- Email: [email protected]
Glycation mediated crosslinking between α-crystallin and MP26 in intact lens membranes
lll&klSOfagehJ
and development Mechanisms of Ageing and Development
ELSEVIER
91 (1996) 65-78
Glycation mediated crosslinking between a-crystallin and MP26 in intact lens membranes Malladi Prabhakaram*, Mason Eye Institute,
Martin L. Katz, Beryl J. Ortwerth
University
of Missouri,
Columbia,
MO 65212, USA
Received 6 June 1996
Abstract With advancing age, progressive crosslinking occurs between lens crystallin proteins and other lenticular components. This crosslinking may be involved in the development of senile cataracl:s. Experiments were conducted to determine whether non-enzymatic glycation could be involved in the crosslinking between lens a-crystallin and MP26, an abundant lens fiber cell membrane intrinsic protein. In vitro crosslinking of a-crystallin and MP26 of bovine lens membranes was observed in presence of two degradation products of ascorbic acid (ASA), dehydroascorbic acid (DHA) and threose. Alkali-washed bovine lens membranes, isolated after glycation with DHA and threose, contained both a-crystallin and MP26, as determined by immunoblot and double immunocytochemical labeling studies. In contrast, membranes incubated without these glycating compounds contained only MP26. SDS-PAGE analysis of [‘251]cc-crystallin incubated with lens membranes in the presence of threose showed a higher amount of radioactivity in high molecular weight aggregates than in the aggregates produced when a-crystallin and threose were incubated without membranes. A slot-blot immunoassay of alkali-washed human lens membranes showed a higher amount of covalently bound cr-crystallin in aged, cataractous or diabetic lens membranes than was present in lens membranes from young normal donors. Based on the in vitro results, we hypothesize that non-enzymatic glycation is one of the in vivo mechanisms in the crosslinking of cr-crystallin to lens membrane proteins, such as MP26. This crosslinking may contribute significantly to the development of age-related and diabetic cataracts. Keywords: cc-crystallin;
* Corresponding 0047-635’4/96/$15.00
Cataract;
Glycation;
Lens; L-threose;
Membranes;
author. Fax: + 1 573 8828474. 0 1996 Elsevier Science Ireland Ltd. All rights reserved
PZI SOO47-6374(96)01781-2
MP26
66
44. Prabhakaram et al. / Mechanisms of’ Ageing and Development 91 (1996) 65-78
1. Introduction
Lens opacification or cataract formation is an age-dependent phenomenon. As a result of several post synthetic modifications, lens crystallin proteins become increasingly insoluble with advancing age [l]. In addition, massive accumulation of crystallins bound to lens fiber cell membranes is observed in senile cataracts [ 1,2]. The insoluble crystallin-membrane components, collectively designated the water-insoluble fraction (WIF), can be obtained by homogenization of vertebrate lenses in aqueous buffers [2]. Although cr-crystallin is the major component of WIF in bovine lens [3], the WIF in human lens also contains p and y crystallins [4,5]. Non-specific hydrophobic interactions between cc-crystallin and the lens membrane proteins [6] or lipids [7] have been proposed as possible mechanisms for the initial binding of a-crystallin to lens membranes in vivo. It appears that a significant fraction of a-crystallin is covalently associated with lens fiber cell membranes in senile cataracts. Recently, it has been hypothesized that membrane bound a-crystallin may act as a nucleation site for the aggregation of other lenticular proteins [8]. Since covalently bound non-membrane proteins affect the functional properties of cellular membranes, elucidation of the mechanism(s) responsible for the covalent interaction between crystallins and lens membrane components may be imthe molecular mechanisms underlying senile portant in understanding cataractogenesis. Previously we have shown that the degradation products of ascorbic acid (ASA), such as dehydroascorbic acid (DHA), diketogulonic acid (DKG) and threose, rapidly glycate and crosslink bovine lens crystallin proteins [9], as well as MP26 or MP22 proteins of bovine lens membranes [lo]. The MP26 protein is a major component of vertebrate lens fiber cell membranes. It is an intrinsic protein with a M, of 26 kD. With advancing age, MP26 is converted to MP22 by endolytic cleavage [l 11. The MP22 protein is at least as readily glycated as MP26 [lo]. Non-enzymatic glycation is a condensation reaction that occurs between reducing sugars and protein free amino groups. The resulting Schiff base undergoes Amadori rearrangement and generates advanced glycation end products or AGES that are involved in crosslinking between the amino groups of proteins [12]. It has been demonstrated that streptozotocin-induced diabetic rat lenses accumulate high molecular weight proteins, resembling those produced in vitro by glycation of lens proteins [13]. In the this study we conducted experiments to determine whether crosslinking between lens soluble or-crystallin and the lens membrane protein MP26 can occur via glycation by the ASA degradation products, DHA and threose.
2. Materials and methods 2.1. Materials Calf lenses were collected from a local abattoir and stored at - 80°C until used.
M. Prabhakaram
et al. / Mechanisms
of Ageing and Development
91 (1996) 65-78
61
Human lenses were obtained from the Missouri Lions Eye Tissue Bank and post-operative cataractous or diabetic human lens nuclei were acquired after surgery from the Mason Eye Institute clinical staff. All the biochemicals were obtained from Sigma, St. Louis, MO. Polyclonal antibodies to bovine a-crystallin were obtained as described previously [14] and polyclonal antibodies to MP26 were acquired as a generous gift from Dr. Larry Takemoto, Kansas State University, Manhattan, Kansas. 2.2. Methods 2.2.1. In vitro glycation of bovine lens a-crystallin and membranes containing MP26 Bovi:ne lens a-crystallin and alkali-washed membranes were isolated as described previously [lo, 15,161. The protein contents of the samples were measured by the BCA assay method (Pierce, Rockford, IL), using BSA as the standard. For glycation experiments, a constant amount of alkali-washed lens membranes (1.0 mg protein) and various amounts of a-crystallin (0.5-5.0 mg) were incubated with DHA or threose (5-20 mM) in 1 ml of 0.1 M phosphate buffer, pH 7.0, containing 1.O mM diethylenetriaminepentaaceticacid (DTPA) as the chelator and 0.02% sodium azide to prevent bacterial contamination. Parallel controls were also prepared without sugars. All the reaction mixtures were incubated at 37°C for 2 weeks under sterile conditions. Aliquots were withdrawn at weekly intervals from each incubation mixture and were stored at - 80°C until used. The reaction mixtures devoid of lens membranes were used directly, but those with membranes were sedimented prior to analysis. While the supernatants were used as such, the pelleted membranes were washed twice with 8 M urea, once with 0.1 N NaOH containing 1 mM b-mercaptoethanol and a final wash with 0.02 M phosphate buffer, pH 7.0. This washing procedure removes non-intrinsic proteins that are not covalently bound to the membranes [10,16]. Alkali-washed lens membranes were solubilized in 0.5% SDS in water (1: 1 w/v) prior to analysis. 2.2.2. bwmunoblot analysis and immunogold labeling of u- crystallin and MP26 For :immunoblot analysis, freshly prepared alkali-washed membranes were solubilizecl in SDS and used immediately to avoid polymerization of MP26/MP22 [ll]. On the other hand, the supernatants obtained from the membrane preparations were processed by denaturing in SDS sample buffer at 100°C for 3 min. The denatured proteins were analyzed by SDS-PAGE [17], electroblotted to PVDF membranes [18], and the blots were probed with antibodies to a-crystallin and Ml?26 [lo]. To further assess the glycation dependent crosslinking between u-crystallin and MP26, double immunogold labeling of the glycated lens membranes was conducted with antibodies to u-crystallin and MP26 [19]. Non-specific binding sites on the
68
M. Prabhakaram et al. / Mechanisms of Ageing and Development 91 (1996) 65-78
alkali-washed lens membranes were blocked with 4% BSA in 50 mM Tris, 150 mM NaCI, pH 7.4 (TBS) for 90 min, followed by centrifugation at 12000 x g for 15 min. The sedimented membranes were washed with TBS and then incubated with antibodies to MP26 in TBS for 3 h. After extensive washes with TBS, the membranes were incubated overnight with protein-A conjugated to 5 nm gold. Following this, the membranes were washed with TBS and incubated with antibodies to a-crystallin in TBS for 3 h. Unbound antibodies were removed by washings in TBS. These membranes were incubated overnight with protein-A conjugated to 20 nm gold, washed with TBS, and finally suspended in water. The membrane suspensions (l-2 ~1) were applied to carbon-coated electron microscope grids, dried in a desiccator and then examined with a JEOL 1200 EX electron microscope. Control samples were also prepared and examined as described above, except that non-immune rabbit serum was substituted for the antibody preparations. 2.2.3. Reactivity of [1251]u-crystallin with bovine lens membranes during glycation In order to detect the binding of a-crystallin to membranes during glycation by an additional method, a-crystallin was iodinated [20]. In these experiments, 1 mg of bovine lens [1251]a-crystallin and membranes containing 1.0 mg of protein were incubated with 20 mM of threose in 1 ml of 0.1 M phosphate buffer, pH 7.0, containing 1 mM DTPA and 0.02% sodium azide. In parallel, [‘251]a-crystallin was also incubated either alone, with threose, or with membranes. Aliquots withdrawn at weekly intervals were analyzed by SDS-PAGE. The [i2’I]a-crystallin was localized on the dried SDS gel by scanning with an AMBIS scanner (Automated Microbiology System, San Diego, CA). 2.2.4. Antibody detection of a-crystallin and A4P26 in human lens membranes Clear human lenses from 26, 41, 67 and 84 year old donors, as well as an 84 year old age-onset cataract and 68 year old diabetic cataracts, were used in this study. Four lenses of each age group were pooled, but only one 84 year old aged-cataractous lens and two lenses from a diabetic individual were processed because of scarcity. Alkali-washed human lens membranes were isolated as described earlier and a slot-blot immunoassay of the SDS-solubilized membranes was performed to detect a-crystallin or MP26. Equal amounts of alkali-washed membranes were analyzed from each sample.
