Binding Capacity of α-Crystallin to Bovine Lens Lipids

Exp. Eye Res. (1996) 63, 407–410

Binding Capacity of α-Crystallin to Bovine Lens Lipids D O U G L A S B O R C H M A N*    D A X I N T A N G Department of Ophthalmology and Visual Sciences, University of Louisville, Louisville, KY 40202, U.S.A. (Received Rochester 11 December 1995 and accepted in revised form 29 March 1996) Three experiments were performed to determine the α-crystallin binding capacity of bovine lens lipid vesicles. In one experiment lipid was kept constant (2±5 mg ml−") and the α-crystallin concentration was changed (0±5 to 3±0 mg ml−"). In another experiment, α-crystallin was kept constant (1 mg ml−") and the concentration of lipid was varied (0±25 to 3 mg ml−"). We calculated the binding capacity of the lipid to be 0±33³0±05 (..) mg α-crystallin (mg lens lipid)−". This was confirmed by changes in the anisotropy and fluorescent intensity of a probe that partitions at the headgroup region of the lipid bilayer. Near 0±33 mg α-crystallin (mg lens lipid)−" the fluorescence intensity and anisotropy of the probe increases and plateaus which indicates that concomitant with α-crystallin binding, water is excluded from the head group region of the bilayer and the headgroup region becomes less mobile. It is possible that α-crystallin binding could protect and stabilize the lipid bilayer and decrease membrane permeability. # 1996 Academic Press Limited Key words : lens ; α-crystallin ; lipids ; binding.

1. Introduction

2. Materials and Methods

Human lens alpha crystallin concentration may be as high as 40 % of the total protein and is the major extrinsic protein of lens membranes (Bloemendal et al., 1972 ; Chandrasekher and Cenedella, 1995 ; Fleschner and Cenedella, 1992). When solubilized, α-crystallin can function as a molecular chaperone that inhibits the heat-induced aggregation of other crystallins and proteins (Horwitz, 1992). Alpha-crystallin–lens membrane binding has been the focus of a number of recent studies (Cenedella and Chandrasekher, 1993 ; Ifeani and Takemoto, 1989, 1990a, 1990b, 1991a, 1991b ; Mulders et al., 1985, 1989 ; Liang and Li, 1992 ; Ramaekers, Versteegen and Bloemendal, 1980 ; Zhang and Augusteyn, 1994). Mulders et al. (1985) showed that association of α-crystallin with lens membranes was temperature, pH and time dependent. The interaction of α-crystallin with bovine lens membranes is age dependent (Ifeani and Takemoto, 1989) and lipid alone may be all that is necessary for αcrystallin–membrane binding since α-crystallin may bind to pure phospholipid vesicles (Ifeani and Takemoto, 1991a). When bound α-crystallin immobiles lipids (Liang and Li, 1992), but does not alter the lipid hydrocarbon chain structure (Sato et al., 1996). In semiquantitative experiments it has been demonstrated that α-crystallin may bind to lens lipids at a capacity five times higher than that of vesicles made from phosphatidylcholine (Sato et al., 1996). The purpose of this study was to determine the binding capacity of bovine α-crystallin to bovine lens lipids.

All reagents and bovine α-crystallin were purchased from the Sigma Chemical Co. (St. Louis, MO, U.S.A.) except where indicated. Bovine lens lipids were extracted from eyes obtained fresh from a slaughterhouse. N-(7-nitrobenz-2-oxa-1,3-diazol-4-yl)-1,2-dihexadecanoyl-sn-glycero-3phosphoethanolamine, triethylammonium salt (NBD-PE) was purchased from Molecular Probes (Eugene, OR, U.S.A.). A monophasic methanolic extraction followed by a hexane} isopropanol purification was used to extract lipid from 60 bovine lenses (Sato et al., 1996). Three experiments were performed to determine the α-crystallin binding capacity of bovine lens lipid vesicles. Binding was determined at 36°C in a 5 m [4-(2-hydroxyethyl)-1-piperazine-ethanesulfonic acid], buffer, pH ¯ 7.5, (Hepes), bubbled with argon gas for 10 min to remove oxygen.

* For correspondence at : Department of Ophthalmology and Visual Sciences, 301 E. Muhammad Ali Blvd., Louisville, KY 40202, U.S.A.

