IIIIIIIIIIIIIII III I II III , I, 1058 4716 LIBRARIES MICHIGAN STATE UNIVERSITY EAST LANSING, MICH. 48824 This is to certify that the dissertation entitled DISRUPTION OF THE REGULATORY CONFORMATION OF CALMODULIN BY ALUMINUM BINDING: A MOLECULAR BASIS FOR ALUMINUM TOXICITY presented by Neal A. Siege] has been accepted towards fulfillment of the requirements for Doctoral Botany degree in 22W Major rofessor Date November I], 1983 MSU is an Affirmative Action/Equal Opportunity Institution 042771 MSU LIBRARIES “- RETURNING MATERIALS: Place in book drop to remove this checkout from your record. FINES will be charged if book ts returned after the date stamped below. W : I DISRUPTION OF THE REGULATORY CONFORMATION OF CALMODULIN BY ALUMINUM BINDING: A MOLECULAR BASIS FOR ALUMINUM TOXICITY By Neal A. Siegel A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Botany and Plant Pathology 1983 ABSTRACT DISRUPTION OF THE REGULATORY CONFORMATION OF CALMODULIN BY ALUMINUM BINDING: A MOLECULAR BASIS FOR ALUMINUM TOXICITY By Neal Siegel The interaction of aluminum ions with bovine brain calmodulin has been examined by fluorescence spectrosc0py, circular dichroic spectrOphotometry, equilibrium dialysis and by the activation of 3',5'-cyclic nucleotide phos- phodiesterase. These experiments show that aluminum binds stoichiometrically and cooperatively to calmodulin. Alum- inum binding at a molar ratio of 2:1 to calmodulin suffices to induce major structural changes. Estimates from spectro- scopic data indicate that the binding affinity for the first mol of aluminum bound to the protein is one order of magni- tude stronger than that of calcium to its comparable site. These estimates agree with a dissociation constant of 0.4 uM derived from equilibrium dialysis experiments. Interaction of aluminum with calmodulin induces a helix-coil transition and enhances the hydrophobic surface area more than does calcium. A molar ratio of 4:1 for [aluminum]/-[calmodulin] completely blocks the activity of the calcium-calmodulin- Neal A. Siegel dependent phosphodiesterase. Highly hydrated aluminum ions apparently promote solvent—rich, disordered polypeptide re- gions in calmodulin which, in turn, profoundly influence the protein's flexibility. EPR spectra of spin-labelled calmo- dulin provide data indicating that aluminum binding causes decreased probe immobilization as compared to the effects of calcium binding. This result of aluminum binding indicates that aluminum-calmodulin is a more random, Open polypeptide relative to the structure of calcium-calmodulin. Calorimet- ric measurements of aluminum binding provide data showing that the first mol of aluminum bound is accompanied by the largest enthalpic change of the three mol bound; the second and third mol of aluminum bound are each entrapically driven. Micromolar concentrations of aluminum ions interfere with calmodulin-stimulated, membrane-bound ATPase activity which plays a role in the maintenance of the transmembrane potential of plasma membrane enriched vesicles isolated from barley roots. At a molar ratio of 3:1 [aluminum]/[calmodu- lin], the calmodulin-stiumlated enzymatic activity, probably associated with a Ca++ Mg2+-ATPase is 95% inhibited. Aluminum-induced changes in calmodulin structure are reflected in reduced formation of the membrane potential when assayed with a fluorescent potential probe, oxonol VI. These data strongly suggest that the aluminum-calmodulin complex represents a primary lesion in toxic responses of plants to this metal. This dissertation is dedicated to my wife, Debbie for her unfailing love, encouragement and support and to our puppy, Corduroy for his loyalty and unselfish devotion. ii ACKNOWLEDGEMENTS I would like to thank the members of my guidance committee for their care and help; Dr. Estelle McGroarty, Dr. Robert Bandurski, Dr. Alfred Haug in whose laboratory my thesis research was conducted, and Dr. Ken Poff, my academic advisor, to whom I owe special thanks for keeping me on the straight and narrow and for his unending patience. I also acknowledge financial support from the United States Department of Energy Plant Research Laboratory under Contract No. DE-ACOZ-76ERO-1338. iii TABLE OF CONTENTS LIST OF FIGURES O O O O O O O O O O O O O O O O O O 0 LIST OF TABLES O O O O O O O O I O O O O O O O O O Page vii dddddddddd —5 ONNU'IUI «b #wwwWNNNN-A ddddd CHAPTER I. GENERAL INTRODUCTION . . . . . . . . . . . . . II. ALUMINUM CHANGES THE CONFORMATION OF CALMODULIN 1. Results and Discussion . . . . . . . . . ii. References . . . . . . . . . . . . . . . III. ALUMINUM INTERACTION WITH CALMODULIN: EVIDENCE FOR ALTERED STRUCTURE AND FUNCTION FROM OPTICAL AND ENZYMATIC STUDIES Summary i. Introduction . . . . . . . . . . . . ii. Materials and Methods . . . . . . . . . . Calmodulin preparation . . . . . . . . . Materials . . . . . . . . . . . . . . . . Phosphodiestrerase assay . . . . . . . . Circular dichroism experim en nts . . . . . Fluorescence measurements . . . . . . . . Metal ion titration . . . . . . . . . . . Equilibrium dialysis . . . . . . . . . . iii. Results . . . . . . . . . . . . CD studies of aluminum-induced changes in calmodulin , , , , , Aluminum binding increases the hydrophobic surface of calmodulin . . . . . . . Intrinsic fluorescence of calmodulin . . Aluminum inhibits phosphodiesterase . . . iv. Discussion . . . . . . . . . . . . . . . v. References . . . . . . . . . . . . . . . IV. CALMODULIN-DEPENDENT FORMATION OF MEMBRANE POTENTIAL IN BARLEY ROOT PLASMA MEMBRANE VESICLES: A BIO- CHEMICAL MODEL OF ALUMINUM TOXICITY IN PLANTS Summary 1. Introduction . . . . . . . . . . . . . . ii. Materials and Methods . . . . . . . . . . Calmodulin preparation . . . . . . . . . Chemicals . . . . . . . . . . . Plant material and growing conditions . . iv Preparation media . . . . . . . . . Membrane isolation . . . . . . Determination of membrane potential ANS fluorescence measurements . . Protein analysis . . . . . . . . Removal of contaminating metals . Data analysis . . . . . . . . . . iii. Results . . . . . . . . . . . Oxonol VI as a probe of transmembran potential . . pH dependence of the .Ca2+- and Mg2+- dependent ATPase activity and i relation to the development of transmembrane potential . . . Effect of divalent cation concentrat potential development . . . . Effect of ionOphores, calmodulin, calmo- dulin antagonists and aluminum the deve10pment of the membrane potential . . . N b e . . 24 t8 the O O O O 25 ion on . . . 25 on O O O 25 Aluminum inhibition of the calmodulin- stimulated ATPase- -dependent formation of the transmembrane potential iv. D18CU881on O O O O O O O I O O O 0 v. References . . . . . . . . . . . . V. A THERMODYNAMIC AND ELECTRON PARAMAGNETIC STUDY OF STRUCTURAL CHANGES IN CALMODULIN ALUMINUM BINDING Summary 1. Introduction . . . . . . . . ii. Sources . . . . . . . . iii. Methods . . . . iv. Results and Discussion v. References . . . . . . VI 0 SUWARY O O O O O O O O O O O O O O O O 0 GENERAL REFERENCES 0 I O O O O O O O O O O O O O O C C 26 O O O O 27 O O O O 27 RESONANCE INDUCED BY 0 O O O 30 O O O O 31 . . . . 31 . . . . 31 . . . . 35 O O O O 37 . . . . 44 APPENDIX. STRUCTURAL DISTINCTIONS BETWEEN BARLEY AND BRAIN CALMODULIN: A STUDY OF STRUCTURAL CHANGES ACCOMPANYING CALCIUM AND ALUMINUM INTERACTION WITH BARLEY CALMODULIN Summary 1. Introduction . . . . . . . ii. Sources . . . . . . . . . iii. Methods . . . . . . . . iv. Results and Discussion . . v. References . . . . . . . . 49 50 50 52 62 LIST OF FIGURES Figure Page 2.1 ANS fluorescence of bovine brain calmodulin as a function of [metal]/[protein] (mol:mol) ratio 8 3.1 Changes in the CD spectrum of bovine brain calmodulin in the presence of (A) calcium and (B) aluminum . . . . . . . . . . . . . . . 14 3.2 Effects of metal titration on the mean residue ellipticity [6], at 222 nm and the helical content of bovine brain calmodulin . . . . . . 15 3.3 Fluorescence of 8-anilino-1-naphthalene sulfonic acid in the presence of bovine brain calmodulin and increasing metal content . . . 16 3.4 Tyrosine fluorescence of bovine brain calmodulin as a function of metal concentration . . . . . 16 3.5 Binding data from equilibrium dialysis experi- ments presented as a Scatchard plot . . . . . 17 3.6 Inhibition of calcium-calmodulin-stimulated 3'-5'-cyc1ic nucleotide phosphodiesterase activity by aluminum . . . . . . . . . . . . . 17 4.1 Changes in the fluorescence intensity of oxynol VI in the presence of membrane vesicles from barley roots . . . . . . . . . . . . . . . . . 24 4.2 pH dependence of the rate constant, k, for the deveIOpment of the transmembrane potential . . 25 4.3 Effects of ionOphores on the development of the transmembrane potential . . . . . . . . . . . 26 4.4 Effect of aluminum on calmodulin-stimulated potential deveIOpment or on ANS partitioning onto isolated calmodulin . . . . . . . . . . . 26 4.5 Changes in spin—probe mobility of covalently labelled calmodulin due to metal addition . . 32 4.6 Binding of aluminum to calmodulin . . . . . . . 32 vi Figure Page A1. Electrophoresis of calmodulin on 15% acrylimide gels in the presence of SDS . . . . . . . . . 53 A2. Electrophoresis of calmodulin on 15% acrylimide gels under non-denaturing conditions . . . . . 54 A3. Changes in the helical content of 10 UM barley calmodulin titrated with CaClz, AlCl3 or AlCl3 in the presence of 70 uM CaC12 . . . . . 56 A4. Changes in ANS fluorescence of 10 uM barley calmodulin titrated with CaClz, AlCl or AlCl3 or A1C13 in the presence of 70 uM Caglz . . . 58 vii LIST OF TABLES Table Page 2.1 a-helical content of bovine brain calmodulin in the presence of Al3+ and Ca2+ . . . . . . . 8 3.1 Structural changes in bovine brain calmodulin induced by metal binding . . . . . . . . . . . 15 4.1 Effects of varying divalent cation concentra- tions on the rate of potential deveIOpment . . 25 4.2 Effects of various compounds on the development of the membrane potential . . . . . . . . . . 26 5.1 Thermodynamic parameters associated with the binding of aluminum to bovine brain calmo- dulin . . . . . . . . . . . . . . . . . . . . 33 Al Changes in [6]22 and helical content of calmodulin derived from bovine brain and barley induced by metal binding . . . . . . . 55 A2 Amino acid content of calmodulin . . . . . . . . 6O viii Chapter 1 General Introduction 2 Aluminum is the most versatile metal in use by in- dustrial mankind today. It is the most abundant metal in our world, comprising eight percent of the Earth's crust. Aluminum is also one of the most toxic of the elements; toxic levels range between one and four parts per million for aluminum in solution at low pH. Forty percent of the world's arable soils are unsuitable for plant growth due to acid levels that maintain toxic amounts of aluminum in solution. The potent toxicity of aluminum is clear whether the measure is taken from crop failure or from lakes which can no longer support plant or animal life due to acidification as a result of acid precipitation caused by the burning of fossil fuels by the industrialized world. As indicated, the toxic effects of aluminum are widespread with the symptoms of poisoning and death appearing below five parts per million of the metal in eukaryotic organisms. In order to understand why aluminum is toxic it is necessary to know something about its physical chemistry. The most unique property of aluminum is that it is bound with water more strongly than any other element of a similar ionic radius. The hydration enthalpy of these bonds is 1144 Kcal/mole (Matheja and Degens, 1971). Six water molecules arranged in a regular octahedron comprise the inner hydra- tion shell; because of its tight association with water, aluminum cannot be considered as a free ion in aqueous solution. In fact, two physical constraints dictate the basic form that the hydrated aluminum ions will have in solution: pH and concentration (Baes and Mesmer, 1976). At low concentration and pH (less than 4 ppm and pH 3), only monomeric, hydrated aluminum ions of +3 charge exist. If this concentration remains below 60-100 “M (approx. 