3. Results In the this study we report on the in vitro formation of glycation crosslinks between a-crystallin and MP26 of bovine lens membranes in the presence of DHA or threose. This is the first direct evidence to show glycation dependent crosslinking between a-crystallin and MP26 in intact bovine lens membranes.
M. Prabhakaram et al. / Mechanisms of Ageing and Development 91 (1996) 65-78
69
3.1. Glycation and crosslinking of a-crystallin and MP26
The effects of glycation on the crosslinking of a-crystallin and MP26 were analyzed by SDS-PAGE (Fig. 1). Alkali-washed lens membranes contained primarily MPZ!6 and MP22 (lanes 2 and 8). Fig. 1 shows that over 2 weeks, in membranes glycated by threose in the absence of a-crystallin (lanes 3 and 4), MP26 and MP22 were crosslinked to form high molecular weight aggregates that did not enter the polyacrylamide gel. These results are consistent with our previous report on the glycating ability of ASA degradation products on intact bovine lens membranes [lo]. Glycation by DHA of the membranes alone (lanes 9 and IO) also led to high molecullar weight aggregate formation, but at a slower rate than observed with threose. The monomeric MP26 was almost completely converted to crosslinked products after 2 weeks when the membranes alone were glycated by threose (lanes 3 and 41). However, when lens membranes were glycated by threose in presence of a-crystallin (lanes 6, 7, 12 and 13) a considerable amount of MP26 remained monomeric after 2 weeks. This suggests that a-crystallin and MP26 compete in the crosslinking reactions mediated by DHA or threose. 12
3 4 56
7891011121314
I
012012012012
Weeks --IIMIP l
YIP
+ 0: +
Threore Threors
MIP+
DtiA
MIP+
a
+ DHA
Fig. 1. SDS-PAGE of the alkali-washed membranes isolated from O-2 week glycation reactions. Alkali-washed bovine lens membranes were incubated either alone (lanes 2-4 and S-lo), or with a-crystallin (lanes 5-7 and ll- 13), in presence of threose (lanes 2-7) or DHA (lanes 8- 13) and analyzed by SDS-PAGE. To indicate the non-covalently bound cc-crystallin in membranes, urea-washed lens membranes were used at zero time (lanes 5 and 1l), instead of alkali-washed membranes. Lanes 1 and 14 rmspresent the standard molecular weight proteins.
70
M. Prabhakaram et al. / Mechanisms of Ageing and Development 91 (1996) 65- 78
a-crystallin incubated with membranes in the absence of glycating agents is present on polyacrylamide gels as two bands of M, 22 and 20 kD (Fig. 1, lanes 5 and 11). Alkali-washing of the membranes incubated with a-crystallin alone did not show any traces of a-crystallin. However, after incubation of a-crystallin with membranes in the presence of either threose or DHA, the cc-crystallin bands completely disappeared from the samples, while high molecular weight aggregates appear that do not enter the gel. This indicates that both threose and DHA mediated the incorporation of a-crystallin into high molecular weight aggregates. 3.2. Immunoreactivity of bovine lens membrane bound cc-crystallin The data presented in Fig. 1 do not indicate whether glycation-mediated incorporation of a-crystallin into high molecular weight aggregates was due solely to polymerization of a-crystallin or also involved crosslinking between cc-crystallin and MP26. Further experiments were conducted to evaluate the latter possibility. Preliminary experiments showed that a-crystallin was present in the glycated lens membranes after alkali washing, as determined by slot-blot immunoassay (data not shown). However, to further assess whether crosslinks formed between MP26 and a-crystallin during glycation, immunoblot analysis of the alkali-washed membranes was performed (Fig. 2). The immunoblots were probed with antibodies to MP26 (Fig. 2A) and a-crystallin (Fig. 2B). In Fig. 2, all the lanes represent SDS solubilized membrane preparations. Lane 7 represents cr-crystallin obtained from the supernatant of non-glycated membranes. When lens membranes were incubated alone or with a-crystallin in the absence of glycating agents, they predominantly exhibited MP26 after alkali washing (lane 2). However, dimers (52 kD), trimers (78 kD) and polymers ( > 100 kD) were readily detectable by antibodies to MP26, in the reactions containing membranes and threose (lane 3) or membranes, a-crystallin and threose (lanes 4-6). Increased formation of polymers and high molecular weight products (HMWP) that did not enter the spacer gel were observed when lens membranes and different amounts of a-crystallin (0.5 mg, lane 4; 1.0 mg, lane 5 and 5.0 mg, lane 6) were glycated by threose. It is also possible that the HMWP that did not enter the spacer gel might be washed off during staining and destaining of the acrylamide gel. The supernatant obtained from the reaction containing lens membranes and a-crystallin (without threose) predominantly consisted of a-crystallin, with no impurities of MP26 (lane 7). This emphasizes that non-covalently bound a-crystallin was eliminated from the membranes during the alkali-wash of those membranes. Lanes 9- 11 represent the membrane preparations, identical to those in lanes 4-6, but probed with antibodies to cr-crystallin. These results show that antibody detectable cr-crystallin was present in the membrane-bound fraction as HMWP, and the intensity of the immunostaining of the HMWP was dependent upon the amount of a-crystallin in the reaction. In another experiment, a-crystallin alone was first glycated with threose and then incubated with non-glycated membranes. Alkali-washed membranes isolated from this preparation did not show any bound a-crystallin by immunoblot assay (data not shown). These results argue for
M. Prabhakaram et al. 1 Mechanisms of Ageing and Development 91 (1996) 65-78
11
HWWP
POLYMERS
0
‘DIYERS
0 c x
r’ .MPPB
-a
123
4
567891011
Fig. 2. Immunoblot showing the antibody reactivity of MP26 (A) and a-crystallin (B) in the alkaliwashed ‘bovine lens membranes isolated from a 2 week reaction containing a-crystallin, membranes and threose as detailed in the Section 2.2. Membranes with N 25 pg of protein were analyzed by SDS-PAGE and then electroblotted to a PVDF membrane. Except for lane 7, all other lanes represent the SDS-solubilized membranes. The blot shows the incubations containing membranes alone or with cc-crystallin (lane 2); membranes and threose (lane 3); membranes, threose and different amounts of a-crystallin (0.5 mg, lane 4; 1.0 mg, lane 5; and 5.0 mg, lane 6). Lane 7 is the initial supematant obtained from the reaction containing membranes and a-crystallin, but not threose. The description for lanes 8-l 1 is the same as for the lanes 3-6, but probed with antibodies to a-crystallin.
the lack of non-specific interaction between pre-glycated a-crystallin and non-glycated MP26 in bovine lens membranes. Therefore, glycation of both cl-crystallin and MP26 is required for the crosslinking of these proteins. 3.3. Distribution of a-crystallin on the glycated membranes To further confirm that a-crystallin was covalently bound to membranes during glycation, immunogold labeling of the alkali- washed glycated membranes was perfonmed. MP26-Ab and CI-crystallin-Ab complexes were labeled with protein-A conjugated to 5 nm and 20 nm gold particles, respectively (Fig. 3A and B). In membranes incubated with a-crystallin in the absence of threose, only anti-MP26 immunolabel was detected, indicating that a-crystallin did not bind to the membranes in an alkali-resistant manner in the absence of threose (Fig. 3B). On the other hand, membranes incubated with both threose and a-crystallin and then subjected to alkali washing were labeled by both anti-MP26 and anti cc-crystallin
72
M. Prabhakaram et al. 1 Mechanisms
ofAge&
and Development 91 (1996) 65- 78
antib odies (Fig. 3A). The immunolabels for MP26 (5 nm) and a-crystallin (20 nm) were for the most part co-localized in the glycated membranes. These results suggt :st that the crosslinking of a-crystallin to the membranes was primarily
Fig. 3. Electron micrographs showing the a-crystallin bound to bovine lens membranes during glycation. Alkali-.washed lens membranes were isolated from 2 week reactions containing cc-crystallin and membranes which were incubated in presence (A) and absence (B) of threose. The membranes were probed with a ntibodies to MP26 or a-crystallin and localized by protein A-conjugated to 5 or 20 nm gold, respecl :ively, as described in Section 2.2. Arrow heads indicate MP26, whereas arrows indicate a-crystallin.
