Preferential Interaction of Alpha Crystallin With Denatured Forms of Gamma Crystallin

5. Gopalakrishnan, D. Boyle, and L. Takemoto

Purpose. To characterize the possible interaction of alpha crystallin with partially denatured forms of gamma crystallin. Methods. Gamma crystallin was denatured in the presence of guanidine hydrochloride, then dialy/ed in the presence or absence of alpha crystallin. The high-molecular-weight complex formed in the presence of alpha was characterized by gel filtration chromatography, electron microscopy, and quantitative Western blot analysis. Results. Relative to native alpha or reconstituted aggregates of purified alpha, the higher molecular weight complex possessed a greater mean diameter and contained increased amounts of gamma crystallin. Conclusions. Alpha crystallin preferentially interacts with partially denatured forms of a lens protein, consistent with its putative role as a functional molecular chaperone in the intact lens. Invest Ophthalmol Vis Sci. 1994;35:382-387.

1 he alpha crystallins are one of the most abundant Because the alpha crystallins can protect other classes of proteins in the lens. They comprise the al- proteins against heat-induced aggregation using an in pha-A and alpha-B chains, that exhibit extensive ho- vitro assay, and because the alpha crystallins are re- mology in their sequences.1 Besides acting as a me- lated in sequence to heat shock proteins, it is highly dium for the refraction of light, recent studies have possible that one of the major functions of the alpha demonstrated that the alpha crystallins can protect crystallins in vivo is to bind to and prevent further other polypeptides against the denaturing effects of denaturation of lens proteins that have been subjected heat-induced aggregation.23 These results are consis- to both physical and chemical forms of stress found in tent with sequence homologies between the alpha the intact lens. In this report, we demonstrate that the crystallins and previously characterized heat shock alpha crystallins do indeed interact directly and prefer- proteins,45 which are expressed in increased amounts entially with denatured forms of the lens gamma crys- during heat-induced stress of nucleated cells. tallins. The molecular chaperones consist of highly con- served groups of polypeptides that serve a variety of roles in the cell. They preferentially bind to partially MATERIALS AND METHODS denatured forms of polypeptides, protecting them Newborn bovine lenses were obtained from Antech, against further denaturation or facilitating protein Inc. (Tyler, TX), and were stored at —75°C until use. renaturation or both. They can also act as "chaper- In the authors' opinion, methods for obtaining tissue ones" by binding to partially denatured proteins and adhered to the ARVO Statement for the Use of Ani- facilitating their translocation to other parts of the cell mals in Ophthalmic and Vision Research. Alpha and (see references 6 and 7 for a review of these func- gamma crystallins were purified from the water-solu- tions). ble fraction of the lens cortex as described previously,8 using a TSK G3000SW column (Perkin-Elmer, Nor- From the Division of liiology. Kansas State University, Manhattan, Kansas. walk, CT). The amount of protein in each fraction was This research was supported by grants from the Nil! and NASA. Submitted for publication: May 13. 1993: revised July 16, 1993; accepted August determined according to the BCA method (Pierce 9, 1993. Chemical Co., Rockford, 1L), using bovine serum al- Proprietary interest category: N. Reprint requests: L. Takemoto, Division of liiology. Kansas State University, Ackert bumin as standard. After dialysis and lyophilization, Hall, Manhattan, KS66506--I901. the gamma crystallin fraction was reduced and carbox-

Investigative Ophthalmology & Visual Science. 1-Vbn.iary 1094, Vol. 3"). No. 2 382 Copyright © Association for Research in Vision and Ophthalmology