2-3 ppm) but the pH increases above 3, protons from the inner hydra- tion shell of water molecules will dissociate, and the net charge of the hydrated complex will decrease to neutrality and even become negative. In addition to charge alteration, polymorphic aluminum hydroxides will form and become insol- uble. Likewise, if the pH remains low, increased aluminum concentration will promote the formation of polymorphic, insoluble aluminum species. At physiological pH ranging between 6 and 7.5, therefore, only aluminum concentrations between 0 and 4 ppm (0-100 HM) will support monomeric, hydrated aluminum species with net charges ranging between +3 to 0. Thus, the biological toxicity of aluminum is clearly a result of constraints imposed by its solution chemistry. Although the effects of aluminum toxicity in plants are manifold, ranging from impairment of root elongation (Foy et al, 1978) to chloroplast membrane degeneration (Hampp and Schnabl, 1975) and plasma membrane degeneration in roots (Hecht-Buccholz and Foy, 1981), the most obvious effect on crOp plants is impaired calcium uptake, distribution and use (Foy et a1, 1978). In recent years calcium involvement as a second messenger by the eukaryotic cells has been shown to be controlled by a ubiquitous, calcium-dependent regulating 4 protein, calmodulin. Processes ranging from plasma-membrane ATPases (Dieter and Marme, 1981, Caldwell and Haug, 1981) and NAD kinase (Muto and Miyachi, 1977) in plants to calcium dependent phosphodiesterase, ATPases and adenylate cyclase in animals (recent reviews: Cormier et al, 1980, Klee, 1980, Klee et a1, 1980, Lin, 1982, Wang and Waisman, 1979) are controlled by calmodulin. Aluminum, known to be a potent calcium antagonist, might exert this antagonism through specific interaction with calmodulin. Because aluminum is antagonistic toward calcium, a detailed investigation into the physical effects of aluminum on calmodulin and attenuation of the regulation of enzymatic processes under its control was initiated. This allowed a comparison to be made between physical alterations in the structure of calmodulin due to aluminum binding and the assesment of these changes on the regulatory functioning of the protein. In this way, calmodulin could be shown to be a key lesion occuring in the broadly defined syndrome of aluminum toxicity. Calmodulin isolated from bovine brain was chosen for study because it represents the best characterized calmo- dulin molecule from all known sources. Its amino acid sequence, order of calcium fill and structural character- istics are well documented. Because of this documentation, even subtle structural changes would be evident as compared with changes occuring in calmodulin from a less studied source. 5 In order to assess the biological impact of aluminum- induced structural changes in calmodulin, two systems known a to be regulated by the protein were assayed for aluminum inhibition; calcium-dependent phosphodiesterase and the Ca2++ Mg2+-ATPase from barley (Hordeum vulgare var. Conquest) root plasma membranes. The physical techniques used to assess relative structural changes of the protein through interaction with aluminum were independent of the biological activity; these techniques included circular dichroism spectroPhotometry, ANS fluorescence spectroscopy, intrinsic tyrosine fluorescence spectroscopy and EPR spectroscopy of spin-labelled calmodulin. Chapter 2 Aluminum Changes the Conformation of Calmodulin Physiol. Chem. & Physics l4 (I982) I65 ALUMINUM CHANGES THE CONFORMATION 0F CALMODULIN NEAL SIEGEL. CHARLES SUHAYDA‘. and ALFRED HAUG MSU-DOE Plant Research Laboratory, and‘ Department of Botany and Plant Pathology, Michigan State University, East Lansing. MI 48824 0 Fluorescence titration experiments indicate that Al" binds stoichiometricall y to electrodialyzed bovine brain calmodulin. There are three binding sites for aluminum on the protein. Application of Al 3' to calmodulin appears to expose larger hydrophobic surface domains as compared with those found for calmodulin in the presence of calcium or gallium. Calmodulin is a ubiquitous and multifunc- tional calcium-regulating protein which has been shown to participate in a variety of calcium-dependent processes including stimulation of ATPases and phosophodi- esterases‘. Calmodulin has a molecular weight of about 17,000 and has been conserved during evolutionz. Dependent upon the ionic strength, the dissociation constants of the four calcium binding sites lie in the micromolar ranges. In the case of bovine brain calmodulin, the high affinity sites 1 and II are devoid of tyrosyl residues, while the low affinity sites 111 and IV are associated with one tyrosine each, viz.. tyr 99 at site 111, and tyr l38 at site IV”. These are the sole tyrosine residues in the entire calmodulin molecule, tryptophan residues are lacking. Therefore measurements of tyrosine fluorescence, or of energy transfer from tyrosine to luminescent lanthanides, can be performed to investigate metal- induced conformational changes of calmo- dulin“. These changes generate domains with considerable hydrophobicity as evi- denced by experiments employing hydro- phobic fluorescence probes like 8-anilino-l- naphthalene sulphonate, ANS“. Aluminum accumulation in animals and man has been implicated in diseases like Alzheimer's disease and dialysis dementia". In plants. micromolar concentrations of aluminum in the soil decreased the rate of root elongation and induced symptoms typi- cal of calcium deficiencies“. Considering the importance of calmodulin in calcium regula- tion, the potent interaction of aluminum with calmodulin shown to occur in this study may represent a crucial biochemical lesion of aluminum toxicity. Calmodulin was prepared from bovine brain acetone powder and purified by pheno- thiazine affinity chromatography". The eluted calmodulin was dialyzed against distilled water, electrodialyzed and then lyophilized. The protein activated 3':5’-cyclic nucleotide phosphodiesterase and migrated as a single band during sodium dodecyl sulphate gel electrophoresis. Fluorescence intensity measurements were carried out on a Perkin-Elmer spectrofluorimeter, model MPF-44A, equipped with a differential corrected spectra unit. The ANS fluorescence intensity of calmodulin, in the absence of metal, was considered as the initial fluores- cence intensity value. Data for ANS fluores- cence, in the presence of the metal, are expressed as relative increase in the initial fluorescence intensity value. Each value represents the mean of at least three separate calmodulin preparations within 5% standard error. Application of aluminum to calmodulin scents to expose larger hydrophobic sur- face domains as compared with those found in the presence of calcium or gallium. At a metal concentration of about 25 uM. calcium enhanced the ANS fluorescence intensity by about 2%. whereas gallium and aluminum increased the intensity by about I66 ”0‘00 d MZMOMMIOCF'H 0 1 2 3 4 5 [METAll/[CaM] FIGURE I. ANS fluorescence of bovine brain cal- modulin as a function of [metal],’[protein] (molzmol) ratio. A 7 pM concentration of calmodulin was prepared in l0 mM morpholino propane sulphonic acid (MOPS) buffer, pH 6.5. The concentration of 8-anilino-l- naphthalene sulphonate, ANS. was 2 uM. Excitation wavelength was 360 nm. ANS fluorescence intensity was recorded at 490 nm. The ordinate lists the metal-induced relative fluorescence intensity. The protein was titrated with the metal ions. viz. Al” (. ). Gaa' (A ), and Ca2' (O l. N. SlliGl-LL, ('. sruavm and A. ”MTG 20 and 400%, respectively (Fig. l). The ANS fluorescence titration curve reached a maximum at a ratio of 3 mol of All“ per mole of calmodulin. Our findings therefore indi- cate that aluminum binds to calmodulin in a stoichiometric manner. Hill plots suggest that the affinity of aluminum to calmodulin is at least one order of magnitude larger than that of calcium to calmodulin. It seems worth noting that aluminum and the closely related gallium have similar binding constants. As calculated by the procedure of Chen et al.‘°, results from our circular dichroism studies indicate that All” application to calmodulin decreases the cit-helical content of the protein in contrast to the observed increase of a-helix content upon binding of calcium in the protein’(Table 1). Upon binding calcium, specific changes in the internal protein struc- ture allow calmodulin to participate in the second messenger systema. Loss of this structure occurs in the presence of aluminum as shown by our studies. We propose that the loss of this structure probably impairs the functioning of calmodulin as a calcium regulator. TABLE I. a-Helical Content of Bovine Brain Calmodulin in the Presence of Al" and Car [Metall’ [Calmodulin] Metal 9i a-Helixg- (mol ; mol) 0 — 37 2 Cd” 40 4 Ca” 49 2 Ar" 28 5 Al" 22 lcalculated according to the method of Chen et al.'°. These values represent the mean of at least three calmodulin preparations within 592 SE. This work was supported by US Department of Energy contract No. DE-ACOl-76EROI338. N. SIEGEL. c. sumvm and A. IIAIIG REFERENCES l. C. B. Klee. T. H. Crouch, and P. G. Rich- man. Calmodulin. Ann. Rev. Biochem. 49, 489 (I980). 2. D. M. Watterson, F. Sharief, and T. C. Vanaman. The complete amino acid sequence of the cab-dependent modulator protein (calmodulin) of bovine brain. J. Biol. Chem. 255. 962 (I980). 3. J. H. Wang and D. M. Waisman. Calmodu- lin and its role in the second messenger system. Current Topics in Cellular Regulab tion. Vol. IS, B. L. Horecker and E. R. Stadtman, Eds. Academic Press, I979, pp. 47-i07. 4. M. C. Kilhoffer, J. G. Demaille, and D. Gerard. Terbium as luminescent probe of calmodulin binding sites. FEBS Letters. 116. 269 (I980). . I67 5. R. W. Wallace. E. A. Tallance. M. E. Dockter, and W. Y. Chcung. Calcium binding domains of calmodulin. J. Biol. Chem, 257. 1845 (I982). 6. D. C. LaPorte, B. M. Wierman, and D. R. Storm. Calcium-induced exposure of a hydrophobic surface on calmodulin. Bio- chem. l9. 38l4 (I980). 7. D. R. Crapper McLachlan. Aluminum in human brain disease: An Overview. Neuro- tox., l. 3 (I980). 8. C. D. Foy, R. L. Chaney, and M. C. White. The physiology of metal toxicity in plants. Ann. Rev. Plant Physiol” 29. 511 (I978). 