M. Prabhakaram et al. / Mechanisms of Ageing and Development 91 (1996) 65-78
a Cryst
13
+ Memb + Thr
I-- ..-_..-a-CrySt
360
i 240 1
1
120
O-
I 92
&ince
I 4
I 6
I a
lo
on SDS gel (cd
Fig. 4. Radioactive detection of iodinated a-crystallin on SDS gels. [‘251]cc-crystallinwas incubated alone (bottom panel), with threose (middle panel), or with membranes and threose (top panel) for 2 weeks as described in Section 2.2. 100 ,ug of aliquot from each reaction was analyzed by SDS-PAGE and radioactivity was determined by AMBIS image scanning of the dried SDS gel.
through MP26. Since the a-crystallin was not removed crosslinking between cc-crystallin and MP26 is considered nature,
by the alkali wash, as being covalent in
3.4. Interaction between [1251]u-crystallin and membranes during glycation In order to evaluate the glycation dependent crosslinking between a-crystallin and MP26, bovine lens membranes and [1251]cr-crystallin were glycated by threose and thlen analyzed by SDS-PAGE. Radioactive scanning of the dried SDS gel (Fig. 4) shows that the incubation of [‘251]a-crystallin alone or with membranes in the absence of threose (Fig. 4, bottom panel) did not produce any crosslinks. Under these conditions, radioactivity was solely present in the subunits of a-crystallin. When [1251]cr-crystallin alone was reacted with threose for 2 weeks, crosslinks were
74
M. Prubhakaram et al. / Mechanisms of Age&
and Deoelopment 91 (1996) 65- 78
readily formed (Fig. 4, middle panel), and the radioactivity was primarily distributed in the oligomers. In contrast, incubation of [“51]a-crystallin, membranes and threose (Fig. 4, top panel) exhibited higher radioactivity in the HMWP fraction than in the oligomers. It appears that in the presence of membranes and threose, glycated a-crystallins not only crosslink among themselves, but they also crosslink with MP26 to favor the formation of increased amounts of HMWP. Moreover, the [i2’I]m-crystallin glycated by threose mainly crosslinked as oligomers, but in presence of threose and membranes, the HMWP fraction was increased 2-3 fold, with this fraction containing both a-crystallin and MP26. Since a-crystallin has more glycation sites (a total of -v 300 lysines/molecule of 8 x 10’ Da a-crystallin) than MP26 (3 lysines/molecule of 26 kD MP26), the formation of HMWP would likely require a-crystallin to crosslink multiple MP26 molecules on the surface of the membrane. 3.5. Slot-blot immunoassay of human lens membranes Previously, it has been shown that both MP26 and a-crystallin can be detected by specific antibodies in the whole lens preparations of human nuclear cataracts [21]. In order to determine whether cr-crystallin is covalently bound to lens membranes in vivo, we have analyzed alkali-washed human lens membranes from donors of different ages. These results show that antibody detectable a-crystallin was present in alkali-washed membranes isolated from young (26 years and 41 years), old (67 years), aged or cataractous (84 years) and diabetic (68 years) human lenses as determined by slot-blot immunoassay (Fig. 5). The amounts of cc-crystallin bound to lens cell membranes were significantly lower in the samples from 26, 41 and 67 year old human donors than in the aged or cataractous (84 years) lens membranes. In addition, the intensity of cr-crystallin immunolabeling was higher in the 68 year old diabetic lens membranes than in the age-matched 67 year old lens membranes from a normal donor. Similar staining intensities were observed when the human lens membranes of different ages were probed with antibodies to MP26 (Fig. 5, bottom panel), indicating that similar amounts of lens membrane were present in each sample.
Anti a-Crystrllin
Anti MP26
Fig. 5. Antibody detection of cc-crystallin and MP26 in different ages of human lens membranes by slot-blot immunoassay. Alkali-washed human lens membranes were isolated as described in Section 2.2 and an equal volume (10 pl/lens) of SDS solubilized membranes was used for the antibody detection of cc-crystallin (top panel) and MP26 (bottom panel).
M. Prabhakaram
et al. / Mechanisms
of Ageing and Development
91 (1996) 65-78
15
4. Discussion The development of senile cataracts is an age-dependent process that can be influenced by environmental insults. Several post-translational modifications of lens proteins may be involved in the formation of high molecular weight aggregates or the water insoluble fraction in the human lens. Modifications, such as those caused by oxidation, UV-irradiation, disulfide formation, hydrophobic interactions and non-enzymatic glycation, have been proposed to be involved in the formation of high molecular weight protein aggregates in vivo [l]. Many of these insults may be involved in the crosslinking of soluble lens crystallins with fiber cell membrane components which accumulate as high molecular weight aggregates during normal aging and in senile cataracts. The purpose of the this work was to investigate the potential contribution of non-enzymatic glycation in the crosslinking of a-crystallin to MP26, an intrinsic lens fiber cell membrane protein. By incubation of soluble bovine lens proteins with the degradation products of ASA, it has been shown that high molecular weight protein aggregates can be generated in vitro [9,10,15]. ASA and its degradation products, such as DHA, DKG, xylosone or threose, have been shown to rapidly glycate lens proteins under in vitro conditions [9]. It has been shown recently that lens proteins glycated by ASA generate UVA sensitizers and high molecular weight proteins which resemble those present in aged human lenses [22]. All the breakdown products of ASA glycate and crosslink lens proteins [9] or membrane proteins [lo]. Of these ascorbate breakdown products, threose, a four carbon compound, is the most reactive with proteins and it can be formed either from ASA or DHA [23]. In general, carbohydrates with shorter chain lengths are more reactive with protein amino groups [12]. We have observed that either a-crystallin or MP26/MP22 reacts with the six-carbon compound DHA at a lower rate th.an with threose (Fig. 1). The physiological levels of threose are difficult to assess because of the rapid reactivity of threose with free amino groups of proteins. By using [14C]threose, we observed previously that threose is readily incorporated into human lens proteins [9]. In the present study, we have used intact bovine lens membranes to assess whether glycation mediated crosslinking can occur between the soluble lens protein cc-cryaallin and the major lens membrane intrinsic protein MP26/22. The membranes used in the crosslinking studies were treated with alkali to remove any non-covalently bound membrane components [ 10,161. However, a small amount ( < 1%) of other membrane components, such as cytoskeletal elements. beaded filaments and MP70, are usually present, even in the alkali washed membranes [l]. These membrane components may also undergo glycation in vivo. By incubating alkali-washed membranes, we have observed that either MP26 or MP22 of bovine lens membranes crosslink and accumulate as high molecular weight aggregates during glycation. Glycation mediated crosslinking of a-crystallin was also observed with these membrane proteins. Alkali-treatment of lens membranes glycated in the presence of a-crystallin eliminated non-covalently bound a-crystallin from membranes. Thus, the presence of cc-crystallin in the membrane fractions after glycation indicates that a fraction of the a-crystallin had become covalently bound.