0014–4835}96}100407­04 $25.00}0

Experiment 1 Lipid was kept constant (2±5 mg ml−") and the αcrystallin concentration was changed (0±5 to 3±0 mg ml−"). Experiment 2 Alpha-crystallin was kept constant (1 mg ml−") and the concentration of lipid was varied (0±25 to 3 mg ml−"). The following protocol was followed for experiments 1 and 2 : stock bovine lipid in methanol (1 mg ml−") was added to 1 ml capacity ultracentrifuge tubes. The samples were frozen in liquid nitrogen and freeze dried # 1996 Academic Press Limited

408

Controls To test whether sonication influenced binding, 0±5 mg ml−" of α-crystallin and lipid were prepared as described above, but in one set of samples the lipid was sonicated in the absence of α-crystallin, and then mixed with the α-crystallin, equilibrated and the binding determined. To determine if α-crystallin became insolubilized during the prolonged incubation period, nine 1 mg ml−" α-crystallin solutions were prepared. Three samples were equilibrated at 36°C for 12 hr, three were equilibrated at 25°C and three samples were prepared within 1 hr of protein measurement. Experiment 3 The fluorophore, NBD-PE, that partitions near the phospholipid head group region was mixed with bovine lipid to indirectly measure the maximum binding capacity of α-crystallin to bovine lipid. When α-crystallin binds to the lipid membrane, the anisotropy (1}wobble) of the probe and the fluorescence would be expected to increase due to the immobility of the probe and the exclusion of water, respectively. The fluorescent probe, NBD-PE, was mixed with bovine lipid in chloroform at a weight ratio between 0±005 and 0±02 to 1 lipid. Samples were prepared as described in the protocol for Experiment 2 except sonication was not used and a buffer solution consisting of 10 m Tris(hydroxymethyl)aminomethane hydrochloride (Tris–HCl), pH 7.5, 0±1 m ethylenediaminetetraacetic acid (EDTA) and 50 m KCl was used to eliminate the possibility of interference due to calcium– fluorophore interactions. Fluorescence Measurements Anisotropy and intensity measurements were performed on an ISS PC1 photon counting spectro-

fluorometer (Champagne, IL, U.S.A.) with a polarization accessory unit. The excitation and emission wavelengths used were 460 and 540 nm, respectively, for the NBD-PE probe to detect environmental and structural changes near the bilayer surface. Fluorescence anisotropy, r, was calculated by r ¯ (Ill®gIv)}(Ill­2gIl)

in which g ¯ Iv}Ill (1)

The fluorescence intensity was measured as the ratio of the sample detector signal and the reference detector signal. 3. Results The plot of α-crystallin bound per lipid verses 1}αcrystallin free (Fig. 1) was used to estimate the maximum binding capacity of α-crystallin to bovine lipid by extrapolating the linear curve through our data to the y axis. The maximal binding capacity of bovine lens lipid to α-crystallin was estimated from the data in Fig. 1 to be 0±34³0±4 (..) mg α-crystallin (mg bovine lens lipid)−". The binding capacity of α-crystallin to bovine lens lipids was determined by varying the concentration of bovine lens lipid and by keeping the concentration of α-crystallin constant and above the saturation concentration. From the data obtained using this methodology, we calculated the binding capacity of the lipid to be 0±33³0±05 (..) mg α-crystallin (mg lens lipid)−", n ¯ 5. The capacity calculated by both methods was almost identical. Samples prepared with and without sonication gave similar binding values (³6 %) which indicates sonication did not influence the binding capacity of αcrystallin to the bovine lipids. The sonication step was eliminated in the fluorescence measurement protocol. 0.4

α-Crystallin bound/lipid (g/g)

in a lyophilizer for 20 min to remove methanol. Hepes buffer then stock α-crystallin (10 mg ml−") was then added. The centrifuge tubes were sealed in an atmosphere of argon mixed and sonicated in a bath sonicator for 15 min then allowed to equilibrate for 12 h at 36°C with gentle shaking. To remove lipid and bound α-crystallin from solution, the samples were centrifuged at 170 000 g for 1±5 h at 36°C. Alpha crystallin in the supernatant was determined by measuring the optical density at 280 nm. The extinction coefficient at 280 nm for the α-crystallin was calculated to be 0±645 O.D. mg ml−"³0±015 (...) n ¯ 28. The absorbance of the supernatant was corrected for a small absorbance due to chromagens from the lipid which had an extinction coefficient of 0±125 O.D. mg−" ml−"³0±002 (...), n ¯ 17. Protein concentration was also quantified using the Peterson assay (Peterson, 1977) with identical results.