9. C. R. Caldwell and A. Haug, Affinity chromatographic isolation of calmodulin from bovine brain acetone powder. Analyt. Biochem.. 116. 325 (l98l). I0. Y. H. Chen, J. T. Yang. and H. M. Martinez. Determination of secondary structures of proteins by circular dichroism and optical rotatory dispersion. Biochem.. II. 4l20 (I972). (Received August 20. l982: revised November 26. l982) Chapter 3 Aluminum Interaction with Calmodulin: Evidence for Altered Structure and Function from Optical and Enzymatic Studies 3b Biochmm'u et Bmphystca Arm. 744 ( 1983) 36-45 Elsevter Biomedical Press BBA 31556 ALUMINUM INTERACTION WITH CALMODULIN EVIDENCE FOR ALTERED STRUCTURE AND FUNCTION FROM OPTICAL AND ENZYMATIC STUDIES N. SIEGEL and A. HAUG ' MSU-DOE Plant Research Laboratory, Michigan State University. East Lansing. MI 48 824 (U. S.A. ) (Received October I2th. I982) Key words: C almodulm; Conformational change: A l 3 I; Metal binding; (Bovine brain) The interaction of aluminum ions with bovine brain calmodulin has been examined by fluorescence spectroscopy, circular dichroic spectrophotometry and equilibrium dialysis. and by the calmodulin-dependent activation of 3',5'-cyclic nucleotide phosphodiesterase. These experiments show that aluminum binds stoichiometrically and cooperatively to calmodulin. Binding of aluminum at a molar ratio of 2: I to calmodulin suffices to induce a major structural change. Estimates from spectroscopic data indicate that the binding affinity for the first mol of aluminum bound to the protein is about one order of magnitude stronger than that of calcium to its comparable site. These estimates agree with a dissociation constant of 0.4 pM derived from equilibrium dialysis experiments. Interaction of aluminum with calmodulin induces a helix-coil transition and enhances the hydrophobic surface area much more than calcium does. A molar ratio of 4: I for IaluminumI/lcalmodulinl is sufficient to block completely the activity of the calcitun-calmodulinodependent phosphodiesterase. Highly hydrated aluminum ions apparently promote solvent-rich. disordered polypeptide regions in calmodulin which. in turn, profoundly influence the protein’s flexibility. Introduction [2—4]. and the list of calmodulin‘s roles is being . constantly expanded (recent reviews. see Refs. Calmodulin is a ubiquitous. evolutionary highly 5-8). Although the detailed mechanisms of conserved, multifunctional. calcium-regulating protein which was originally discovered as an activator of cyclic nucleotide phosphodiesterase [I]. Some of the processes in which calmodulin is known to participate involve stimulation of vari- ous ATPases, adenylate cyclase, plant NAD kinase ‘ Present address: Pesticide Research Center. Michigan State University. East Lansing, MI 48824. U.S.A. Abbreviations: EGTA. ethylene glycol bist fl-aminoethyl ether)- N.N.N'.N’-tetraacetic acid: ANS. 8-anilino-l-naphthalene sul- fonic acid: Mops. 4-morpholinepropanesulfonic acid: Mes. 4. morpholineethanesulfontc acid: PMSF. phenylmethylsulfonyl fluoride; cGM P. guanosine 3'.5'-cyclic phosphoric acid. 0l67-4838/83/0000-0000/50300 L I983 Elsevier Science Publishers calmodulin action remain largely unknown. this molecule is thought to modulate cellular processes in which calcium is the second messenger [6.9]. Four specific calcium-binding sites exist on calmodulin and a detailed investigation of the positive cooperativity exhibited by calcium bind- ing to calmodulin has recently been published [ID]. The ability of other metals to compete with calcium for binding to calmodulin has been demonstrated for magnesium. manganese and the sequential bin- ding of terbium [I I-l3]. Conformational changes induced by calcium binding enhanced the hydro- phobic surface exposure of calmodulin [l4]. Calmodulin is found in eukaryotic animal and 12 plant tissues. and in high concentrations in brain tissue. especially in the frontal cortex [4.15]. Anti- psychotic drugs and certain neuropeptides have been found to modify calmodulin-mediated neuro- nal functions [16]. Moreover. changes in calcium levels have been correlated with aluminum accu- mulation in Alzheimer's disease. which is a neuro- fibrillary degenerative process in the human central nervous system [l7]. In plants. micromolar con- centrations of aluminum in the soil and subse- quent uptake diminished the rate of root elonga- tion and induced symptoms typical of calcium deficiency [18]. In this report we present results of a study on the interaction of aluminum with bovine brain calmodulin. The aim of the investigation was to detect aluminum-induced changes in calmodulin conformation. We found that aluminum binds stoichiometrically to calmodulin and enhances the hydrophobic surface domain relative to that ex- posed in the presence of calcium. These changes are reflected in the inhibition of the calmodulin- dependent activity of 3’.5’-cyclic nucleotide phos- phodiesterase. Aluminum-induced conformational changes in calmodulin are examined by fluores- cence spectroscopy and circular dichroic spectro- photometry. Materials and Methods Calmodulin preparation. Calmodulin was pre- pared from bovine brain acetone powder as de- scribed previously [l9]. However. to increase the purity of the isolated product. the protein-loaded affinity column was washed with a buffer contain- ing 20 mM Mes/NaOI-I. pH 7. 500 mM NaCl and I mM mercaptoethanol. Subsequently. the cal- modulin was eluted with the same buffer contain- ing l0 mM EGTA. CaCl2 was immediately added to the calmodulin fraction collected to give a final concentration of IS mM. Following dialysis against IO mM ammonium carbonate and distilled water [19]. the calmodulin fraction was further subjected to electrodialysis. applying a current between 5 and 10 mA for about 4 h. The final solution was lyophilized and the product was stored in a desic- cator at -40°C. This product was reconstituted in deionized water and dialyzed first against lOO [1M EDTA. then exhaustively against deionized water. 37 and finally against deionized buffer containing Chelex-IOO resin. This material stimulated 3'.5'- cyclic nucleotide phosphodiesterase. had an ultra- violet absorption spectrum typical to that of calmodulin [l9]. and migrated as a single band on SDS-polyacrylamide gels (15%) as visualized by the silver staining technique [20]. Protein con- centrations were determined by a modification [21] of the method of Lowry et al. [42]. or by measur- ing an absorption spectrum with a Gilford spec- trOphotometer. model 2400: the molar extinction coefficient is 3300 M" -cm". at 277 nm [13]. Calmodulin isolated as described above contained less than 10’8 M calcium. magnesium. manganese and aluminun. This quantitative analysis was per- formed with a Jarrell-Ash plasma emission spec- trometer. model 955 Atomcomp. and a Varian atomic absorption spectrophotometer. model 1475. Materials. Bovine brain acetone powder. Tris, Mes. PSMF. EDTA. EGTA. Mops, 5'-nucleoti- dase and 3’.5'-cyclic nucleotide phosphodiesterase were purchased from Sigma Chemical Co. (St. Louis. MO). [8-3H]Guanosine 3'.5’—cyclic phos- phate was obtained from New England Nuclear Corp. (Boston. MA). AIC13. CaCl2 and NaCl were obtained from Mallinckrodt Science Products (St. Louis. MO). Chelex-IOO. AG lX-8. and Affigel phenothiazine were purchased from Bio-Rad Laboratories (Richmond. CA). The sodium salt of ANS was obtained from K & K Laboratories (Plainview. NY). (+ )-lO-Camphorsulfonic acid was obtained from Aldrich Chemical Co. (Milwaukee. WI). All other chemicals were of the highest quality available. All glassware and quartz cuvettes were washed in concentrated nitric acid. Buffer solutions were prepared in double. glass-distilled deionized water and passed over columns (2 X 30 cm) of Chelex- lOO. Metal stock solutions of calcium and aluminum chloride were freshly prepared in de- contaminated buffers. Plasma emission and atomic absorption spectroscopy indicated buffer solutions typically to contain less than 10‘“ M calcium. magnesium. manganese and aluminum. Phosphodiesterase assay. Hydrolysis of cyclic GMP by 3'.5’-cyclic nucleotide phosphodiesterase was determined by the procedure described by Wolff et al. [I I]. with the following modifications. 300441 reaction volumes containing 25 ,uM 13 38 carrier-labelled cGMP and 5.5 uM calmodulin in 10 mM Tris-HCI. pH 6.5. were constructed in polyethylene containers. An aliquot of activator-deficient 3'.5’-cyclic nucleotide phos- phodiesterase was added to start the reaction: reaction times were allowed to vary between 15 and 30 min. at 37°C. The reaction was stopped by boiling the sample for 2 min. After the solution was cooled to ambient temperature. 5’-nucleoti- dase was added and the sample was incubated for 20 min at 37°C. This reaction was stopped by adding a l-ml aliquot of AG lX-8 resin. diluted 1 :2 (v/v) in 50% isopropanol. The reaction vials were centrifuged and the supernatants analyzed in a Beckman scintillation counter. model LS 7000. to determine the amount of hydrolysis. Circular dichroism experiments. Some CD spec- tra were recorded at ambient temperature on a Jasco spectropolarimeter. model ORD/ UV / CD-S. modified by Sproul Scientific Instruments (Boulder Creek. CA). For purposes of digital signal process- ing. CD spectra were also recorded on a Jasco automatic recording spectropolarimeter. model J-40 C. interfaced with a Data General Nova-3 processor and a Tracor-Northern digital signal analyzer. model TN-ISOO. Both instruments were calibrated to a molar ellipticity of [0]: +7260 deg-cmz-dmol". at 290.5 nm. for a 0.1% aque- ous solution of ( +)-IO-camphorsulfonic acid [22]. Samples of calmodulin were prepared from the lyophilized protein reconstituted as described in Materials and Methods. in deionized 10 mM Tris- HCl. adjusted to pH 6.5 with concentrated HCl. CD spectra were obtained from samples in rectan- gular quartz cells of l-cm path length. Calmodulin solutions of 3-ml volume were titrated with 3—10—p1 aliquots of calcium or aluminum chloride salts from stock solutions. Mean residue ellipticities of calmodulin. [0]. were calculated from the relation [0] -= OobsM/ 100 lc, where 00b3 is the observed elliptical value. M is the mean residue molecular weight, taken to be 117 [l 1]. I is the optical path length in decimeters. and c is the protein concentration in g/ml. The relative structural contents of calmodulin were calculated by the procedure of Chen et al. [23]. The helical content was estimated from the rela- tionship % a-helix -= —([0]222 + 2 340)/ 303. where 101222 is the mean residue ellipticity at 222 nm. Relative contents of a-helix were also estimated according to the procedure of Greenfield and Fas- man [24] for the purpose of comparing the mea- sured values with those reported in the literature. The Hill coefficient. h. was determined from the expression Rx = 81”". where the cooperativity in- dex. R‘. has been derived from experimental data [25]. Fluorescence measurements. Fluorescence inten- sity measurements were performed on a Perkin- Elmer spectrofluorimeter. model MPF-44A. equipped with a differential corrected spectra unit. The excitation and emission wavelengths for tyro- sine fluorescence