16
M. Prabhakaram et al. / Mechanisms of Ageing and Deuelopment 91 (1996) 65-78
Our results also show that during glycation of lens membranes in the presence of a-crystallin, a fraction of MP26 remained monomeric, whereas no monomeric MP26 remained after glycation of the membranes in the absence of a-crystallin (Fig. 1). This suggests that a-crystallin and MP26 (or MP22) compete for the available DHA or threose. In bovine lens, a-crystallin has a J4, of 800 kD and is composed of 40 subunits with a total of 280-290 lysines [l]. The intrinsic membrane protein MP26 has a M, of 26 kD and it contains 3 lysines along with a free N-terminal amino group [24]. Due to the differences in molecular mass of a-crystallin and MP26, stoichiometrically, even a single a-crystallin molecule can span a number of MP26 molecules on membranes. Therefore, the freely accessible lysines on cc-crystallin are glycated initially, followed by the crosslinking of these lysines to the lysines of additional cr-crystallins or those of MP26/MP22 on the membranes. With time, these crosslinked proteins aggregate and accumulate as high molecular weight complexes. Recently it has been shown that cc-crystallin contains preferential, but common, glycation sites to react with ASA [25] or glucose [26]. By employing double immunolabeling techniques and electron microscopic studies, we have observed both MP26 and M-crystallin on the glycated bovine lens membranes. By the BCA assay method we have estimated that not more than - 40-50 ,ug of cr-crystallin was covalently bound to 1 mg of membrane bound MP26 during glycation. In contrast, it has been shown that as much as 5 mg of native a-crystallin can bind non-specifically to 1 mg of MP26 in non-alkali washed bovine lens membranes [27]. We have also observed an age-dependent increase in the protein content of WIF in the human lenses by the BCA assay method. The protein content in the WIF of 26 year human lens was 1 mg/lens, whereas 4.2 mg/lens was present in the WIF of 85 year old cataractous lenses. The increased intensity of a-crystallin immunostaining in the alkali-washed membranes of human cataracts could represent the increased amount of cr-crystallin bound to the aged lens membranes in the WIF. The increased amounts of a-crystallin bound to the diabetic lens membranes suggests that glycation may be involved, since non-enzymatic glycation is implicated in diabetics [12,13]. Increased levels of DHA, the precursor of threose, have been reported in nuclear and diabetic cataracts [28,29], suggesting the possibility of ascorbate mediated glycation in vivo. A massive accumulation of cr-crystallin in aged or cataractous human lens membranes [4,5], and also the interaction of phospholipid vesicles containing MP26 with cc-crystallin [6], suggest the crosslinking ability of a-crystallin to membrane components in vivo. The failure of alkali washing to remove the bound a-crystallin in the glycated bovine lens membranes or in cataractous human lens membranes suggests that a-crystallin is covalently bound to membranes during glycation or aging, respectively. In addition to cr-crystallin, significant amounts of j3 and y-crystallins have been detected in the WIF of both normal and cataractous human lens membranes [7]. The presence of these crystallins on membranes, even after treatment of P-mercaptoethanol or alkali washing [21], indicate that a mechanism(s) other than simple hydrophobic interaction or disulfide crosslinking is involved in the crosslinking of crystallins to membrane components in vivo.
M. Prabhakaram et al. / Mechanisms of Ageing and Development 91 (1996) 65-78
71
Previously, we have shown that blocking the available -SH groups in the bovine lens soluble proteins did not prevent the crosslinking properties of cr-crystallin in the presence of ASA or its breakdown products [30]. Based upon the present findings, we hypothesize that the in vivo crosslinking of crystallins to lens membrane-bound proteins involves glycation. With advancing age, the highly crosslinked membrane-protein complexes appear to aggregate and contribute to the lens opacification in senile cataractogenesis.
Acknowledgements The authors thank Dr. Larry Takemoto for generously supplying the MP26 antibodies and Chun-Lan Gao for her assistance in electron microscopy. This work was supported by NIH grants EY07070 (BJO), EY08813 (MLK) and in part by Research to Prevent Blindness.
References [l] E.R. Berman, In E.R. Berman (ed.), Biochemistry of the Eye, Plenum Press, New York, 1991, pp. 201-290. [2] A. Pirie, Color and solubility of the proteins of human cataracts. Invest. Ophthalmol., 7 (1968) 634-650. [3] J. Bours, Isotachophoresis
and immunoelectrophoresis of water soluble and insoluble crystallins of the ageing bovine lens. Curr. Eye Res., 3 (1984) 691-697. [4] M. Takehana and L. Takemoto, Quantitation of membrane-associated crystallins from aging and cataractous human lenses. Invest. Ophthalmol. Vis. Sci., 28 (1987) 780-784. [5] G. Chandrasekhar and R.J. Cenedella, Protein associated with human lens ‘native’ membrane during aging and cataract formation. Exp. Eye Res., 60 (1995) 707-717. [6] J.W.M. Mulders, J. Stokkermans, J.A.M. Leunissen, E. Benedetti, H. Blomendal and W.W. DeJong, Interaction of cc-crystallin with lens plasma membranes. Affinity for MP26. Eur. J. Biochem., 152 (1985) 721-728. [7] F. Ifeanyi and L. Takemoto, Interaction
[8] [9] [lo] [ll] [12] [13]
[14]
of lens crystallins with lipid vesicles. Exp. Eye Res., 53 (1931) 535-538. CF:. Fleschner and R.J. Cenedella, Examination of a lens native plasma membrane fraction and its associated crystallins. Curr. Eye Res., I1 (1992) 739-752. B.J. Ortwerth, J.A. Speaker, M. Lopez, E. Li and M.S. Feather, Ascorbic acid glycation; the rea’ction of L-threose in lens tissue. Exp. Eye Res., 58 (1994) 665-674. M. Prabhakaram and B.J. Ortwerth, Glycation of MP26 and MP 22 in bovine lens membranes. Biochem. Biophys. Res. Commn., 185 (1992a) 4966504. L.J. Takemoto, J.S. Hansen and J. Horwitz, Inter species conservation of the main intrinsic polypeptide (MIP) of the lens membrane. Camp. Biochem. Physiol., 68B (1981) 101-106. V.M. Monnier, In J.E. Baynes and V.M. Monnier (eds), The Maillard Reaction in Aging, Diabetes and Nutrition, Alan R. Liss, New York, 1989, pp. l-22. R.E. Perry, M.S. Swamy and E.C. Abraham, Progressive changes in lens crystallin glycation and high molecular weight aggregate formation leading to cataract development in streptozotocin-diabetic rats. Exp. Eye Res., 44 (1987) 269-282. B.J. Ortwerth, P.R. Olesen, K.K. Sharma and M. Prabhakaram, Chemical modification of alpha crystallin. Exp. Eye Res., 56 (1993) 107-l 14.
78
M. Prabhakaram et al. 1 Mechanisms 01 Ageing and Development 91 (1996) 65-78
[15] M. Prabhakaram and B.J. Ortwerth, The glycation and cross-linking of isolated lens crystallins by ascorbic acid. Exp. Eye Res., 55 (1992b) 451-459. [I61 P. Russell, W.G. Robison and J.H. Kinoshita, A new method for rapid isolation of the intrinsic membrane proteins from lens. Exp. Eye Res., 32 (1981) 51 l-516. [17] U.K. Laemmli, Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature, 277 (1970) 680-685. [18] H. Towbin, T. Staehelin and J. Gordon, Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose: procedure and some applications. Proc. Natl. Acad. Sci. USA, 76 (1979) 4350-4354. [19] M.L. Katz, C. Gao and H.J. Stientjes, Regulation of the inter photoreceptor protein content of the retina by vitamin A. Exp. Eye Res., 57 (1993) 393-401.
retinoid-binding
[20] F.C. Greenwood, W.M. Hunter and J.S. Glover. The preparation of ‘311-labeled human growth hormone of high specific radioactivity. Biochem. J., 89 (1963) 114-123. [21] L. J. Takemoto, J.S. Hansen and J. Horwitz, High molecular weight aggregates from human cataracts: characterization by Western Blot analysis. Biochem. Biophys. Res. Commn., 122 (1984) 1028-1033. [22] B.J. Ortwerth, M. Linetsky and P.R. Olesen, Ascorbic acid glycation of lens proteins produces UVA sensitizers similar to those in human lens. Photochem. Photobiol., 62 (1995) 454-462. [23] M.G. Lopez and M.S. Feather, The production of threose as a degradation product from L-ascorbic acid. J. Carbohydr. Chem., II (1992) 799-800. [24] M.B. Gorin, S.B. Yancey, J. Cline, J.P. Revel and J. Horwitz, The major intrinsic protein (MIP) of the bovine lens fiber membrane. Characterization and structure based on cDNA cloning. Cell, 39 (1984) 49-59. [25] B.J. Ortwerth, S.H. Slight, M. Prabhakaram, Y. Sun and J.B. Smith, Site specific glycation of lens crystallins by ascorbic acid. Biochim. Biophys. Acta, 1 I1 7 (1992) 207-215. [26] E.C. Abraham, M. Cherian and J.B. Smith, Site-specificity in the glycation of stA and ctB crystallins by glucose. Biochem. Biophys. Res. Commn., 201 (1994) 1451-1456. [27] R.J. Cenedella and G. Chandrasekhar, High capacity binding of alpha crystallins to various bovine lens membrane preparations. Curr. Eye Res., 12 (1993) 1025-1038. [28] W. Lohmann, W. Schmehl and J. Strobel, Nuclear cataract. Oxidative damage to the lens. Exp. Eye Res., 43 (1986) 859-862. [29] S. Som, D. Mukherjee, S. Deb, R. Choudhury, S.N. Mukherjee, S.N. Chatterjee and I.B. Chatterjee, Ascorbic acid metabolism in diabetes mellitus. Metabolism, 30 (1981) 572-577. [30] M. Prabhakaram and B.J. Ortwerth, The glycation-associated crosslinking of lens proteins by ascorbic acid is not mediated by oxygen free radicals. Exp. Eye Res., 53 (1991) 261-168.