Downloaded from iovs.arvojournals.org on 09/29/2021 Alpha-Gamma Interaction 383

yamidated (RCA) in the presence of 7 M guanidine 1 = Void Volume hydrochloride,9 followed by dialysis and lyophiliza- tion. To determine the possible interaction of alpha 1.0- with different forms of gamma crystallin, 1.80 mg al- pha was mixed with 0.36 mg native or RCA gamma in 0.50 ml solution containing 6.0 M guanidine hydro- 0.5 • chloride, 10 mM Tris-HCl, pH 7.4. After incubation at 22°C for 15 minutes, the solution was dialyzed for 30 to 36 hours at 4°C against a solution containing 10 0.0- mM Tris-HCl, pH 7.4, with several changes of the 1.0- buffer, then dialyzed for 4 to 6 hours against TSK buffer (0.06 M sodium phosphate, 0.1 M sodium sul- E fate, pH 7.0). The dialysate was centrifuged at 10,000g 0.5 • for 5 minutes, and 50% of the supernatant was in- c jected into a Biosep S4000 gel permeation column o (300 mm X 7.8 mm, Phenomenex, Torrance, CA). 00 0.0- Proteins were resolved at a How rate of 0.5 ml/min, CM 1.0- using TSK buffer. LJJ The major peaks were collected, and 5% of the o material was precipitated in acetone to remove salts, 0.5 • then dissolved in sample buffer and analyzed by West- Z ern blot analysis using 15% (v/v) polyacrylamide gel < 10 electrophoresis. For the purposes of quantitation, CD 0.0- known amounts of alpha crystallins, gamma crystallins, 1.0 - or RCA gamma crystallins were resolved at the same (£ time. The blots were probed with rabbit polyclonal 0 antisera made against human alpha and bovine gamma Eye Institute. After binding of radioiodinated protein A, protein bands of the resulting autoradio- 0.0- graph were quantitated using scanning densitometry. Approximately 10% of the remaining material 1.0" from the S-4000 column was applied to Formvar- and carbon-coated copper grids, stained with uranyl ace- 0.5 • tate, exposed to osmium tetroxide vapor, and visual- ized by electron microscopy as previously de- scribed.^1) Statistical analysis of aggregate diameters was carried out according to the Mann-Whitney test, 0.0 - using the Number Cruncher Statistical Package (J.L. 0 10 15 20 25 Hintze, Kaysville, UT). RETENTION TIME (MINUTES) FIGURE 1. Gel permeation chromatography of alpha plus RESULTS gamma crystallin, with and without guanidine hydrochloride Figure .1 shows the profiles of various mixtures of al- treatment.. See Materials and Methods for details of renatur- pha and gamma crystallins, after their dialysis and reso- aiion and chromatography. (A) Alpha plus gamma, no guan- lution using a Biosep S4000 gel permeation column. idine hydrochloride treatment, the vertical arrow designates Figure 1A shows the elution times for a mixture of the void volume of the column; (B) alpha alone, guanidine hydrochloride treatment plus dialysis; (C) alpha plus gamma, alpha and gamma crystallins that was not subjected to guanidine hydrochloride treatment plus dialysis; (D) alpha prior denaturation with guanidine hydrochloride. The plus RCA gamma, guanidine hydrochloride treatment plus 15.6-minute and 21.6-minute elution times represent dialysis; (E) gamma alone, guanidine hydrochloride treat- the peaks for native alpha and native gamma crystal- ment plus dialysis, the vertical arrow designates the expected lins, respectively. Guanidine hydrochloride treatment elution time of the HMVVA. and dialysis of alpha alone (Fig. IB), followed by S-4000 chromatography results in a peak eluting at a similar time (15.4 minutes). lure, is a commonly used procedure to study their po- Denaturation of the binding protein by guanidine tential interactions.12 Under these conditions, the hydrochloride, followed by dialysis or dilution to rena- chaperone binds to proteins in various stages of rena-

Downloaded from iovs.arvojournals.org on 09/29/2021 384 Investigative Ophthalmology & Visual Science, February 1994, Vol. 35, No. 2