D. B O R C H M A N A N D D. T A NG

0.3

0.2

0.1

0

1 2 1/α-Crystallin free (ml mg–1)

3

F. 1. Binding of bovine lens α-crystallin to bovine lipids, 36°C. The concentration of lipid was kept constant.

α-CRYSTALLIN–L I P I D B I N D I NG 1.20

409

(A)

1.18 1.16 Fluorescence

1.14 1.12 1.10 1.08 1.06 1.04 1.02 1.00 0.0 0.10

0.2

0.4

0.6

0.8

1.0

(B)

Anisotropy

0.09

0.08

0.07

0.06

0.05 0.0

0.2 0.4 0.6 0.8 1.0 Alpha-crystallin/lens lipid (w/w)

F. 2. Change in the (A) fluorescence intensity and (B) anisotropy of the head group lipid probe, NBD-PE, with α-crystallin binding. An increase in fluorescence indicates exclusion of water. An increase in anisotropy indicates a decrease in the mobility of the probe.

Less than 1 % of the α-crystallin became insoluble after incubation in an atmosphere of argon after a 12 hr incubation at either 21 or 36°C. An incubation period of over 3 hr at 36°C was beneficial for increasing the solubility of the α-crystallin by about 50 %, even after sonication. Figure 2(A) and (B) shows the change in NBD-PE probe fluorescence and anisotropy as the ratio of αcrystallin to lens lipid was increased from 0 to 1±0 (w}w), respectively. Note the plateau that occurs above the α-crystallin to lipid weight ratio of 0±3 to 0±4 which indicates that the maximum binding occurs at this weight ratio. The increase in fluorescence with αcrystallin binding [Fig. 2(A)] indicates that in the process of binding, water is excluded from the head group region of the lipid bilayer. The increase in anisotropy with α-crystallin binding [Fig. 2(B)] indicates that the probe becomes less mobile. 4. Discussion We found that bovine lens lipids have a high capacity to bind α-crystallin, 0±33 mg α-crystallin (mg lens lipid)−". Considering that the human lens contains

at most, only 1±2 mg of lipid, from our binding capacity data from bovine lens material, if we may speculate about the binding in the human lens, we calculate that only a small proportion, 0±4 mg, of the total 32 mg of human lens α-crystallin would be directly bound to the membrane lipid. It has been suggested that HMW aggregates of crystallins associated with human lens membranes could perhaps assemble on the membrane after binding (Chandrasekher and Cenedella, 1995). Thus a large amount of α-crystallin could be indirectly bound to the membrane via the small amount of α-crystallin bound directly to the membrane lipids. This study supports the finding using non-lens lipids, that intrinsic proteins may not be necessary for α-crystallin binding (Ifeanyi and Takemoto, 1991a). The α-crystallin–lipid binding capacity reported in this paper is about five times higher than previously reported (Ifeanyi and Takemoto, 1991a) possibly because we used a much higher concentration of αcrystallin, 0±5–3 mg α-crystallin ml−" compared to 0±36 mg α-crystallin ml−" used in previous studies and the amount of lipid used in this study was also often higher, 2±5 mg ml−", compared to 0±784 mg phospholipid ml−" used in previous studies and}or purified bovine lipid membranes devoid of protein were used in this study. Perhaps more important than the actual amount of α-crystallin binding, is the possibility that α-crystallin could stabilize and protect the lipid bilayer. Our fluorescent probe data [Fig. 2(A)] indicate that upon bind of α-crystallin, water is excluded from the lipid headgroup region of the bilayer which could protect the lipid from hydrophilic oxidants such as H O and # # oxidized ascorbate. Exclusion of water and immobilization of the lipid head groups with αcrystallin binding would also be expected to decrease the permeability of the bilayer to hydrophilic cations and water. This hypothesis is currently being tested using large unilamellar vesicles. Head group interactions appear to be more important to α-crystallin binding since binding does not affect the hydrocarbon structure of the membrane at 36°C (Sato et al., 1996). At 36°C, lens lipids are near the center of their order to disorder transition, the most sensitive point for changes in lipid structure, (Borchman et al., 1993). Perhaps the role of α-crystallin-lipid binding is to stabilize the membrane lipid structure as it does to lens proteins. Further studies are needed to determine factors influencing α-crystallin–lipid binding, the affect of α-crystallin on lipid structural stability and the influence of intrinsic proteins on the binding capacity and binding constant.