were 280 i 2 and 320 j; 5 nm; for ANS fluorescence studies 360 i 4 and 490 i 4 nm. respectively. For measurements of tyrosine fluorescence. calmodulin was dissolved in 100 mM NaCl. and maintained at pH 6.5 with NaOH. For ANS fluorescence measurements. the sample was prepared in 10 mM Mops buffer. at a 10 pM concentration. adjusted to pH 6.5 with NaOH. ANS concentration of 2 pM. A cuvette of 1 cm optical pathlength was used. The spectra were recorded at room temperature. The absorbance at the excitation and emission wavelengths was less than 0.05. Metal ion titration. Maximum fluorescence is defined as fluorescence intensity of tyrosine emis- sion of calmodulin. in the absence of EGTA and metal ions. Upon addition of 100 pM EGTA. tyrosine fluorescence was quenched. and upon subsequent addition of l—Z—pl aliquots of metal stock solution to the 2 ml sample in the cuvette. a fluorescence increase of tyrosine emission was ob- served. Data for tyrosine fluorescence. in the pres- ence of metal ions. are expressed as percent of maximum fluorescence intensity. The ANS fluo- rescence intensity of a calmodulin solution. in the absence of metal ions. was considered as the initial fluorescence intensity value. This initial value in- creased upon addition of l—Z-pl aliquots of metal stock solution. Data for ANS fluorescence. in the presence of metal ions. are expressed as the rela- tive increase of the initial fluorescence intensity value. A value of zero is defined as the fluores- cence value with protein absent. Equilibrium dialysis. For equilibrium dialysis calmodulin samples were reconstituted in 10 mM Tris-HCI. pH 6.5. Sample volumes of 2 ml were 14 put in dialysis membranes having a molecular weight cutoff at 3400 and dialyzed against 100-ml volumes of various aluminum concentrations in the same buffer for 24 h at room temparature. Aluminum concentrations were determined ac~ cording to the following procedure. 5-iil aliquots of concentrated nitric acid were added to 1 ml sample volumes. After 10 min. 1 ml of l M acetate buffer. pH 6.0. was added. followed by 0.25 ml of 0.01 mg/ml Eriochrome cyanine R in water. After 20 min the absorbance was measured at 535 nm. Standards were prepared by dilution of a 1000 pg/ml aluminum atomic absorption standard solution (Aldrich Chemical Co.. Milwaukee. WI). The value of the dye absorbance was shown to be accurate between 0.1 and 1 ppm aluminum. as compared with the value obtained from direct measurements of aluminum in a plasma emission spectrometer. Interference caused by the presence of protein was accounted for. Results CD studies of aluminum-induced changes in cal- modulin The ultraviolet circular dichroism spectra of calmodulin show increasing negative ellipticities upon titration of calcium to the protein solution. Minima of negative ellipticity appear at 222 and 207 nm (Fig. IA). consistent with CD spectra reported for calmodulin [l 1]. From these spectral features relative amounts of secondary structures can be estimated [23]. We find that a protein structure containing 37% a-helix, 11% B-sheet and 52% random coil generates a CD spectrum which fits best to that observed for metal-free calmodu- lin. These values agree with those reported previ- ously; the helical content has been found to vary between 28 and 45% for metal-free calmodulin and to increase by 10-15% upon calcium binding to calmodulin. at a molar ratio of 4: 1 [3.6.26]. Applying the Greenfield and Fasman procedure [24]. a helix content of 31% can be calculated from our data for the native protein. while Wolff et al. [1 1] determined a value of 28%. Analysis of the CD spectra of calmodulin titrated with aluminum (Fig. 13) indicates a spec- tral shift towards decreasing values of negative ellipticity. contrary to the shift observed upon MEAN RESIOUE ELLIPTICITY (deg cm? / declrnole x10") 39 TICITY 3) (deg cm! I declmole x 10‘ MEAN RESIDUE ELLIP 6 I '16 1 1 l 2.65 210 220 230 240 250 X (nml -2). -4s- ‘16 "i 1 1 l 1 2300 210 220 230 240 250 Mum) Fig. 1. Changes in the CD spectrum of bovine brain calmodulin in the presence of (A) calcium and (B) aluminum. Calmodulin was reconstituted in metal-free 10 mM Tris-HCl buffer. pH 6.5. at a final concentration of 10 MW. A. CaCl, was titrated to a final concentration of (a) 0. (b) 20 and (c) 40 pM. B. MO, was titrated to a final concentration of (a) 0. (b) 20 and (c) 50 pH. The spectra were invariant with respect to time. Error bars indicate the actual noise levels observed during multiple scans. 40 application of increasing calcium concentrations. The aluminium-induced spectral shift even takes place in the presence of saturating levels of calcium (data not shown). Experimentally determined val- ues of [0]222 and calculated values of the helix content are plotted vs. the metal concentration (Fig. 2). Relative to values derived for metal-free calmodulin. the helical content increased by 17% upon calcium binding. whereas it decreased by 30% upon binding of aluminum to calmodulin at a molar ratio of 4: 1. The aluminum-induced change even occurred in the presence of 100 pM CaCl2 (data not shown). From respective metal titration curves. midpoint ratios (mol/mol) of 2: l for [calciuml/[calmodulin] and 2.5 : l for [alumi- numl/[calmodulin] can be derived. Additional in- formation is listed in Table 1. CD spectra for calcium titration at pH 7.5 were identical to those measured at pH 6.5 (data not shown). From plots of [0]222 vs. metal concentration. Hill coefficients of 1.54 and 1.63 can be calculated for calcium and aluminum binding to calmodulin —— these values -9 I I r I r j I I '0 :r t “115 9 -.o_ I ' 420 E a.» 3 125 - ’T o\ \ 42» a I "E 730 E 9 ~I3- r:. o X g = «35 .— '14" \ , j 1 ‘40 9'15: . .. . l ' 1415 net 1 L l I l l l 2 o l 2 3 4 s 6 7 3° [METAL] / [CALMODULIN] Fig. 2. Effects of metal titration on the mean residue ellipticity. [0]. at 222 nm. and the helical content of bovine brain calmodulin. Conditions were the same as those described in the' legend to Fig. l. Titration was performed with CaCl, (0) or AlCl, (0). Mean values for at least four separate calmodulin preparations are plotted; error 5%. Helical content is estimated from the relationship: % helix - —([0]222 +2340)/303. Inset: Computed helical content according to the procedure of Chen et a1. [23]. taking into account the presence of fl-sheets and random coils. TABLE I STRUCTURAL CHANGES IN BOVINE BRAIN CAL- MODULIN INDUCED BY METAL BINDING Results were calculated by the method of Chen et al. [23]. [metal]/ Metal % % % [calmodulin] helix sheet coil (mol/mol) o - 37 l l 52 2 Can2 t 40 9 so 4 Ca2 * 44 9 47 2 Al" 28 7 65 5 Al 3 ° 22 s 73 are indicative of positive cooperativity for the changes observed. Aluminum binding increases the hydrophobic surface of calmodulin Titration of calmodulin with aluminum or calcium induces the exposure of a hydrophobic surface on the protein. as evidenced by partition studies which employed the fluorescent hydro- phobic probe. ANS. This molecule has been used to investigate the expression of hydrophobic surface domains on calmodulin by calcium bind- ing [14]. Aluminum enhanced the ANS fluores- cence intensity more effectively than did calcium in the presence of calmodulin (Fig. 3). The fluores- cence intensity reached saturation at an [aluminum]/[calmodulin] ratio of 3: I (mol/mol). The aluminum-induced fluorescence changes is five times that produced by calcium and remained at this value even in the presence of 120 pM CaCl2 (data not shown). This aluminum-induced change can be characterized by a Hill coefficient of 1.86. indicative of positive cooperativity. Intrinsic fluorescence of calmodulin Bovine brain calmodulin harbors two tyrosyl residues. viz.. Tyr 99 and Tyr 138. which are located at the calcium-binding sites III and IV. respectively [12]. Tryptophan residues are lacking Therefore. measurements of tyrosine fluorescence can be performed to investigate induced confor- mational changes of the protein. The aluminum-induced increase in tyrosine flu- orescence intensity of calmodulin saturated at a 16 .. 5 _—T"’".'“T ' __,_1_ fl (0 P _ g t c >- 4 4 K 1 C l E : 33* ‘ l; 2% - h g . A ’. ——-<> 8 ' fi V V A ll. 1 l 1 1 O 1 2 3 4 El [METAL] / [c ALMODULIN] Fig. 3. Fluorescence of 8-anilino—l-naphthalene sulfonic acid in the presence of bovine brain calmodulin and increasing metal content. CaC 12 (O) or AlCl 3 (O) was titrated into a solution of 10 pM calmodulin in 10 mM MOPS buffer. pH 6.5. as de- scribed in Materials and Methods. The fluorescent probe was present at 2 pM. The data shown are from four separate calmodulin preparations and represent mean values within 5% error. The zero fluorescence value is defined as that of ANS in the absence of calmodulin molar ratio of 2: 1. The first phase of fluorescence enhancement occurred at a ratio of 1 : 1 [aluminum]/[Calmodulin]; this initial phase was characterized by a large and steep increase (Fig. 4). The steepness of the increase is reflected in a Hill coefficient of 1.2. The aluminum concentration at which a 50% increase in tyrosine fluorescence oc- curred is al least one order of magnitude lower than the respective calcium concentration (Fig. 4). The initial phase of aluminum-induced fluores- cence enhancement probably results from binding of the metal near or at Tyr 138. since this residue has been found to be most sensitive to metal-in- duced conformational changes [27]. This interpre- tation is also consistent with results from NMR experiments demonstrating that binding of one calcium ion per calmodulin shifted the resonances of Tyr 138 [28.29]. The aluminum-induced changes in tyrosine fluorescence may also be a consequence of long-range conformational changes. The steep- ness of the aluminum-induced response indicates that at least the first twoaluminum ions are bound to calmodulin in a positive cooperative manner. This observation is consistent with results derived from kinetic data of calcium binding to the protein [30]. 41 IOOP an m 0 . O 1 is O FLUORESCENCE (7. Max i mum) N O 7 l 1 L l 0 l 2 3 4 5 [METAL] / [CALMODULIN] Fig. 4. Tyrosine fluorescence of bovine brain calmodulin as a function of metal concentration. A 9pM concentration of calmodulin was prepared in 100 mM NaCl. pH 6.5. After addition of EGTA to give a final concentration of 100 uM. the protein solution was titrated with CaClz (O) or AlCl, (0). 