turation. In Figure JC, alpha crysiallins and unmodi- shown in Table 1. The HMWA plus native alpha peaks fied gamma crystallins were incubated in the presence from Figures 1C and ID contain significant amounts of 6.0 M guanidine hydrochloride, followed by dialysis of gamma crystallin (326 ± 24 /zg and 328 ± 12 ^g, and S-4000 column chromatography. Compared with respectively), when compared with the amounts of the 15.6-minute peak for native alpha in Figure 1A, gamma found in the major alpha peak obtained from guanidine hydrochloride treatment results in an ear- reconstituted alpha alone (3.4 ±0.3 /ug) or from native lier eluting peak (13.5 minutes), termed the high-mo- alpha incubated with native gamma (20.4 ± 3.0 fig). lecular-weight aggregate (HMWA) peak. Also present The amount of gamma crystallins present in the is a smaller peak eluting at 15.5 minutes, which proba- HMWA plus native alpha peaks accounts for approxi- bly represents uncomplexed alpha, and is termed the mately 90% of the material added to the original guan- native alpha peak. idine-hydrochloride-treated mixture, demonstrating In Figure ID, the gamma preparation was first that almost all the gamma has remained in solution by reduced and carboxyamidated before dialysis, to en- being complexed with alpha. sure that complete renaturation was blocked. Under these conditions, the HMWA eluting at 13.6 minutes DISCUSSION was also present when the dialysate was resolved by Recent studies using an in vitro assay have demon- S-4000 chromatography. strated that the alpha crystallins are able to protect Figure IE shows that guanidine hydrochloride other proteins from heat-induced denaturation and treatment of gamma alone, followed by dialysis, does aggregation.2-3 This observation suggests that in the not result in the HMA peak. The absence of a detect- intact lens, one of the major functions of alpha crysial- able peak represents the probable precipitation of lins is to protect lens polypeptides against the exten- gamma. Identical results were obtained for RCA sive amounts of denaturation that could eventually re- gamma (results not shown). The lack of a gamma or sult in cataractogenesis. Based on studies of molecular HMWA peak in Figure IE reflects the absence of al- chaperones from other cell types,13 the mechanism of pha, which binds to gamma and prevents precipitation this protection must involve direct interaction of alpha during the dialysis renaturation procedure. with the partially denatured lens protein. Together, the results shown in Figure 1 suggest To test this hypothesis, we characterized the bind- that after denaturation and partial renaturation in the ing of alpha crystallin to native versus denatured presence of alpha, some of the gamma may bind to forms of the gamma crystallin. We dissolved both al- alpha crystallins, preventing its precipitation, and pro- pha crystallin and gamma crystallin in 6.0 M guanidine ducing an alpha-gamma aggregate of higher molecu- hydrochloride, then renatured them by dialysis. The lar weight. results in Figures 1 to 3 clearly show that guanidine To verify the presence of these higher molecular hydrochloride treatment and subsequent dialysis of a weight aggregates, electron microscopy was used to mixture of alpha and gamma crystallin results in the compare their size with aggregates from native alpha. production of larger aggregates than those obtained Negative staining (Fig. 2) shows the presence of parti- after identical treatment of alpha alone. cles that could be measured and quantitated as shown The results suggest that, formation of the high-mo- in the histogram in Figure 3. Relative to native alpha lecular-weight aggregates is due to the preferential or reconstituted alpha alone (Fig. 3, A and B), guani- binding of alpha to denatured forms of gamma crystal- dine hydrochloride treatment and dialysis of alpha in lin, to produce a supramolecular complex containing the presence of gamma (Fig- 3C) or RCA gamma (Fig. both alpha and partially denatured forms of gamma 3D) produced aggregates with a greater range of size, crystallin. This conclusion is supported by the results resulting in a larger mean value (12.01 ± 3.66 in Fig. of Table 1, which demonstrate that the high-molecu- 3C; 11.98 ± 3.51 in Fig. 3D) as compared with 9.53 ± lar-weight aggregates contain much larger amounts of 2.31 in Figure 3A and 10.6 ± 3.07 in Figure 3B. Statis- gamma or RCA gamma than do the aggregates of na- tical analysis using the Mann-Whitney test demon- tive alpha or alpha reconstituted in the absence of strated that the populations of aggregates shown in gamma. Using identical conditions to denature and Figure 3C and Figure 3D were significantly larger than renature gamma crystallin in the presence of alpha, we those shown in Figure 3B (P

Downloaded from iovs.arvojournals.org on 09/29/2021 Alpha-Gamma Interaction 385

1

FIGURE 2, Electron microscopy of aggregate peaks from S-4000 chromatography. See Mate- rials and Methods for details of electron microscopy. (A) Alpha crystallin peak from alpha plus gamma, no guanidine hydrochioride treatment; (B) alpha alone, guanidine hydrochlo- rtde treatment plus dialysis; (C) HMWA plus native alpha peaks, alpha plus gamma, guani- dine hydrochioride treatment plus dialysis; (D) HMWA plus native alpha peaks, alpha plus RCA gamma, guanidine hydrochioride treatment plus dialysis. The inserted bar in panel (D) represents the distance of 50 nm.