Acknowledgements Supported by Public Health Service research grant EYO7975 (Bethesda, MD, U.S.A.) and the Kentucky Lions Eye Foundation (Louisville, KY, U.S.A.), and an unrestricted

410

grant from Research to Prevent Blindness, Inc. Chris John should be acknowledged for his technical help.

References Borchman, D., Lamba, O. P. and Yappert, M. C. (1993). Structural characterization of human lens membrane clear and cataractous lipid. Exp. Eye Res. 57, 199–208. Bloemendal, H., Zweers, A., Vermorken, F., Dunia, I. and Benedetti, E. L. (1972). The plasma membranes of eye lens fibers. Biochemical and structural characterization. Cell Diff. 35, 61–7. Fleschner, C. R. and Cenedella, R. J. (1992). Examination of ‘ native ’ plasma membrane fractions and its associated crystallins. Curr. Eye Res. 11, 739–52. Cenedella, R. J. and Chandrasekher, G. (1993). High capacity binding of alpha crystallins to various bovine lens membrane preparations. Curr. Eye Res. 12, 1025–38. Chandrasekher, G. and Cenedella, R. J. (1995). Protein associated with human lens ‘ native ’ membrane during aging and cataract formation. Exp. Eye Res. 60, 707–17. Horwitz, J. (1992). α-Crystallin can function as a molecular chaperone. Proc. Natl. Acad. Sci. U.S.A. 89, 10449–53. Ifeanyi, F. and Takemoto, L. (1989). Differential binding of α-crystallins to bovine lens membrane. Exp. Eye Res. 49, 143–7. Ifeanyi, F. and Takemoto, L. (1990a). Specificity of αcrystallin binding to the lens membranes. Curr. Eye Res. 9, 259–65.

D. B O R C H M A N A N D D. T A NG

Ifeanyi, F. and Takemoto, L. (1990b). α-Crystallin from human cataractous versus normal lenses : change in binding to lens membrane. Exp. Eye Res. 50, 113–6. Ifeanyi, F. and Takemoto, L. (1991a). Interaction of lens αcrystallin with lipid vesicles. Exp. Eye Res. 53, 535–8. Ifeanyi, F. and Takemoto, L. (1991b). Involvement of the Nterminal region in α-crystallin–lens membrane recognition. Exp. Eye Res. 52, 305–8. Liang, J. J. N. and Li, X. (1992). Spectroscopic studies on the interaction of calf lens membranes with α-crystallins. Exp. Eye Res. 54, 719–24. Mulders, J. W. M., Stokkermans, J., Leunissen, J. A. M., Benedetti, E. L., Bloemendal, H. and De Jong, W. W. (1985). Interaction of α-crystallin with plasma membranes. Eur. J. Biochem. 152, 721–8. Mulders, J. W. M., Wojcik, E., Bloemendal, H. and De Jong, W. W. (1989). Loss of high-affinity binding of bovine nuclear α-crystallin. Exp. Eye Res. 49, 149–52. Peterson, G. L. (1977). A simplification of the protein assay method of Lowry et al. which is more generally applicable. Anal. Biochem. 83, 346–57. Ramaekers, F. C. S., Selten-Versteegen, A. E. and Bloemendal, H. (1980). Interaction of newly synthesized α-crystallin with isolated lens plasma membranes. Biochim. Biophys. Acta 596, 57–63. Sato, H., Borchman, D., Ozaki, Y., Lamba, O. P., Byrdwell, W. C., Yappert, M. C. and Paterson, C. A. (1996). Lipid–protein interactions in human and bovine lens membranes by Fourier transform Raman and infrared spectroscopes. Exp. Eye Res. 62, 47–53. Zhang, W. Z. and Augusteyn, R. C. (1994). On the interaction of α-crystallin with membranes. Curr. Eye Res. 13, 225–30.