100% fluorescence intensity corresponds to the intensity emitted from calmodulin prior to EGTA and metal additions. Excita- tion was performed at 280 nm. emission was recorded at 320 nm. The dequenching experiments were carried out in the presence of a chelator. This chelator. EGTA. does not bind significantly to calmodulin [27]. Furthermore. the removal of calcium by excess chelator from calmodulin was independent of che- lator type and concentration [30]. Binding of EGTA to calcium-binding proteins has been de- scribed [12]. Stability constants. expressed as log K. for aluminum-EGTA and calcium-EGTA. at the pH used. are 3.97 and 3.79. respectively [31]. Therefore. the observed changes are not a result of differences in metaloEGTA interactions. Extrapolation of the data presented in Fig. 4 indicates a value for the dissociation constant of about 0.1 pM. for the first mol of aluminum bound. This compares well with similarly obtained values for aluminum dissoCiation constants de- rived from data obtained from CD and ANS experiments. viz.. 0.2 and 0.4 pM. respectively. Dissociation constants for calcium binding derived from these data correspond to published values of about 1 uM [3]. Considering the cooperative char- acter of metal binding. we confined ourselves to estimating the dissociation constant for the first 17 42 mol of aluminum bound to calmodulin. These spectroscopically derived dissociation constants for aluminum binding agree with a value of 0.4 pM derived from equilibrium dialysis studies. Results from the latter studies also show the existence of three metal-binding sites on the protein (Fig. 5). We believe that the dissociation constant. ranging from 0.1 to 0.4 uM. is indicative of the order of magnitude of the high-affinity binding constant for metal binding. Clearly. a thermodynamic anal- ysis is necessary for a detailed evaluation of dis- sociation constants. Aluminum inhibits phosphodiesterase Calciumodependent calmodulin activation of 3’.5’-cyclic nucleotide phosphodiesterase is a prime example of the modulatory role for calmodulin [6.32]. Under our experimental conditions. addi- KD - 3.74 x 10‘7M r2'0.96 aouuo ([ALUMINIM] / EALNDDULND/FREE ([ALUMINUM]1 0.8 _ 0.6 - 0.4 r- 02 r’ 00 l 42 3 4 BOUND ([ALUMINUM] / [CALMODULIN]) Fig. 5. Binding data from equilibrium dialysis experiments. presented as a Scatchard plot. 5 pM calmodulin was dialyzed against various amounts of MCI, in 10 mM Tris-HCI. pH 6.5. for 24 h. 1-ml aliquots of solution. each from inside and outside of the dialysis bag. were assayed for aluminum as described in Materials and Methods. r2 is the correlation coefficient. ' CALMODULIN ' STIMULATED ' 1 1 1 1 1 l 40" ‘ 1 1 1 1 1 PHOSPHODIESTERASE ACTIVITY % CALCIUM [ALLMINUM] /[cawoouu~] Fig. 6. Inhibition of calcium-calmodulin-stimulated 3'.5'ocyclic nucleotide phosphodiesterase activity by aluminum. The en- zyme was incubated with 25 jiM cyclic GMP. 5.5 pM calmodulin and 25 uM CaClz. in 10 mM Tris~HC1. pH 6.5. Incubation times varied between 15 and 30 min. 100% inhibi- tion equals the basal enzyme activity observed with calcium absent. The enzymatic activity was determined as described in Materials and Methods. The enzymatic activity was measured in the presence (0) or absence of calcium (I). tion of calcium doubled the initial enzymatic activ- ity from 0.28 to 0.61 nmol cyclic GMP hydro- lyzed/ml per min. Elevated activation has been shown to depend on the presence of imidazole and millimolar concentrations of Mg2+ [6.26]. Titra- tion of aluminum into the assay system containing 25 uM calcium lowered the hydrolyzing activity to levels equivalent to basal activity in the absence of calcium. The aluminum concentration which inhibited the enzymatic activity by 50% is calcu- lated to be 15 pM. representing a molar ratio of 3: 1 for [aluminuml/[calmodulin]. Results from separate experiments indicated that in the absence of Ca2+ aluminum did not interfere with the basal phosphodiesterase activity. Aluminum appeared to interact with calmodulin rather than the enzymatic protein (Fig. 6). Discussion The results presented in this study show that aluminum binds stoichiometrically to bovine brain 18 calmodulin. This indicates that aluminum ions are bound to specific sites on calmodulin. Considering the difference in charge and that of crystal or hydrated radii [33]. it is doubtful whether the aluminum-binding sites are identical to those for calcium. The binding regions for aluminum may overlap those for calcium and the curvature of the respective binding loops may also vary. There seem to exist two ‘high-affinity‘ binding sites. as demonstrated. for example. in our fluorescence experiments. Our estimates indicate that the ‘high-affinity‘ binding for aluminum to calmodu- lin is about one order of magnitude stronger than that of calcium to its comparable site. Similar to the interaction of Ca2+ to calmodulin. two aluminum ions per protein are sufficient to induce the major structural change as evidenced by CD and ANS fluorescence studies. However. a molar ratio of 4: 1 is required to block the phos- phodiesterase activity maximally. Considering the pronounced aluminum-induced conformational changes. at least for the first two ions. we have to consider the possibility that the higher binding constant. relative to that for calcium. results from a positive entropy contribu- tion. A major. factor for such a positive change is probably the release of coordinated water mole- cules around the metal ion upon complex forma- tion with the protein. Both Al“ and Ca2+ have a primary coordination number of 6. which repre- sents the effective hydration number of the inner- most hydration shell [34]. It has been suggested that there exists a single water molecule at the high-affinity binding site of troponin C. a protein which is closely related to calmodulin [35]. There- fore. five coordinated water molecules are to be released for each calcium ion bound to the protein. The entropy of calcium binding to the troponin sites 1 and 2 has been measured to be about 15 e.u. [36]. The strength of these coordinated bonds. responsible for ion /water interactions. is apprecia- bly higher for A1“. which has an hydration en- thalpy of 1 144 kcal/mol of ion. as compared to a value of 399 kcal/mol for Ca2+ [37]. Moreover. the value of the formal charge over ionic radius and the intermediate electronegativity are both a factor of 2 higher for A13+ relative to Ca“ [37]. Also the number of water molecules in the outer hydration shells of Al“ is significantly higher 43 than that for Ca2+ [34]. Considering these physical data. interactions of aluminum ions with calmodulin are expected to differ appreciably from those of Ca2+. In particular. the solvent structure around protein ligands should depend on the re- spective coordinated metal ion. As far as calcium is concerned. it is known that this ion appears to be coordinated to oxygen ligands and mobility of the ion is restricted considerably when bound to the protein [38.39]. A13+ also forms stable com- plexes with electronegative ligands such as oxygen and nitrogen. In contrast to calcium. the higher electronegativity of Al3+ would result in an in- creased covalent character of the coordinate bond established. Consistent with our observations. the stability of the aluminumocalmodulin complex would therefore be increased as compared to that of the calcium complex. Aluminum-induced conformational changes in calmodulin are also documented by our CD stud- ies. While Ca“. and the spatially isomorphous lanthanide. Tb“. promote helix formation [27]. the helix content is decreased by about 30% when Al3+ is present with calmodulin at a molar ratio of 4: 1. Charge differences per se cannot account for this metal type-dependent change in helix struc- ture since the trivalent ion. Tb“. promotes helical formation. just as Ca2+. Rather. the unique physi- cochemical characteristics of the highly solvated aluminum ion. as discussed above. are hypothe- sized to be responsible for the observed structural alterations. It is thought that upon binding of Ca2+ to calmodulin negative charges on the pro- tein will be neutralized. This in turn results in weakening of constraining forces. thus permitting the formation of additional helical elements. Fi- nally. a more compact calcium-calmodulin com- plex emerges [29]. As far as interactions of Al3+ with calmodulin are concerned. their molecular origin and significance are presently unclear. and further experiments are necessary. What can be stated is that under our experimental conditions mononuclear. hydrated aluminum species are pre- sent. as opposed to polynuclear species existing at higher pH values and elevated aluminum con- centrations [40]. As to thermodynamic changes. application of aluminum ions to calmodulin leads to a helix-coil transition. which is accompanied by a strong enhancement of the protein‘s hydro- 19 44 phobic surface domains. Such behavior is to be expected for a randomly coiled polypeptide with an increased portion of hydrophobic components. Simply stated. the hydrated aluminum ions pro- mote an open. solvent-rich. disordered polypeptide region whereas calcium ions promote a peptide environment where intramolecular interactions be- tween adjacent peptide elements are favored. The importance of ion-dependent changes in solvent structure is further exemplified by qualitative ex- periments showing that calmodulin aggregates above a molar ratio of 10:1 for [aluminum]/ [calmodulin]. in contrast to calcium-calmodulin complexes. In summary. it appears that aluminum binding to specific regions of calmodulin results in local structural changes which. in turn. have profound consequences for the relative motion of distinct. internal structural domains [41]. As a result. the protein's flexibility and its ability to interact with various proteins is impaired. These kinds of changes in calmodulin may explain why the aluminum-calmodulin complex lost at least part of its regulatory character. This complex may thus be a key lesion that occurs in the broadly defined syndrome of aluminum toxicity. if viewed in terms of lost regulatory capacity. Acknowledgements We thank Mr. Charles G. Suhayda for provid- ing the tyrosine fluorescence data. Dr. J. Speck. Biochemistry Department. Michigan State Univer- sity. for his technical assistance with the CD in- strument. Dr. J. Shafer. Department of Biological Chemistry. University of Michigan. for the use of the computerized CD instrument. Dr. M. Lhedin for her assistance with atomic absorption measure- ments. and Dr. P. Wagner for the use of the Perkin-Elmer spectrofluorimeter. This work was supported by US. Department of Energy Contract No. DE-AC02-76ER01338. References - 1 Cheung. W.Y. (1970) Biochem. Biophys. Res. Comm. 38. 533-537 2 Klee. CB. (1980) in Calcium and Cell Function (Cheung. W.Y.. ed.). Vol. 1. pp. 59-77. Academic Press. New York 3 Klee. C.B.. Crouch. T.H. and Richman. PG. (1980) Annu. Rev. Biochem. 49. 489-515 4 Cormier. M.J.. Anderson. J.M.. ('harbonneau. H.. Jones. HP. and McCann. RD. (1980) in Calcium and Cell Func- tion (Cheung. W.Y.. ed.). Vol. I. pp. 201-218. Academic Press. New York 5 Wolff. DJ. and Brostrom. C .O. (1979) Adv. Cyclic Nucleo- tide Res. 11. 28-88 6 Wang. J.H. and Waisman. D.M. (1979) Curr. Topics Cell. Reg. 15. 47-107 7 ‘Schulman. H.. Huttner. W.B. and Greengard. P. (1980) in Calcium and Cell Function (Cheung. W.Y.. ed). Vol. 1. pp. 219-250. Academic Press. New York 8 Lin. Y.M. (1982) Mol. Cell. Biochem. 45. 101—112 9 Rasmussen. H.. Goodman. D.B.P. and Tenenhouse. A. (1972) CRC Crit. Rev. Biochem. 1. 95-148 10 Crouch. T.H. and Klee. CB. (1980) Biochem. 19. 3692-3698 11 Wolff. DJ.. Poirier. P.G.. Brostrom. CO. and Brostrom. M.A. (1977) J. Biol. Chem. 252. 4108-4118 12 Wallace. R.W.. Tallant. E.A.. Dockter. ME. and Cheung. W.Y. (1982) J. Biol. Chem. 257. 1845-1854 13 Kilhoffer. M.C.. Demaille. J.G. and Gerard. D. (1980) FEBS Lett. 116. 269-272 14 LaPorte. D.C.. Wierman. B.M. and Strom. DR. (1980) Biochemistry 19. 