and development Mechanisms of Ageing and Development
ELSEVIER
91 (1996) 65-78
Glycation mediated crosslinking between a-crystallin and MP26 in intact lens membranes Malladi Prabhakaram*, Mason Eye Institute,
Martin L. Katz, Beryl J. Ortwerth
University
of Missouri,
Columbia,
MO 65212, USA
Received 6 June 1996
Abstract With advancing age, progressive crosslinking occurs between lens crystallin proteins and other lenticular components. This crosslinking may be involved in the development of senile cataracl:s. Experiments were conducted to determine whether non-enzymatic glycation could be involved in the crosslinking between lens a-crystallin and MP26, an abundant lens fiber cell membrane intrinsic protein. In vitro crosslinking of a-crystallin and MP26 of bovine lens membranes was observed in presence of two degradation products of ascorbic acid (ASA), dehydroascorbic acid (DHA) and threose. Alkali-washed bovine lens membranes, isolated after glycation with DHA and threose, contained both a-crystallin and MP26, as determined by immunoblot and double immunocytochemical labeling studies. In contrast, membranes incubated without these glycating compounds contained only MP26. SDS-PAGE analysis of [‘251]cc-crystallin incubated with lens membranes in the presence of threose showed a higher amount of radioactivity in high molecular weight aggregates than in the aggregates produced when a-crystallin and threose were incubated without membranes. A slot-blot immunoassay of alkali-washed human lens membranes showed a higher amount of covalently bound cr-crystallin in aged, cataractous or diabetic lens membranes than was present in lens membranes from young normal donors. Based on the in vitro results, we hypothesize that non-enzymatic glycation is one of the in vivo mechanisms in the crosslinking of cr-crystallin to lens membrane proteins, such as MP26. This crosslinking may contribute significantly to the development of age-related and diabetic cataracts. Keywords: cc-crystallin;
* Corresponding 0047-635’4/96/$15.00
Cataract;
Glycation;
Lens; L-threose;
Membranes;
author. Fax: + 1 573 8828474. 0 1996 Elsevier Science Ireland Ltd. All rights reserved
PZI SOO47-6374(96)01781-2
MP26
66
44. Prabhakaram et al. / Mechanisms of’ Ageing and Development 91 (1996) 65-78
1. Introduction
Lens opacification or cataract formation is an age-dependent phenomenon. As a result of several post synthetic modifications, lens crystallin proteins become increasingly insoluble with advancing age [l]. In addition, massive accumulation of crystallins bound to lens fiber cell membranes is observed in senile cataracts [ 1,2]. The insoluble crystallin-membrane components, collectively designated the water-insoluble fraction (WIF), can be obtained by homogenization of vertebrate lenses in aqueous buffers [2]. Although cr-crystallin is the major component of WIF in bovine lens [3], the WIF in human lens also contains p and y crystallins [4,5]. Non-specific hydrophobic interactions between cc-crystallin and the lens membrane proteins [6] or lipids [7] have been proposed as possible mechanisms for the initial binding of a-crystallin to lens membranes in vivo. It appears that a significant fraction of a-crystallin is covalently associated with lens fiber cell membranes in senile cataracts. Recently, it has been hypothesized that membrane bound a-crystallin may act as a nucleation site for the aggregation of other lenticular proteins [8]. Since covalently bound non-membrane proteins affect the functional properties of cellular membranes, elucidation of the mechanism(s) responsible for the covalent interaction between crystallins and lens membrane components may be imthe molecular mechanisms underlying senile portant in understanding cataractogenesis. Previously we have shown that the degradation products of ascorbic acid (ASA), such as dehydroascorbic acid (DHA), diketogulonic acid (DKG) and threose, rapidly glycate and crosslink bovine lens crystallin proteins [9], as well as MP26 or MP22 proteins of bovine lens membranes [lo]. The MP26 protein is a major component of vertebrate lens fiber cell membranes. It is an intrinsic protein with a M, of 26 kD. With advancing age, MP26 is converted to MP22 by endolytic cleavage [l 11. The MP22 protein is at least as readily glycated as MP26 [lo]. Non-enzymatic glycation is a condensation reaction that occurs between reducing sugars and protein free amino groups. The resulting Schiff base undergoes Amadori rearrangement and generates advanced glycation end products or AGES that are involved in crosslinking between the amino groups of proteins [12]. It has been demonstrated that streptozotocin-induced diabetic rat lenses accumulate high molecular weight proteins, resembling those produced in vitro by glycation of lens proteins [13]. In the this study we conducted experiments to determine whether crosslinking between lens soluble or-crystallin and the lens membrane protein MP26 can occur via glycation by the ASA degradation products, DHA and threose.
2. Materials and methods 2.1. Materials Calf lenses were collected from a local abattoir and stored at - 80°C until used.
M. Prabhakaram
et al. / Mechanisms
of Ageing and Development
91 (1996) 65-78
61
Human lenses were obtained from the Missouri Lions Eye Tissue Bank and post-operative cataractous or diabetic human lens nuclei were acquired after surgery from the Mason Eye Institute clinical staff. All the biochemicals were obtained from Sigma, St. Louis, MO. Polyclonal antibodies to bovine a-crystallin were obtained as described previously [14] and polyclonal antibodies to MP26 were acquired as a generous gift from Dr. Larry Takemoto, Kansas State University, Manhattan, Kansas. 2.2. Methods 2.2.1. In vitro glycation of bovine lens a-crystallin and membranes containing MP26 Bovi:ne lens a-crystallin and alkali-washed membranes were isolated as described previously [lo, 15,161. The protein contents of the samples were measured by the BCA assay method (Pierce, Rockford, IL), using BSA as the standard. For glycation experiments, a constant amount of alkali-washed lens membranes (1.0 mg protein) and various amounts of a-crystallin (0.5-5.0 mg) were incubated with DHA or threose (5-20 mM) in 1 ml of 0.1 M phosphate buffer, pH 7.0, containing 1.O mM diethylenetriaminepentaaceticacid (DTPA) as the chelator and 0.02% sodium azide to prevent bacterial contamination. Parallel controls were also prepared without sugars. All the reaction mixtures were incubated at 37°C for 2 weeks under sterile conditions. Aliquots were withdrawn at weekly intervals from each incubation mixture and were stored at - 80°C until used. The reaction mixtures devoid of lens membranes were used directly, but those with membranes were sedimented prior to analysis. While the supernatants were used as such, the pelleted membranes were washed twice with 8 M urea, once with 0.1 N NaOH containing 1 mM b-mercaptoethanol and a final wash with 0.02 M phosphate buffer, pH 7.0. This washing procedure removes non-intrinsic proteins that are not covalently bound to the membranes [10,16]. Alkali-washed lens membranes were solubilized in 0.5% SDS in water (1: 1 w/v) prior to analysis. 2.2.2. bwmunoblot analysis and immunogold labeling of u- crystallin and MP26 For :immunoblot analysis, freshly prepared alkali-washed membranes were solubilizecl in SDS and used immediately to avoid polymerization of MP26/MP22 [ll]. On the other hand, the supernatants obtained from the membrane preparations were processed by denaturing in SDS sample buffer at 100°C for 3 min. The denatured proteins were analyzed by SDS-PAGE [17], electroblotted to PVDF membranes [18], and the blots were probed with antibodies to a-crystallin and Ml?26 [lo]. To further assess the glycation dependent crosslinking between u-crystallin and MP26, double immunogold labeling of the glycated lens membranes was conducted with antibodies to u-crystallin and MP26 [19]. Non-specific binding sites on the
68
M. Prabhakaram et al. / Mechanisms of Ageing and Development 91 (1996) 65-78
alkali-washed lens membranes were blocked with 4% BSA in 50 mM Tris, 150 mM NaCI, pH 7.4 (TBS) for 90 min, followed by centrifugation at 12000 x g for 15 min. The sedimented membranes were washed with TBS and then incubated with antibodies to MP26 in TBS for 3 h. After extensive washes with TBS, the membranes were incubated overnight with protein-A conjugated to 5 nm gold. Following this, the membranes were washed with TBS and incubated with antibodies to a-crystallin in TBS for 3 h. Unbound antibodies were removed by washings in TBS. These membranes were incubated overnight with protein-A conjugated to 20 nm gold, washed with TBS, and finally suspended in water. The membrane suspensions (l-2 ~1) were applied to carbon-coated electron microscope grids, dried in a desiccator and then examined with a JEOL 1200 EX electron microscope. Control samples were also prepared and examined as described above, except that non-immune rabbit serum was substituted for the antibody preparations. 2.2.3. Reactivity of [1251]u-crystallin with bovine lens membranes during glycation In order to detect the binding of a-crystallin to membranes during glycation by an additional method, a-crystallin was iodinated [20]. In these experiments, 1 mg of bovine lens [1251]a-crystallin and membranes containing 1.0 mg of protein were incubated with 20 mM of threose in 1 ml of 0.1 M phosphate buffer, pH 7.0, containing 1 mM DTPA and 0.02% sodium azide. In parallel, [‘251]a-crystallin was also incubated either alone, with threose, or with membranes. Aliquots withdrawn at weekly intervals were analyzed by SDS-PAGE. The [i2’I]a-crystallin was localized on the dried SDS gel by scanning with an AMBIS scanner (Automated Microbiology System, San Diego, CA). 2.2.4. Antibody detection of a-crystallin and A4P26 in human lens membranes Clear human lenses from 26, 41, 67 and 84 year old donors, as well as an 84 year old age-onset cataract and 68 year old diabetic cataracts, were used in this study. Four lenses of each age group were pooled, but only one 84 year old aged-cataractous lens and two lenses from a diabetic individual were processed because of scarcity. Alkali-washed human lens membranes were isolated as described earlier and a slot-blot immunoassay of the SDS-solubilized membranes was performed to detect a-crystallin or MP26. Equal amounts of alkali-washed membranes were analyzed from each sample.
