, the relative resistance of αB-crystallin to truncation is due to the difference in tertiary structure between the αA- and αB-crystallin. αB-crystallin contains more α-helix and β-pleated sheet than the αA-crystallin. ^{54 55}  

One of the posttranslational modifications expected in diabetic lenses is glycation of the lens crystallins. However, in the present study the only evidence suggestive of glycation was a small peak (<5%) observed in the spectrum for αA-crystallin from the 57/59-year-old diabetic lenses (results not shown). A possible reason that glycated αA- or αB-crystallin is not observed is that diabetes in humans is usually controlled well with appropriate medication. It was surprising that αA- and αB-crystallins from uncontrolled diabetic rat lenses did not show any glycated components. An earlier study ³⁵ has shown that as little as 5% glycation of α-crystallins is detectable by the technique used for analysis in the current study. A possible explanation is that glycation of α-crystallins in these lenses was below the 5% detection limit. Another possibility is that hexose is not the major glycating agent. Recent studies show that such dicarbonyl intermediates as glyoxal and methylglyoxal are the major substrates for in vivo glycation. ^{56 57} Previous examination of the modifications of water-insoluble human lens α-crystallin demonstrated increased degradation of crystallins, deamidation at Gln and Asn, oxidation of Met, backbone cleavage, and extensive disulfide bonding as principal modifications. ^{50 58} In these studies, other posttranslational modifications, including glycation, carbamylation, glutathione adducts, and various cross-links were not detected, possibly because they were present at a level of less than 5% of the total proteins. 

The present study shows enhanced formation of α-crystallins with C-terminal truncation in diabetic lenses. αA-Crystallin is the major source of the truncated α-crystallins. That the truncated α-crystallins are concentrated in the αH fraction strongly suggests a significant role for these posttranslationally modified α-crystallins in the increased formation of the αH fraction in the diabetic lenses. It should be pointed out that an earlier study of the αH fraction has shown a nearly 50% loss of chaperone activity. ³⁸ Kelley et al. ⁵⁹ have shown that proteolysis of α-crystallin by m-calpain results in loss of chaperone activity. Recent studies of recombinant human αA-crystallin and truncated αA_1-172 by Thampi and Abraham ⁶⁰ and of recombinant human αB-crystallin and αB_1-170 by Cho and Abraham ⁶¹ have shown that the chaperone activity of α-crystallin is compromised after proteolysis, because of the C-terminal truncation of αA-crystallin rather than αB-crystallin. 

 Supported by National Eye Institute Grants EY11352 (ECA) and EY07609 (JBS).

 Submitted for publication March 18, 2002; revised May 28, 2002; accepted June 6, 2002.

 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.

 

Separation of water-soluble crystallins by gel-permeation chromatography from 70-year-old diabetic human lens.

Figure 2.

 

ESI mass spectra of reversed-phase HPLC–purified αA-crystallin from the (A, B) αL and (C, D) αH fractions of control and diabetic lenses (72-year-old) illustrating intact unmodified (M_r 19,951), phosphorylated (M_r 20,031), and C-terminal–truncated (αA_1-172, M_r 19,864) αA-crystallins.

Table 1.

 

Ratios of the ESIMS Responses for αA_1-172 to Intact αA-Crystallin

	Age	Control	Diabetic
αL	57,59	0.35	0.38
	69	0.28	0.30
	72	0.28	0.48
αH	57,59	0.38	0.82
	69	0.38	0.55
	72	0.38	0.70

  P < 0.05 for αH in control versus diabetic; P < 0.05 for αL versus αH in diabetics.

Figure 3.

 

ESI mass spectra of αB-crystallin purified from the (A, B) αL and (C, D) αH fractions of lens crystallins from 72-year-old diabetic and age-matched control human lenses showing the αB (M_r 20,201), monophosphorylated αB (αB+1PO₄, M_r 20,281), and diphosphorylated αB (αB+2PO₄, M_r 20,361).

Table 2.

 

Molecular Weights of Water-Soluble Rat Lens α-Crystallins

Identity	Calculated Mr
αA1-173	19,834
αA1-168	19,405
αA1-163	18,834
αA1-162	18,678
αA1-157	18,211
αA1-151	17,573
αA1-173 + 1PO4	19,912
αB1-175	20,131
αB1-170	19,636
αB1-170 + 1PO4	19,716
αB1-175 + 1PO4	20,211
αB1-175 + 2PO4	20,291

Figure 4.

 

ESI mass spectra of αA-crystallin from (A, B) αL- and (C, D) αH-crystallin fractions from control and diabetic rat lenses showing the presence of five degradation products. (See Table 2 for a list of all the products.)

Figure 5.

 

ESI mass spectra of the αB-crystallin from the (A, B) αL- and (C, D) αH-crystallin fractions from control and diabetic rat lenses. (See Table 2 for a list of all the products.)