3814-3819 15 Smoake. .I.A.. Song. S.Y. and C heung. W.Y. (1974) Bio- chim. Biophys. Acta 341. 402-411 16 Malencik. DA. and Anderson. S.R. (1982) Biochemistry 21. 3480-3486 17 Crapper. D.R.. Karlik. S. and De Boni. U. (1978) in Alz- heimer's Disease: Senile Dementia and Related Disorders (Katzman. R.. Terry. RD. and Dick. K.L.. eds). Aging. Vol. 7. pp. 471-485. Raven Press. New York 18 Foy. C.D.. Chaney. R.L. and White. M.C. (1978) Annu. Rev. Plant Physiol. 29. 511-566 19 Caldwell. CR. and Haug. A. (1981) Anal. Biochem. 116. 325-330 20 Merril. C .R.. Goldman. D.. Sedman. SA. and Ebert. M.I-l. (1981) Science 211. 1437-1438 Wang. CS. and Smith. R.L. (1975) Anal. Biochem. 63. 414-417 22 Lambert. J.B.. Shurwell. H.F.. Verbit. L.. C ooks. RC. and Stout. CH. (1976) Organic Structural Analysis. p. 341. MacMillan. New York 23 Chen. Y.H.. Yang. J.T.. and Martinez. H.M. (1972) Bio- chemistry 11. 4120-4131 24 Greenfield. N. and Fasman. GD. (1969) Biochem. 8. 4108-4116 25 C ornish-Bowden. A. (1979) Fundamentals of Enzyme Kinetics. pp. 147-176. Butterworths. London , 26 Dedman. J.R.. Potter. J.D.. Jackson. R.L.. Johnson. 10. and Means. A.R: (1977) J. Biol. Chem. 252. 8415-8422 27 Wang. C .L.A.. Aquaron. R.R.. Leavis. RC. and Gergeley. J. (1982) Eur. J. Biochem. 124. 7-12 28 Smith. M.B.. Niemeczura. W.P.. Murtaugh. TJ.. Siegel. FL. and Gibbons. W.A. (1981) Fed. Proc. 40. 1871 29 Seamon. KB. (1980) Biochemistry 19. 207-215 2 31 32 35 20 White. H.. Tudor. M.. Bose. K. and Markley. D. (1982) Biophys. J. 37. 51a Martell. A.E. and Smith. R.M. (1974) Critical Stability Constants. Vol. 1. p. 271. Plenum Press. New York Goldberg. ND. and Haddox. M.K. (1977) Annu. Rev. Biochem. 46. 823-896 Nightingale. ER. (1959) J. Phys. Chem. 63. 1381-1387 Hindman. J.C. and Sullivan. J.C. (1971) in Coordination Chemistry (Martell. A.E.. ed.). Vol. 1. pp. 393-426. Van Nostrand Reinhold Co.. New York Wang. C.L.A.. Leavis. P.C.. Harrocks. W.D.W. and Gerge- Icy. J. (1981) Biochemistry 20. 2439-2444 36 37 38 39 41 42 45 Potter. J.D.. Hsu. FJ. and Pownall. HJ. (1977) J. Biol. Chem. 252. 2452-2454 Noller. H. (1982) Acta Chim. Acad. Sci. Hung. 109. 429-448 Andersson. T.. Drakenberg. T.. Forsen. S.. Thulin. E. and Sward. M. (1982) J. Am. Chem. Soc. 104. 576-580 Krebs. J. (1981) Cell Calcium 2. 295-311 Baes. CF. and Mesrner. R.E (I976) The Hydrolysis of Cations. pp. 112-123. J. Wiley and Sons. New York Karplus. M. and McCammon. J.A. (1981) CRC Crit. Rev. Biochem. 11. 293-349 Lowry. O.H.. Rosebrough. NJ.. Farr. A.L and Randall. R.J. (1951) J. Biol. Chem. 193. 265-275 Chapter 4 Calmodulin—Dependent Formation of Membrane Potential in Barley Root Plasma Membrane Vesicles: A Biochemical Model of Aluminum Toxicity in Plants PHYSIOL. PLANT. 59: 285—291. CW 1983 Calmodulin-dependent formation of membrane potential in barley root plasma membrane vesicles: A biochemical model of aluminum toxicity in plants NealSiegelandAlfredHaag Siegel. N. and Haug, A. 1983. Calmodulinodependent formation of membrane po~ tential in barley root plasma membrane vesicles: A biochemical model of aluminum toxicity in plants - Physiol. Plant. 59: 285-291. Micromolar concentrations of aluminum ions interfere with calmodulin-stimulated. membrane-bound ATPase activity which plays a role in the maintenance of the transmembrane potential of plasma membrane-enriched vesicles isolated from barley roots. Calmodulin appears to be the major target for aluminum interaction resulting in pronounced changes in the exposure of a large. hydrophobic surface on this protein as determined with a fluorescent. hydrophobic surface probe. At a molar ratio of 3:1 [aluminum]/[calmodu|in]. the calmodulin-stimulated enzymatic activity. probably associated with a Ca” + Mg‘*-ATPase. is about 95 % inhibited. Aluminum-induced changes in calmodulin structure are reflected in reduced formation of the membrane potential when assayed with a fluorescent potential probe. oxonol Vl. We hypothesize that the aluminum-calmodulin complex represents a primary lesion in toxic responses of plants to this metal. Additional key words - Ca" + Mg’*-ATPase activity. N. Siegel, MSU-DOE Plant Research Laboratory. Michigan State Univ., Easr Lans- ing. MI 48824. USA; A. Hang (present address and reprint requests), Pesticide Re- search Center. Michigan State Univ., East Lansing, MI 48824, USA. lamdactioa Following mobilization at acidic pH. aluminum ions are potent, toxic agents to plants. Elevated soil aluminum levels were shown to impair root elongation and to in- terfere with the uptake, distribution and use of calcium, magnesium. phosphorus and other essential minerals (Foy et al. 1978). Elevated tissue aluminum levels were reported to cause chloroplast membrane degeneration and to decrease CO, fixation in spinach (Hampp and Schnabl 1975). Upon application of aluminum to barley roots, plasma membrane degeneration seems to be an early indication that aluminum is present at toxic levels (Hecht-Buchholz and Fay 1981). The mode of interac- tion of aluminum with living tissue is unknown although it has been observed that aluminum binds to ATP, which in turn impairs yeast hexokinase (Viola et al. 1980). Aluminum can also bind to DNA and impair genetic expression (Foy et al. 1978). Among several Received 29 March. 2983'. revised 24 May. 1983 m m. :9. l9” varieties of wheat. barley and soybeans. aluminum tol- erance seems to be associated with resistance against aluminum-induced calcium deficiency or reduced cal- cium transport, and has been shown in barley to be linked to a single. dominant gene (Foy et al. 1978). As a result of aluminum-induced changes in calcium uptake and utilization, biochemical systems dependent on calcium for regulation may be targets of aluminum ions. In plant as in animal cells. free calcium levels are strictly regulated and do not exceed micromolar con- centrations (Clarkson and Hanson 1980). Activation of key, calcium-regulated processes in many instances in- , volves calmodulin. This small (MW = 17 000). acidic, calcium-dependent regulatory protein responds to transient increases in intracellular calcium levels and has been shown to regulate Ca“-ATPase activity in plants (Caldwell and Haug 1981a. Dieter and Marine 1980, 1981). In addition. aluminum ions have been shown to interact with calmodulin in such a way that the 285 23 protein's regulatory capacity diminishes and becomes lost with increasing molar ratios of [aluminum]/[cal- modulin] (Siegel and Haug 1983). In the present study. we intend to demonstrate that the calmodulin-regulated. Ca“— and Mg’*—dependent ATPase activity in barley root plasma membranes is electrogenic and interfered with by aluminum. Effects of various metals. pH, calmodulin, ionophores and aluminum on the electrogenic activity are described. These data will be dicussed in terms of lost regulatory capacity by calmodulin resulting in the failure of the plasma membrane to maintain an elctrical potential necessary for proper cell maintenance. Abbreviations - ANS. 8-anilino-l-naphthalene sulfonic acid: CCCP. carbonyl cyanide m-chlorophenylhydrazone; DTT. dithiothreitol; MES, morpholinoethane sulfonic acid; MOPS. 4omorpholinopropane sulfonic acid; oxonol V1. propyl oxonol; PMSF. phenylmethylsulfonylfluoride; SDS, sodium dodecyl sulfate. Materialsand methob Caludaflnpreparfllaa Calmodulin was prepared from bovine brain acetone powder as previously described (Caldwell and Haug 1981b). However, to increase purity of the yield, the protein-loaded affinity column was washed with a buf- fer containing 500 mM NaCl (Siegel and Haug 1983). For our experiments we used bovine brain calmodulin since its structure is well known (Klee et al. 1980) and because this protein stimulates barley root plasma membrane ATPase activity in the identical way to cal- modulin extracted from barley (Caldwell and Haug 1981a) Che-leak Bovine brain acetone powder, D'IT, EDTA, MES MOPS, PMSF. sodium-ATP and Tris were purchased form Sigma Chemical Co. (St. Louis, MO). MCI, and CaCl, were purchased from Mallinckrodt Science Pro- ducts (St. Louis, MO). Affigclphenothiazine and Chelex-lOO were purchased from Bio-Rad Laborato- ries (Richmond, CA). The sodium salt of ANS was obtained from K&K Laboratories (Plainview, NY). A-23187 and nigericin were obtained from Calbio- chem-Behring Corp. (La Jolla. CA). Chlorpromazine HCl and trifluoperazine HCI were obtained from Smith. Kline and French Labs. (Philadelphia. PA). Oxonol VI was purchased from Molecular Probes. inc. (Junction City, OR). All other chemicals were of the highest purity commercially available. Pharmaterialaadgrowiagcondltioas Barley seeds (Hordeum vulgare L. cv. Conquest) were 286 germinated and grown in the dark at 16°C over aerated solutions of 0.25 mM C3804. adjusted to pH 5.0 with H,SO.. The growth medium was changed daily. Six days after imbibition the primary roots were washed in chilled, distilled water and excised. All handling of the material was done at 4°C. Preparatioamedia Homogenizing medium: 0.25 M sucrose, 3 mM EDTA, 1 mM PMSF. and 1 mM Naz-ATP were prepared in 25 mM Tris-MES buffer adjusted to pH 7.2. Sucrose gra- dient: The discontinuous gradient was prepared from 34% and 40% (w/w) sucrose solutions in 1 mM Tris- MES. pH 7.2. Wash solution for the isolated mem- branes; 0.25 M sucrose in 1 mM Tris-MES buffer, pH 6.5. Mutt-abomin- Ar plasma membrane enriched microsome fraction was isolated according to the method of Nagahashi et al. (1978), as modified by Caldwell and Haug (1980) with minor modifications, using the following scheme: 1. Roots (approximately 200 g) were homogenized in 3 ml of homogenizing solution per gram of tissue by hand using a ceramic mortar and pestle without the addition of any abrasive material. . Filtration through miracloth. . Centrifugation l 300 g, 15 min, discard pellet. . Centrifugation 80 000 g, 30 min. supernatant dis- carded. . Pellet resuspended in wash solution. . Resuspended membranes loaded onto discontinu- ous gradient. . Centrifugation 80 000 g. 2 h. . Membrane at interface (34%I40%) removed. . Interface resuspended in wash solution. . Centrifugation 80 000 g, 30 min. . Resuspend pellet in wash solution at a concentra- tion of 0.5 mg vesicle protein ml". . Membranes were stored for a maximum of 24 h on ice in a cold room kept at 4°C. O‘Ut #UN HOOMQ t—a—a Determination of membrane potential Fluorescence measurements of oxonol Vl werd made in a 2 ml reaction volume containing various concentra- tions of divalent cations. calmodulin, aluminum or - ionophores at the desired pH in the wash medium. Vesicles were diluted to 50 pg vesicle protein ml“. All experiments were carried out at 16°C. After a 15 min preincubation of the vesicles in the adjusted wash medium, oxonol VI was added from a concentrated stock solution to a final concentration of 0.3 W. The solution was then transferred to a‘ quartz cuvette with a 10 mm optical path length which was placed into a thermostatted holder of a fluorimeter (Jen and Haug 1981). The excitation and emission wavelengths were Physiol. Hunt. 59. m3 24 580 nm and 640 nm. respectively. Light scattering was negligible. A baseline was established by adjusting the suppres- sion