3. Results In the this study we report on the in vitro formation of glycation crosslinks between a-crystallin and MP26 of bovine lens membranes in the presence of DHA or threose. This is the first direct evidence to show glycation dependent crosslinking between a-crystallin and MP26 in intact bovine lens membranes.
M. Prabhakaram et al. / Mechanisms of Ageing and Development 91 (1996) 65-78
69
3.1. Glycation and crosslinking of a-crystallin and MP26
The effects of glycation on the crosslinking of a-crystallin and MP26 were analyzed by SDS-PAGE (Fig. 1). Alkali-washed lens membranes contained primarily MPZ!6 and MP22 (lanes 2 and 8). Fig. 1 shows that over 2 weeks, in membranes glycated by threose in the absence of a-crystallin (lanes 3 and 4), MP26 and MP22 were crosslinked to form high molecular weight aggregates that did not enter the polyacrylamide gel. These results are consistent with our previous report on the glycating ability of ASA degradation products on intact bovine lens membranes [lo]. Glycation by DHA of the membranes alone (lanes 9 and IO) also led to high molecullar weight aggregate formation, but at a slower rate than observed with threose. The monomeric MP26 was almost completely converted to crosslinked products after 2 weeks when the membranes alone were glycated by threose (lanes 3 and 41). However, when lens membranes were glycated by threose in presence of a-crystallin (lanes 6, 7, 12 and 13) a considerable amount of MP26 remained monomeric after 2 weeks. This suggests that a-crystallin and MP26 compete in the crosslinking reactions mediated by DHA or threose. 12
3 4 56
7891011121314
I
012012012012
Weeks --IIMIP l
YIP
+ 0: +
Threore Threors
MIP+
DtiA
MIP+
a
+ DHA
Fig. 1. SDS-PAGE of the alkali-washed membranes isolated from O-2 week glycation reactions. Alkali-washed bovine lens membranes were incubated either alone (lanes 2-4 and S-lo), or with a-crystallin (lanes 5-7 and ll- 13), in presence of threose (lanes 2-7) or DHA (lanes 8- 13) and analyzed by SDS-PAGE. To indicate the non-covalently bound cc-crystallin in membranes, urea-washed lens membranes were used at zero time (lanes 5 and 1l), instead of alkali-washed membranes. Lanes 1 and 14 rmspresent the standard molecular weight proteins.
70
M. Prabhakaram et al. / Mechanisms of Ageing and Development 91 (1996) 65- 78
a-crystallin incubated with membranes in the absence of glycating agents is present on polyacrylamide gels as two bands of M, 22 and 20 kD (Fig. 1, lanes 5 and 11). Alkali-washing of the membranes incubated with a-crystallin alone did not show any traces of a-crystallin. However, after incubation of a-crystallin with membranes in the presence of either threose or DHA, the cc-crystallin bands completely disappeared from the samples, while high molecular weight aggregates appear that do not enter the gel. This indicates that both threose and DHA mediated the incorporation of a-crystallin into high molecular weight aggregates. 3.2. Immunoreactivity of bovine lens membrane bound cc-crystallin The data presented in Fig. 1 do not indicate whether glycation-mediated incorporation of a-crystallin into high molecular weight aggregates was due solely to polymerization of a-crystallin or also involved crosslinking between cc-crystallin and MP26. Further experiments were conducted to evaluate the latter possibility. Preliminary experiments showed that a-crystallin was present in the glycated lens membranes after alkali washing, as determined by slot-blot immunoassay (data not shown). However, to further assess whether crosslinks formed between MP26 and a-crystallin during glycation, immunoblot analysis of the alkali-washed membranes was performed (Fig. 2). The immunoblots were probed with antibodies to MP26 (Fig. 2A) and a-crystallin (Fig. 2B). In Fig. 2, all the lanes represent SDS solubilized membrane preparations. Lane 7 represents cr-crystallin obtained from the supernatant of non-glycated membranes. When lens membranes were incubated alone or with a-crystallin in the absence of glycating agents, they predominantly exhibited MP26 after alkali washing (lane 2). However, dimers (52 kD), trimers (78 kD) and polymers ( > 100 kD) were readily detectable by antibodies to MP26, in the reactions containing membranes and threose (lane 3) or membranes, a-crystallin and threose (lanes 4-6). Increased formation of polymers and high molecular weight products (HMWP) that did not enter the spacer gel were observed when lens membranes and different amounts of a-crystallin (0.5 mg, lane 4; 1.0 mg, lane 5 and 5.0 mg, lane 6) were glycated by threose. It is also possible that the HMWP that did not enter the spacer gel might be washed off during staining and destaining of the acrylamide gel. The supernatant obtained from the reaction containing lens membranes and a-crystallin (without threose) predominantly consisted of a-crystallin, with no impurities of MP26 (lane 7). This emphasizes that non-covalently bound a-crystallin was eliminated from the membranes during the alkali-wash of those membranes. Lanes 9- 11 represent the membrane preparations, identical to those in lanes 4-6, but probed with antibodies to cr-crystallin. These results show that antibody detectable cr-crystallin was present in the membrane-bound fraction as HMWP, and the intensity of the immunostaining of the HMWP was dependent upon the amount of a-crystallin in the reaction. In another experiment, a-crystallin alone was first glycated with threose and then incubated with non-glycated membranes. Alkali-washed membranes isolated from this preparation did not show any bound a-crystallin by immunoblot assay (data not shown). These results argue for
M. Prabhakaram et al. 1 Mechanisms of Ageing and Development 91 (1996) 65-78
11
HWWP
POLYMERS
0
‘DIYERS
0 c x
r’ .MPPB
-a
123
4
567891011
Fig. 2. Immunoblot showing the antibody reactivity of MP26 (A) and a-crystallin (B) in the alkaliwashed ‘bovine lens membranes isolated from a 2 week reaction containing a-crystallin, membranes and threose as detailed in the Section 2.2. Membranes with N 25 pg of protein were analyzed by SDS-PAGE and then electroblotted to a PVDF membrane. Except for lane 7, all other lanes represent the SDS-solubilized membranes. The blot shows the incubations containing membranes alone or with cc-crystallin (lane 2); membranes and threose (lane 3); membranes, threose and different amounts of a-crystallin (0.5 mg, lane 4; 1.0 mg, lane 5; and 5.0 mg, lane 6). Lane 7 is the initial supematant obtained from the reaction containing membranes and a-crystallin, but not threose. The description for lanes 8-l 1 is the same as for the lanes 3-6, but probed with antibodies to a-crystallin.
the lack of non-specific interaction between pre-glycated a-crystallin and non-glycated MP26 in bovine lens membranes. Therefore, glycation of both cl-crystallin and MP26 is required for the crosslinking of these proteins. 3.3. Distribution of a-crystallin on the glycated membranes To further confirm that a-crystallin was covalently bound to membranes during glycation, immunogold labeling of the alkali- washed glycated membranes was perfonmed. MP26-Ab and CI-crystallin-Ab complexes were labeled with protein-A conjugated to 5 nm and 20 nm gold particles, respectively (Fig. 3A and B). In membranes incubated with a-crystallin in the absence of threose, only anti-MP26 immunolabel was detected, indicating that a-crystallin did not bind to the membranes in an alkali-resistant manner in the absence of threose (Fig. 3B). On the other hand, membranes incubated with both threose and a-crystallin and then subjected to alkali washing were labeled by both anti-MP26 and anti cc-crystallin
72
M. Prabhakaram et al. 1 Mechanisms
ofAge&
and Development 91 (1996) 65- 78
antib odies (Fig. 3A). The immunolabels for MP26 (5 nm) and a-crystallin (20 nm) were for the most part co-localized in the glycated membranes. These results suggt :st that the crosslinking of a-crystallin to the membranes was primarily
Fig. 3. Electron micrographs showing the a-crystallin bound to bovine lens membranes during glycation. Alkali-.washed lens membranes were isolated from 2 week reactions containing cc-crystallin and membranes which were incubated in presence (A) and absence (B) of threose. The membranes were probed with a ntibodies to MP26 or a-crystallin and localized by protein A-conjugated to 5 or 20 nm gold, respecl :ively, as described in Section 2.2. Arrow heads indicate MP26, whereas arrows indicate a-crystallin.