Because the gamma crystallins contain an unusu- act directly with the enzyme carbonic anhydrase after ally large number of cysteine and half-cystine resi- its denaturation by heat.17 dues,1 reduced alkylation of these amino acids with It should be realized, however, that the amount of iodoacetamide would be expected to result in forms of gamma crystallin that binds alpha crystallin after gamma that would not completely renature under any guanidine hydrochioride denaturation and dialysis will condition. The observation that alpha binds to the depend on several parameters. These include the same amounts of unalkylated versus RCA gamma after weight ratio of alpha to gamma, the concentration of guanidine hydrochioride treatment supports the con- the denaturant, and the identity of the protein being clusion that alpha is indeed preferentially recognizing denatured. For example, it was previously shown that denatured forms of the gamma crystallin structure. alpha crystallin facilitated the renaturation of gammas This conclusion is consistent with the results of an- crystallin to a conformation similar to that of native other recent study, which showed that alpha can inter- protein.18 Consistent with these earlier findings, we

Downloaded from iovs.arvojournals.org on 09/29/2021 SSi Investigative Ophthalmology & Visual Science, February 1994, Vol. 35, No. 2

B 12 3 4 5 1 2 3 4 5

FIGURE 4. Western blot analysis of major peaks obtained from S-4000 chromatography. Approximately 2% (A) or 1% (B) of the collected peaks were analyzed as described in Mate- rials and Methods, using polyclonal antisera made against o gamma crystallin (A) or alpha crystallin (B). Arrows desig- nate the bands quantitated by densitometry. Lane 1 of (A) represents 2.5 Mg of gamma crystallin standard, and lane 1 of (B) represents 2.5 /ig of alpha crystallin standard. Lanes 2 to 5 of (A) and (B) represent the same samples. Lane 2, native alpha peak from alpha plus gamma, no guanidine treatment; lane 3, native alpha peak from alpha alone, guan- idine treatment plus dialysis; lane 4, HMWA plus native al- pha peak, alpha plus gamma, guanidine treatment plus dialy- llilliili...... sis; lane 5, HMWA plus native alpha peak, alpha plus RCA III gamma, guanidine treatment plus dialysis.

20 D. X =11.98 +3.51 nm lens proteins. Previous studies of other molecular 15 chaperones have demonstrated that their binding to 10 denatured forms of proteins can be reversed in the presence of adenosine triphosphate.67*13 Based on the 5 results of our in vitro binding studies, most, if not all, 0 .Illlllllhll.. of the binding between alpha and the denatured pro- -5 J I I I tein cannot be reversed with the addition of adenosine 5 10 15 20 25 triphosphate (results not shown). If a similar situation Diameter (nm) exists in the intact lens then the relatively large amounts of free alpha found in the water-soluble frac- FIGURE 3. Size distribution of aggregates. The panels corre- tion of young lenses may over the lifetime of the organ- spond with the same samples shown in Figure 2. For each ism be irreversibly complexed with partially denatured panel, 100 aggregates were measured, and their relative size proteins produced during the aging process. This pos- was plotted as a function of percentage abundance. The in- sert represents the mean ± SD. TABLE l, Quantitation of Alpha and Gamma Crystallins in the HMWA and/or have found that after guanidine hydrochloride treat- Native Alpha Peaks from S-4000 Gel ment and dialysis, only about 20% to 30% of gammas binds to alpha, whereas the rest is eluted in the un- Permeation Chromatography complexed form (results not shown). Sample Alpha (fig)* Gamma (fxg)* Previous studies have demonstrated the presence of increased amounts of high-molecular-weight ag- Native alpha + native 19 gamma 1396 ± 46 20.4 ± 3.0 gregate material in the aging and cataractous lens. Guanidine-treated alpha 1446 ± 46 3.4 ± 0.3 Some of this material could be the result of alpha bind- Guanidine-treated alpha ing to partially denatured forms of the other lens crys- + gamma 1456 ± 120 326 ± 24 tallins. Based on the results of this report, we hypothe- Guanidine-treated alpha size that during aging, the large amounts of alpha pres- + RCA gamma 1376 ± 114 328 ± 12 ent in lens fiber cells are necessary for the purpose of * Average of three separate determinations ± SD. Values normal- binding to and preventing further denaturation of ized to starting amounts of alpha and gamma crysiallin.