current of the picoammeter until the value had stabilized. An aliquot (20 ul) of 100 mM Tris-ATP (Hodges and Leonard 1974) was injected from a mi- crosyringe into the sample to begin the reaction. Fluorescence signal changes were monitored and re- corded for up to 20 min. First order rate constants, k. were calculated from the relationship F = Fo° exp(-kt). using a Tektronix 4051 computer and non-linear, least- squares equation fitting programming F and F0 are the fluorescence intensities at times t and t - 0, respec- tively. ANSI-oresceaeetae-are-eats Fluorescence intensity measurements were performed on a Perkin-Elmer spectrofluorimeter, model MPF-44A, equipped with a differential corrected spec- tral unit. The excitation and emission wavelengths were set at 360 t 4 nm and 490 i 4 nm, respectively. Cal- modulin was prepared in 10 mM MOPS, pH 6.5. at a final concentration of 10 W. ANS was added from a concentrated stock solution to a final concentration of 2 M. A quartz cuvette of 10 mm optical pathlength was used. The ANS fluorescence intensity of a calmodulin solution in the absence of metal ions was considered to be the initial fluorescence intensity value. Data for ANS fluorescence in the presence of metal ions are expressed as the relative increase of the initial fluorescence inten- sity value. A value of zero is defined as the fluorescence of the solution without calmodulin present. ma. aadysh Protein was quantified with bovine serum albumin as a standard (Wang and Smith 1975). Calmodulin con- centrations were also adjusted spectrophotometrically using a value for the molar extinction coefficient of 3 300 M" cm“ at 277 nm (Crouch and Klee 1980). Rwanda-hating“ All glassware and quartz cuvettes were washed with concentrated nitric acid. Buffer solutions were prepared in double. glass-distilled, deionized water and passed over columns (2 cm x 30 cm) of Chelex-IOO resin. Metal stock solutions were freshly prepared in metal- decontaminated buffers. Plasma emission and atomic absorption spectroscOpy indicated that the buffer solu- tions and the Tris-ATP preparations typically contained less than 10" M calcium, magnesium. and aluminum. This quantitative analysis was performed on a Janell- Ash plasma emission spectrometer, model 955 Atom- -comp. and a Varian atomic absorption spectrophoto- meter. model 1475. m. run. so. toss Data-alyh The presented data are means of at least three indepen- dent experiments : standard deviation except when otherwise noted. Results OxonolWQapmbeoftraasmembranepoteatial The fluorescent oxonol probe is voltage sensitive and is therefore useful for evaluating changes in transmem- brane potentials (Beeler et al. 1981. Smith et al. 1981). Subsequent redistribution of the dye in beef-heart sub- mitochondrial vesicles was shown to produce a fluores- cence intensity loss upon Mg3°~ATP stimulation of ATPase activity involved in establishing an electro- chemical gradient across these membranes. Our results derived from time-dependent fluorescence signal changes are qualitatively consistent with these data. The fluorescence decay curves from our experiments were best fit by a single exponential function which is consis- tent with data from the literature and is a consequence of the low affinity of oxonol V1 for phospholipid mem- branes (Smith et al. 1981 ). The direction of fluorophore distribution is across membranes with an inside-positive potential as a result of this probe‘s delocalized negative charge (Beeler et al. 1981). Addition of ATP to the vesicle suspension activated the Ca“-ATPase resulting in a sustained. fluorescence intensity decrease (Fig. 1). As a control, vesicles were boiled for 2 min; immediately upon addition of ATP, the fluorescence intensity of oxonol increased slightly, but quickly reached a new steady-state. Upon preincu- bation of vesicles in 0.01% SDS or Triton X-100 (re- FLUORESCENCC INTENSHY —. Fig 1. Changes in fluorescence intensity of oxonol V1 in the presence of membrane vesicles from barley roots. Traces show responses of (A) vesicles which were boiled for 2 min. or (B) vesicles which were untreated. Conditions for assay: 0.05 mg vesicle protein ml“, 0.25 M sucrose. 1 mM Tris-MES (pH 5.5), 1 mM CaCl,. 0.3 nM oxonol Vl. 16°C. The vesicles were allowed to pre-incubate until a stable value for the fluores- cence signal was established. Tris-ATP was then added to 1 mM (arrow). 287 25 sults not shown). the fluorescence signal was indistin- guishable from that of the dye in the absence of vesicles. Analysis of the fluorescence signal after ATP activation of the vesicle suspension indicated that. in some cases, two components were present. The first. fast component became negligible within 1 to 2 min and was considered as resulting from turbulence in the assay solution after mixing. First order rate constants were calculated from all traces after 5 min to avoid interference. pl-l dependenceoftheCa"--dMg’°-depeodeat ATPaseactivltyaadltstelatioatothedevelopmeatofths Mamie-spots.“ Ca"- and Mg“-ATPase activities were independently activated by the respective cations and the effect of each on transmembrane potential development compared (Fig. 2). The observed pH response profiles for both activities are consistent with those results previously derived for the high-affinity Ca"- and Mg’°-ATP hy- drolyzing activities of the ATPase in these same mem- brane vesicles (Caldwell and Haug 1980). Calcium at 1 mM sustained a high rate of potential development at pH 5, but dropped below the value for 1 mM Mg“ at increasing pH values. Above pH 5.5. Mg2+ sustained a fairly constant rate between 0.12 min" and 0.11 min“. whereas the rate sustained by Ca’+ decreased 72% from 0.11 min" to 0.03 min" in the same pH range. Under these conditions the concentration of free cal- cium is higher than that found intracellularly in vivo. it is for the purpose of comparison with previous enzyma- tic studies that these concentrations of the divalent ca- tions were used (Caldwell and Haug 1980).'1he decline in the ATPase-controlled membrane potential change may be related to the effects of positively charged en- zyme-substrate complexes at lower pH for both Ca”- and Mg’°-ATPase activities, whereas the rates meas- ured at elevated pH values (above pH 6.0) may be the result of deprotonation leading to charge neutralization of the enzyme-substrate complex. The pH dependence L L t 4 5 6 7 8 DH Fig. 2. pH dependence of the rate constant. k. for the de- velopment of the transmembrane potential. Vesicles were pre- -incubated under the conditions described in the legend to Fig. l. with 1 mM CaCl, (0). or 1 mM MgCl, (0). Tris-ATP was added to 1 mM. Vertical bars showthe 288 of the development of the transmembrane potential un- der conditions in which the Ca“- or Mg’°-ATPase ac- tivity is stimulated coincided with the enzymatic hy- drolyzing activity as previously described (Caldwell and Haug 1980). Effectofdivalsatcatloacoaceatratioaoapoteat'nl development Development of the membrane potential was more sensitive to changes in MgCl, concentration than to changes in CaCl, concentration (Tab. 1). At a physiological pH of 6.5, there was a two-fold decrease in the rate of potential development when the Mg“ concentration was changed from 1 mM to 100 W, with Ca3+ absent. Addition of 10 w Ca” to either Mg2° concentration did not significantly alter the rate of po- tential formation. The effect of various Ca2+ concentra- ~ tions in the absence of Mg’° was slight and less than doubled over a three order of magnitude concentration range from 1 M to 1 mM. Where it has been studied, the concentration of free cytosolic Ca3+ ranges between 0.1 and 10 pH with Mg2+ tending to be 10- to 100-fold more concentrated (Clarkson and Hanson 1980). Mg‘° and Ca2+ concentrations of 100 nM and 10 W, respec- tively. were chosen as representative of physiological concentrations of these divalent cations; the stimulated ATPase activity developed a potential change with a rate constant of 0.08 min“ and is referred to as the control value. Results summarized in Tab. 2 indicate that both the H‘lMe° ionophore nigericin and the 2H*/Me’° ionophore A-23187 effectively prevented the formation of a transmembrane potential. This is true whether the vesicle suspension is preincubated with the respective ionophore or, as shown in Fig. 3, if the respective ionophore is added after the potential was allowed to develop. in either case the ionophoretic concentration Tab. 1. Effects of varying divalent cation concentrations on the rate of potential development. Conditions. 0. 05 mg vesicle protein ml“, 0. 25 M sucrose. 1 mM Tris-MES buffer (pH 6. 5). 0. 3 uM oxonol VI. 16°C. After a stable fluorescence signal was established. Tris-ATP was added to 1 mM. Divalent cation Rate. k. (min")tso 1 mM Mg’° ..................... 011610.014 1 mM Mg" + 10 w Ca” ....... 010710.043 100 w Mg” ................... 0.063 .+.0.021 100 Mt Mg” + 10 MM c." ..... \ 008010.009 1 w Ca” ...................... 0019:0002 10 W Ca” ..................... 002910.003 100 W Ca” .................... 0038:0004 1 mM Ca" ..................... 004210.003 fly“. I“. 59. I”) 26 Tab. 2. Effects of various compounds on the development of the membrane potential. Conditions were as indicated in Tab. 1. except 50 pH Ca’° and 100 M Mg2° were used. Valuesare within 5% ss Compounds Rate. k. % Control (min") activity No addition (control) ................ 0.08 100 A-23187 (10 W) ................... 0 0 Nigericin (10 W) ................... 0 0 Trifluoperazine (10 W) ............. 0.07 87 Chlorpromazine (10 W) ............ 0.08 100 AlCl, (10 pM) ...................... 0.08 100 Calmodulin (10 M) ................ 0.24 300 + Chlorpromazine (10 W) ......... 0.09 112 + Trifluoperazine (10 W) ......... 0.08 100 + AlCl, (10 W) ................. 0.17 210 + CCCP (5 uM) .................. 0.04 16 was below that which would cause detergent-like dis- ruption of the membranes. The protonophore CCCP is somewhat less effective in preventing potential de- velopment than the other ionophores, but it did reduce the calmodulin-stimulated rate by 85%. In the absence of calmodulin. incubation of the vesicle suspensions with the calmodulin-antagonists chlorpromazine or trifluoperazine (Klee et al. 1980) allowed normal de- velopment of the ATP-stimulated potential, as did in- cubation with 10 M AlCl,. Incorporation of 10 14M calmodulin in the assay suspension increased the rate of potential development by 200% upon ATP activation. Chlorpromazine and trifluoperazine both reversed the calmodulin-dependent activity at a molar ratio of 1:1 for [drug]/[calmodulin], whereas the decrease was only 30% at the same ratio of [aluminum]/[calmodulin]. mmumwa, AWMofthem-br-e psteathl The viability of seedlings growing in aluminum-toxic soils is known to decrease dramatically as compared with those seedlings growing in soils low in aluminum (Foy et a1. 