M. Prabhakaram et al. / Mechanisms of Ageing and Development 91 (1996) 65-78
a Cryst
13
+ Memb + Thr
I-- ..-_..-a-CrySt
360
i 240 1
1
120
O-
I 92
&ince
I 4
I 6
I a
lo
on SDS gel (cd
Fig. 4. Radioactive detection of iodinated a-crystallin on SDS gels. [‘251]cc-crystallinwas incubated alone (bottom panel), with threose (middle panel), or with membranes and threose (top panel) for 2 weeks as described in Section 2.2. 100 ,ug of aliquot from each reaction was analyzed by SDS-PAGE and radioactivity was determined by AMBIS image scanning of the dried SDS gel.
through MP26. Since the a-crystallin was not removed crosslinking between cc-crystallin and MP26 is considered nature,
by the alkali wash, as being covalent in
3.4. Interaction between [1251]u-crystallin and membranes during glycation In order to evaluate the glycation dependent crosslinking between a-crystallin and MP26, bovine lens membranes and [1251]cr-crystallin were glycated by threose and thlen analyzed by SDS-PAGE. Radioactive scanning of the dried SDS gel (Fig. 4) shows that the incubation of [‘251]a-crystallin alone or with membranes in the absence of threose (Fig. 4, bottom panel) did not produce any crosslinks. Under these conditions, radioactivity was solely present in the subunits of a-crystallin. When [1251]cr-crystallin alone was reacted with threose for 2 weeks, crosslinks were
74
M. Prubhakaram et al. / Mechanisms of Age&
and Deoelopment 91 (1996) 65- 78
readily formed (Fig. 4, middle panel), and the radioactivity was primarily distributed in the oligomers. In contrast, incubation of [“51]a-crystallin, membranes and threose (Fig. 4, top panel) exhibited higher radioactivity in the HMWP fraction than in the oligomers. It appears that in the presence of membranes and threose, glycated a-crystallins not only crosslink among themselves, but they also crosslink with MP26 to favor the formation of increased amounts of HMWP. Moreover, the [i2’I]m-crystallin glycated by threose mainly crosslinked as oligomers, but in presence of threose and membranes, the HMWP fraction was increased 2-3 fold, with this fraction containing both a-crystallin and MP26. Since a-crystallin has more glycation sites (a total of -v 300 lysines/molecule of 8 x 10’ Da a-crystallin) than MP26 (3 lysines/molecule of 26 kD MP26), the formation of HMWP would likely require a-crystallin to crosslink multiple MP26 molecules on the surface of the membrane. 3.5. Slot-blot immunoassay of human lens membranes Previously, it has been shown that both MP26 and a-crystallin can be detected by specific antibodies in the whole lens preparations of human nuclear cataracts [21]. In order to determine whether cr-crystallin is covalently bound to lens membranes in vivo, we have analyzed alkali-washed human lens membranes from donors of different ages. These results show that antibody detectable a-crystallin was present in alkali-washed membranes isolated from young (26 years and 41 years), old (67 years), aged or cataractous (84 years) and diabetic (68 years) human lenses as determined by slot-blot immunoassay (Fig. 5). The amounts of cc-crystallin bound to lens cell membranes were significantly lower in the samples from 26, 41 and 67 year old human donors than in the aged or cataractous (84 years) lens membranes. In addition, the intensity of cr-crystallin immunolabeling was higher in the 68 year old diabetic lens membranes than in the age-matched 67 year old lens membranes from a normal donor. Similar staining intensities were observed when the human lens membranes of different ages were probed with antibodies to MP26 (Fig. 5, bottom panel), indicating that similar amounts of lens membrane were present in each sample.
Anti a-Crystrllin
Anti MP26
Fig. 5. Antibody detection of cc-crystallin and MP26 in different ages of human lens membranes by slot-blot immunoassay. Alkali-washed human lens membranes were isolated as described in Section 2.2 and an equal volume (10 pl/lens) of SDS solubilized membranes was used for the antibody detection of cc-crystallin (top panel) and MP26 (bottom panel).
M. Prabhakaram
et al. / Mechanisms
of Ageing and Development
91 (1996) 65-78
15
4. Discussion The development of senile cataracts is an age-dependent process that can be influenced by environmental insults. Several post-translational modifications of lens proteins may be involved in the formation of high molecular weight aggregates or the water insoluble fraction in the human lens. Modifications, such as those caused by oxidation, UV-irradiation, disulfide formation, hydrophobic interactions and non-enzymatic glycation, have been proposed to be involved in the formation of high molecular weight protein aggregates in vivo [l]. Many of these insults may be involved in the crosslinking of soluble lens crystallins with fiber cell membrane components which accumulate as high molecular weight aggregates during normal aging and in senile cataracts. The purpose of the this work was to investigate the potential contribution of non-enzymatic glycation in the crosslinking of a-crystallin to MP26, an intrinsic lens fiber cell membrane protein. By incubation of soluble bovine lens proteins with the degradation products of ASA, it has been shown that high molecular weight protein aggregates can be generated in vitro [9,10,15]. ASA and its degradation products, such as DHA, DKG, xylosone or threose, have been shown to rapidly glycate lens proteins under in vitro conditions [9]. It has been shown recently that lens proteins glycated by ASA generate UVA sensitizers and high molecular weight proteins which resemble those present in aged human lenses [22]. All the breakdown products of ASA glycate and crosslink lens proteins [9] or membrane proteins [lo]. Of these ascorbate breakdown products, threose, a four carbon compound, is the most reactive with proteins and it can be formed either from ASA or DHA [23]. In general, carbohydrates with shorter chain lengths are more reactive with protein amino groups [12]. We have observed that either a-crystallin or MP26/MP22 reacts with the six-carbon compound DHA at a lower rate th.an with threose (Fig. 1). The physiological levels of threose are difficult to assess because of the rapid reactivity of threose with free amino groups of proteins. By using [14C]threose, we observed previously that threose is readily incorporated into human lens proteins [9]. In the present study, we have used intact bovine lens membranes to assess whether glycation mediated crosslinking can occur between the soluble lens protein cc-cryaallin and the major lens membrane intrinsic protein MP26/22. The membranes used in the crosslinking studies were treated with alkali to remove any non-covalently bound membrane components [ 10,161. However, a small amount ( < 1%) of other membrane components, such as cytoskeletal elements. beaded filaments and MP70, are usually present, even in the alkali washed membranes [l]. These membrane components may also undergo glycation in vivo. By incubating alkali-washed membranes, we have observed that either MP26 or MP22 of bovine lens membranes crosslink and accumulate as high molecular weight aggregates during glycation. Glycation mediated crosslinking of a-crystallin was also observed with these membrane proteins. Alkali-treatment of lens membranes glycated in the presence of a-crystallin eliminated non-covalently bound a-crystallin from membranes. Thus, the presence of cc-crystallin in the membrane fractions after glycation indicates that a fraction of the a-crystallin had become covalently bound.