Downloaded from iovs.arvojournals.org on 09/29/2021 Alpha-Gamma Interaction 387

sibility is consistent with the known, age-dependent 10. Laemmli U. Cleavage of structural proteins during as- decrease of alpha crystallin in the water-soluble frac- sembly of the head of bacteriophage T4. Nature. tion of the lens homogenate.20'21 As a result of this 1970:227:680-686. binding and insolubilization process, the aged lens I 1. Koretz J, Augusteyn R. Electron microscopy of native contains much lower amounts of free alpha, resulting and reconstituted alpha crystallin aggregates. Curr Eye in a diminished ability of the lens to prevent further Res. 1988:7:25-30. J2. VViech H, Buchner J, Zimmerman R, Jakob U. Hsp90 denaturation of its proteins. chaperones protein folding in vitro. Nature. 1992; 358:169-170. Key Words 13. Ostermann J, Horwich A, Neupert W, Haiti F-U. Pro- alpha crystallin, molecular chaperone interaction tein folding in mitochondria requires complex forma- tion with hsp60 and ATP hydrolysis. Nature. 1989:341:125-130. References 14. Boyle D, Gopalakrishnan S, Takemoto L. Localization 1. Bloemcndal H. The vertebrate eye lens. Science of the chaperone binding site. Biochem Biophys Res 1977; 197:127-137. Commun. 1993; 192:1147-1 154. 2. HorwitzJ. Alpha crystallin can function as a molecular 15. Braig K, Simon M, Furuya F, HainfeldJ, Horwich A. chaperone. Proc Nat Acad Sci USA. 1992;89:10449- A polypeptidc bound by the chaperonin groEL is local- 10453. ized with a central cavity. Proc Nat.1. Acad Sci USA. 3. Horwitz J, Emmons T, Takemoto L. The ability to pro- 1993;90:3978-3982. tect against heat-induced aggregation is age-depen- 16. Creiglnon F. Unfolding protein folding. Nature dent. Curr Eye Res. 1992; 1.1:81 7-822. 1991:352:17-18. 4. Ingolia T, Craig E. Four small Drosophila heat shock 1 7. Rao V, Horwitz J, Zigler S. Alpha-crystallin, a molecu- proteins arc related to each other and to mammalian lar chaperone, forms a stable complex with carbonic alpha-crystallin. Proc Nad Acad Sci USA. 1982;79: anhydrase upon heat denaturation. Biochem Biophys 2360-2364. Res Commun. 1993; 190:786-793. 5. de Jong W, Leunissen J, Leenen P, Zwecrs A, Ver- 18. Horwitz J. The function of alpha crystallin. Invest Oph- steeg M. Dogfish alpha-crystallin sequences. J Hiol thalmol Vis Sci. 1993;34:10-22. Chain. 1988;263:514 1 -5149. 19. Jedziniak J, Kinosliita J, Yates E, Hocker L, Benedek 6. Schlesinger M. Heat shock proteins. J Biol Chem. G. On the presence and mechanism of formation of 1990:265:121 J 1-121 14. heavy molecular weight aggregates in aging and cata- 7. Cething M-J, Sambrook J. Protein folding in the cell. ractous human lenses. Exp Eye Res. 1975;20: 367- Nature. 1992;355:33-45. 369. 8. Hanseri J, Takemoto D, Takemoto L. Monoclonal an- 20. Roy D, Spcctor A. Absence of low-molecular-weighi liserum to gamma crystallin: Characterization and use alpha-crystallin in nuclear region of old human lenses. as a probe of changes in the aging human lens. Ijms Proc Natl Acad Sci USA. 1976:73:3484-3487. Res. I984;2:2O7-219. 21. McFall-Ngai M, Ding LL, Takemoto L, Horwitz J. 9. Takemoto L, Azari P. Role of sulfliydryl groups in the Spatial and temporal mapping of the age-related formation of a hereditary cataract in the rat. Exp Eye changes in human lens crystallins. Exp Eye Res. Res. 1975:20:1-13. 1985:41:745-758.

Downloaded from iovs.arvojournals.org on 09/29/2021