16
M. Prabhakaram et al. / Mechanisms of Ageing and Deuelopment 91 (1996) 65-78
Our results also show that during glycation of lens membranes in the presence of a-crystallin, a fraction of MP26 remained monomeric, whereas no monomeric MP26 remained after glycation of the membranes in the absence of a-crystallin (Fig. 1). This suggests that a-crystallin and MP26 (or MP22) compete for the available DHA or threose. In bovine lens, a-crystallin has a J4, of 800 kD and is composed of 40 subunits with a total of 280-290 lysines [l]. The intrinsic membrane protein MP26 has a M, of 26 kD and it contains 3 lysines along with a free N-terminal amino group [24]. Due to the differences in molecular mass of a-crystallin and MP26, stoichiometrically, even a single a-crystallin molecule can span a number of MP26 molecules on membranes. Therefore, the freely accessible lysines on cc-crystallin are glycated initially, followed by the crosslinking of these lysines to the lysines of additional cr-crystallins or those of MP26/MP22 on the membranes. With time, these crosslinked proteins aggregate and accumulate as high molecular weight complexes. Recently it has been shown that cc-crystallin contains preferential, but common, glycation sites to react with ASA [25] or glucose [26]. By employing double immunolabeling techniques and electron microscopic studies, we have observed both MP26 and M-crystallin on the glycated bovine lens membranes. By the BCA assay method we have estimated that not more than - 40-50 ,ug of cr-crystallin was covalently bound to 1 mg of membrane bound MP26 during glycation. In contrast, it has been shown that as much as 5 mg of native a-crystallin can bind non-specifically to 1 mg of MP26 in non-alkali washed bovine lens membranes [27]. We have also observed an age-dependent increase in the protein content of WIF in the human lenses by the BCA assay method. The protein content in the WIF of 26 year human lens was 1 mg/lens, whereas 4.2 mg/lens was present in the WIF of 85 year old cataractous lenses. The increased intensity of a-crystallin immunostaining in the alkali-washed membranes of human cataracts could represent the increased amount of cr-crystallin bound to the aged lens membranes in the WIF. The increased amounts of a-crystallin bound to the diabetic lens membranes suggests that glycation may be involved, since non-enzymatic glycation is implicated in diabetics [12,13]. Increased levels of DHA, the precursor of threose, have been reported in nuclear and diabetic cataracts [28,29], suggesting the possibility of ascorbate mediated glycation in vivo. A massive accumulation of cr-crystallin in aged or cataractous human lens membranes [4,5], and also the interaction of phospholipid vesicles containing MP26 with cc-crystallin [6], suggest the crosslinking ability of a-crystallin to membrane components in vivo. The failure of alkali washing to remove the bound a-crystallin in the glycated bovine lens membranes or in cataractous human lens membranes suggests that a-crystallin is covalently bound to membranes during glycation or aging, respectively. In addition to cr-crystallin, significant amounts of j3 and y-crystallins have been detected in the WIF of both normal and cataractous human lens membranes [7]. The presence of these crystallins on membranes, even after treatment of P-mercaptoethanol or alkali washing [21], indicate that a mechanism(s) other than simple hydrophobic interaction or disulfide crosslinking is involved in the crosslinking of crystallins to membrane components in vivo.
M. Prabhakaram et al. / Mechanisms of Ageing and Development 91 (1996) 65-78
71
Previously, we have shown that blocking the available -SH groups in the bovine lens soluble proteins did not prevent the crosslinking properties of cr-crystallin in the presence of ASA or its breakdown products [30]. Based upon the present findings, we hypothesize that the in vivo crosslinking of crystallins to lens membrane-bound proteins involves glycation. With advancing age, the highly crosslinked membrane-protein complexes appear to aggregate and contribute to the lens opacification in senile cataractogenesis.
Acknowledgements The authors thank Dr. Larry Takemoto for generously supplying the MP26 antibodies and Chun-Lan Gao for her assistance in electron microscopy. This work was supported by NIH grants EY07070 (BJO), EY08813 (MLK) and in part by Research to Prevent Blindness.
References [l] E.R. Berman, In E.R. Berman (ed.), Biochemistry of the Eye, Plenum Press, New York, 1991, pp. 201-290. [2] A. Pirie, Color and solubility of the proteins of human cataracts. Invest. Ophthalmol., 7 (1968) 634-650. [3] J. Bours, Isotachophoresis
and immunoelectrophoresis of water soluble and insoluble crystallins of the ageing bovine lens. Curr. Eye Res., 3 (1984) 691-697. [4] M. Takehana and L. Takemoto, Quantitation of membrane-associated crystallins from aging and cataractous human lenses. Invest. Ophthalmol. Vis. Sci., 28 (1987) 780-784. [5] G. Chandrasekhar and R.J. Cenedella, Protein associated with human lens ‘native’ membrane during aging and cataract formation. Exp. Eye Res., 60 (1995) 707-717. [6] J.W.M. Mulders, J. Stokkermans, J.A.M. Leunissen, E. Benedetti, H. Blomendal and W.W. DeJong, Interaction of cc-crystallin with lens plasma membranes. Affinity for MP26. Eur. J. Biochem., 152 (1985) 721-728. [7] F. Ifeanyi and L. Takemoto, Interaction
[8] [9] [lo] [ll] [12] [13]
[14]
of lens crystallins with lipid vesicles. Exp. Eye Res., 53 (1931) 535-538. CF:. Fleschner and R.J. Cenedella, Examination of a lens native plasma membrane fraction and its associated crystallins. Curr. Eye Res., I1 (1992) 739-752. B.J. Ortwerth, J.A. Speaker, M. Lopez, E. Li and M.S. Feather, Ascorbic acid glycation; the rea’ction of L-threose in lens tissue. Exp. Eye Res., 58 (1994) 665-674. M. Prabhakaram and B.J. Ortwerth, Glycation of MP26 and MP 22 in bovine lens membranes. Biochem. Biophys. Res. Commn., 185 (1992a) 4966504. L.J. Takemoto, J.S. Hansen and J. Horwitz, Inter species conservation of the main intrinsic polypeptide (MIP) of the lens membrane. Camp. Biochem. Physiol., 68B (1981) 101-106. V.M. Monnier, In J.E. Baynes and V.M. Monnier (eds), The Maillard Reaction in Aging, Diabetes and Nutrition, Alan R. Liss, New York, 1989, pp. l-22. R.E. Perry, M.S. Swamy and E.C. Abraham, Progressive changes in lens crystallin glycation and high molecular weight aggregate formation leading to cataract development in streptozotocin-diabetic rats. Exp. Eye Res., 44 (1987) 269-282. B.J. Ortwerth, P.R. Olesen, K.K. Sharma and M. Prabhakaram, Chemical modification of alpha crystallin. Exp. Eye Res., 56 (1993) 107-l 14.
78
M. Prabhakaram et al. 1 Mechanisms 01 Ageing and Development 91 (1996) 65-78
[15] M. Prabhakaram and B.J. Ortwerth, The glycation and cross-linking of isolated lens crystallins by ascorbic acid. Exp. Eye Res., 55 (1992b) 451-459. [I61 P. Russell, W.G. Robison and J.H. Kinoshita, A new method for rapid isolation of the intrinsic membrane proteins from lens. Exp. Eye Res., 32 (1981) 51 l-516. [17] U.K. Laemmli, Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature, 277 (1970) 680-685. [18] H. Towbin, T. Staehelin and J. Gordon, Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose: procedure and some applications. Proc. Natl. Acad. Sci. USA, 76 (1979) 4350-4354. [19] M.L. Katz, C. Gao and H.J. Stientjes, Regulation of the inter photoreceptor protein content of the retina by vitamin A. Exp. Eye Res., 57 (1993) 393-401.
retinoid-binding
[20] F.C. Greenwood, W.M. Hunter and J.S. Glover. The preparation of ‘311-labeled human growth hormone of high specific radioactivity. Biochem. J., 89 (1963) 114-123. [21] L. J. Takemoto, J.S. Hansen and J. Horwitz, High molecular weight aggregates from human cataracts: characterization by Western Blot analysis. Biochem. Biophys. Res. Commn., 122 (1984) 1028-1033. [22] B.J. Ortwerth, M. Linetsky and P.R. Olesen, Ascorbic acid glycation of lens proteins produces UVA sensitizers similar to those in human lens. Photochem. Photobiol., 62 (1995) 454-462. [23] M.G. Lopez and M.S. Feather, The production of threose as a degradation product from L-ascorbic acid. J. Carbohydr. Chem., II (1992) 799-800. [24] M.B. Gorin, S.B. Yancey, J. Cline, J.P. Revel and J. Horwitz, The major intrinsic protein (MIP) of the bovine lens fiber membrane. Characterization and structure based on cDNA cloning. Cell, 39 (1984) 49-59. [25] B.J. Ortwerth, S.H. Slight, M. Prabhakaram, Y. Sun and J.B. Smith, Site specific glycation of lens crystallins by ascorbic acid. Biochim. Biophys. Acta, 1 I1 7 (1992) 207-215. [26] E.C. Abraham, M. Cherian and J.B. Smith, Site-specificity in the glycation of stA and ctB crystallins by glucose. Biochem. Biophys. Res. Commn., 201 (1994) 1451-1456. [27] R.J. Cenedella and G. Chandrasekhar, High capacity binding of alpha crystallins to various bovine lens membrane preparations. Curr. Eye Res., 12 (1993) 1025-1038. [28] W. Lohmann, W. Schmehl and J. Strobel, Nuclear cataract. Oxidative damage to the lens. Exp. Eye Res., 43 (1986) 859-862. [29] S. Som, D. Mukherjee, S. Deb, R. Choudhury, S.N. Mukherjee, S.N. Chatterjee and I.B. Chatterjee, Ascorbic acid metabolism in diabetes mellitus. Metabolism, 30 (1981) 572-577. [30] M. Prabhakaram and B.J. Ortwerth, The glycation-associated crosslinking of lens proteins by ascorbic acid is not mediated by oxygen free radicals. Exp. Eye Res., 53 (1991) 261-168.