1978). Primary effects of aluminum toxicity apparently occur at the plasma membrane (Hecht- Buchholz and Foy 1981). Thus the Ca’*- and Mg’°-de- pendent ATPase activity described in this report may serve as a potential marker for detrimental actions of ' aluminum on membrane maintenance machinery. Fig- ure 4 shows the results of experiments in which vesicle suspensions were incubated with 50 w Ca“. Under these conditions the rate of potential build-up did not differ from that rate measured at 10 uMCa". Incuba- tion of the vesicle suspensions with 10 M calmodulin allowed the potential to form at a rate of 0.24 min" with 50 W Ca“ present. Addition of up to 40 («M AlCl, accelerated the loss of the calmodulin-stimulated 19 Physiol. m. 59. 1903 ATP i l ’2 § i S I AEP a E; Fig 3. Effects of ionophores on the development of the trans- membrane potential. Vesicles were pre-incubated as described in the legend to Fig. 1. with 100 M CaC1,. Tris-ATP was added to 1 mM (arrow). (A). addition of 10 Mt A-23187 or nigericin (l); (B) pre-incubation with A-23187 or nigericin prior to ATP addition. activity with a 50% value falling at a molar ratio of' 1.4/1 [aluminum]/[calmodulin]. On the other hand, no calmodulin-stimulated activity was detected above 30 W AlCl3 although these aluminum concentrations did not interfere with the basal ATPodependent potential development (in the absence of calmodulin). ' The observation that the nonocalmodulin stimulated activity was not interfered with prompted the investiga— tion of the interaction of aluminum with the isolated protein. The experimental results are reported as the relative fluorescence intensity of ANS. a probe used for monitoring the hydrophobic surface properties of pro- teins (La Porte et al. 1980). Maximum hydrophobic surface exposure on the protein was achieved at a ratio U 1 b 0‘- mm wrumrsccucrt-i 42 0.08; 1 o :1 ° 0 i0 20 30 4‘“ [All pM ‘ Fig. 4. Effect of aluminum on calmodulin-stimulated potential development or on ANS partitioning onto isolated calmodulin. Vesicles were preincubated as described in the legend to Fig. 1, with 100 W MgCl,. 50 W CaCl,. 10 M4 calmodulin and various AlCl, concentrations (0). A rate constant. k. of 0.08 min" represents non-calmodulin-stimulated activity; a rate constant of 0.24 min“ represents maximal calmodulin-stimu~ lated activity. Relative ANS fluorescence was measured in the presence of 10 pM calmodulin and various AlCl, concentra- tions (I). Values are within 5% st. 289 27 of 3:1 [aluminumlllcalmodulin] (mol:mol). the 50% ratio is 1.5/ 1. Thus. a parallel was established between loss of the calmodulin-stimulated potential develop- ment and the loss of structural integrity of calmodulin in the presence of aluminum. Discussion The results of this study demonstrate that toxic aluminum ions interfere with calmodulin-stimulated ATPase activity which plays a role in the maintenance of the membrane potential. Calmodulin appears to be the major target for aluminum. This protein has been shown to mediate calcium reg- ulation in plant enzyme systems (Cormier et a1. 1980. Dieter and Marnie 1981). Calmodulin has the capacity to bind four calcium ions at specific sites on each molecule. The resulting conformational changes en- hance the hydrophobic surface exposure which is ap- parently necessary for proper interfacing of cal- cium-calmodulin and its target protein (Crouch and Klee 1980. Lin 1982). One such target protein seems to be the barley root plasma membrane-bound ATPase found in plasma membraneenriched vesicles as de- scribed in this report, since our results indicate that, at physiological calmodulin concentrations of 10 Ml (Klee et al. 1980, Wang and Waisman 1979), electrogenic activity is stimulated by 200% over the activity observed in the absence of calmodulin. As to the cal- modulin-stimulated electrogenic activity, our results are consistent with the existence of a membrane-bound ATPase involved in pumping protons or in coupled proton fluxes. The pumping activity leads to the estab- lishment of a membrane potential which, in turn, de- pends critically on the complete regulatory capacity of calcium-calmodulin. Aluminum-induced changes of calmodulin therefore lead to a reduction in electrogenic activity accompanied by a decreased membrane poten- tial. Aluminum ions interact stoichiometrically with calmodulin (Siegel and Haug 1983). The resulting changes in surface hydrophobicity and helix content lead to a loss of regulatory properties of calmodulin as exemplified by the aluminum-induced inhibition of calmodulin-activated phosphodiesterase activity. Questions arise whether the ATPase activity present in the vesicle system used in this study is derived solely from the plasma membrane. The activity of the ATPase as measured by inorganic phosphate release is inhibited 50% at a 100 W N. N‘-dicyclohexylcarbodiimide con- centration, under conditions of pH and divalent metal concentration similar to those used in our studies (A. Lesniak. personal communication). This result is also consistent with the ATPase activity measured in the plasma membrane of corn leaf, corn root and cat root (Perlin and Spanswick 1981). Possible contamination by other membranes including tonoplast may exist. For 290 this reason we refer to the vesicle system as a plasma membrane-enriched microsome fraction. For the purpose of the present discussion we refer to the various species of aluminum present in solution collectively as Al. Under the conditions used in the present study we can state that mononuclear. hydrated aluminum species are present. as opposed to polynu- clear species existing at higher pH values and elevated aluminum concentrations. We cannot be certain. how- ever. of the charge on these hydrated aluminum species (Baes and Mesmer 1976). Although ATP represents a potential chelator of aluminum. no inhibitory effect was evident when aluminum was included in the assay system. with cal- modulin absent. Viola et al. (1980) showed that yeast hexokinase is strongly inhibited by the presence of Al-ATP with an inhibition constant, K,, of 0.16 pH. at pH 7. These data supported the earlier findings by Womack and Colowick (1979) who described Al-ATP asillfibition of yeast and brain hexokinases. A striking characteristic of the aluminum inhibition in these sys- tents was the observation that when the same amounts of aluminum were added to the reaction mixtures, separately from ATP. more than 10 times as much aluminum was required to get a comparable effect. Since we added ATP separately to the reaction mixture our results therefore cannot be compared to those studies in which ATP pre-mixed with aluminum is ad- ded as a substrate. Further support of our results comes from a personal communication by R. Post, cited in Wornack and Colowick’s article (1979), that aluminum has little effect on the Na’. K”-ATPase of guinea pig kidney membranes. We conclude that under our ex- perimental conditions. AloATP is not an inhibiting sub- strate for the ATPase activity. Our hypothesis is that the Al-calmodulin complex represents a primary biochemical lesion in toxic re- spouses of plants to aluminum. Considering the pivotal role of calmodulin in calcium regulation. aluminum in- terference is expected to result in severe imbalances of .cellular processes. such as maintenance of the mem- brane potential. cell growth, root elongation and chloroplast function. This hypothesis is consistent with observations that aluminum-induced toxic responses of plants resemble. in part. those occuring as a result of calcium deficiencies. Acknowledgements - We thank Dr P. Wagner for the use of the Perkin-Elmer spectrofluorimeter. We gratefully acknow- ' ledge the generous gift of trifluoperazine from Smith. Kline and French Laboratories. This work was supported by the US Department of Energy contract No. DE-AC02-76ERO-1338. References Baes. C. F. & Mesmer. R. E. 1976. - The Hydrolysis of Ca- tions. 1. Wiley and Sons. New York. pp. 112-123. ISBN 0-471-03985-3. Physicl. m. 59. ms 28 Beeler. T. J.. Farrnen. R. H. at Martonosi. A. N. 1981. The mechanism of voltage-sensitive dye responses on sarco- plasmic reticulum. - J. Membr. Biol. 62: 113-137. Caldwell. C. R. & Haug. A. 1980. Kinetic characterization of barley root plasma membrane-bound Ca’° and Mg’°-de- pendent adenosine triphosphatase activities. - Physiol. Plant. 50: 183-193. — & Haug. A. 1981a. Calmodulin. stimulation of the barley root plasma membrane-bound Ca”. Mg’*-ATPase. - Plant Physiol. 67 (Suppl): 136. - & Haug. A. 1981b. Affinity chromatographic isolation of calmodulin from bovine brain acetone powder. - Anal. Biochem. 116: 325-330. Clarkson. D. T. & Hanson. J. B. 1980. The mineral nutrition of higher plants. - Annu. Rev. Plant Physiol. 31: 239—298. Cormier. M. J .. Anderson. J. M.. Charbonneau. H.. Jones, H. P. & McCann. R. O. 1980. Plant and fungal calmodulin and the regulation of plant NAD kinase. — In Calcium and Cell Function (W. Y. Cheung. ed.). Vol. 1. pp. 201—218. Academic Press. New York. ISBN 0-12-171401-2. Crouch. T. H. & Klee, C. B. 1980. Positive cooperative bind- ing of calcium to bovine brain calmodulin. - Biochemistry 19: 3692-3698. Dieter. P. & Marine, D. 1980. Calmodulin activation of plant microsomal Ca2° uptake. — Proc. Natl. Acad. Sci. USA 77: 731 1—73 14. - a Marine, D. 1981. A calmodulin-dc ndent microsonial ATPase from corn roots (Zea mays L. .- FEBS Lett. 125: 245-248. Foy. C. D.. Chaney. R. L. & White. M. C. 1978. The physiology of metal toxicity in plants. - Annu. Rev. Plant Physiol. 29: 511-566. Hampg. R. & Schnabl, H. 1975. Effect of aluminum ions on ‘ 0,-fixation and membrane system of isolated spinach chloroplasts. - Z. Pflanzenphysiol. 76: 300—306. Hecht-Buchholz. C. & Foy. C. D. 1981. Effect of aluminum toxicity on root morphology of barley. - Plant Soil 63: 93-95. ‘ Hodges, T. K. & Leonard. R. T. 1974. Purification of a plasma membrane-bound adenosine triphosphatase from plant roots. - In Methods in Enzymology (S. Fleischer and L. Packer. eds). Vol. 328. pp. 392-406. Academic Press, New York. ISBN 0-12-181895-0. Edited by P. Nissen 19‘ Physiol. Plant. 59. 1983 Jen. C. J. & Haug. A. 1981. Potassium-induced depolarization of the transmembrane potential in Blasmcla-Jtella emer- sonii zoospores precedes encystment. - Exp. Cell Res. 131: 79—87. Klee. C. B.. Crouch. T. H. 8: Richman. P. G. 1980. Calmodu- lin. - Annu. Rev. Biochem. 49: 489—515. LaPorte. D. C., Wierman. B. M. & Storm. D. R. 1980. Cal- cium-induced exposure of a hydrophobic surface on cal- modulin. — Biochemistry 19: 3814-3819. Lin. Y. M. 1982. Calmodulin. - Mol. Cell Biochem. 45: 101-1 12. Nagahashi. 6.. Leonard. R. T. & Thompson. W. W. 1978. Purification of plasma membranes from roots of barley: Specificity of the phosphotungstic acid-chromic acid stain. - Plant Physiol. 61: 993—999. Perlin. D. 5. 8L Spanswick. R. M. 1981. Characterization of ATPase activity associated with corn leaf plasma mem- branes. - Plant Physiol. 68: 521-526. Siegel. N. & Haug. A. 1983. Aluminum interaction with cal- modulin: Evidence for altered structure and function from optical and enzymatic studies - Biochim. Biophys. Acta 744: 36-45. Smith, J. C.. Halliday. L. & Topp, M. R. 1981. The behvaior of the fluorescence lifetime and polarization of oxonol po- tential-sensitive extrinsic probes in solution and in beef heart submitochondrial particles. - J. Membr. Biol. 60: 173-185. Viola, R. E... Morrison. J. F. & Cleland. W. W. 1980. Interac- tion of metal (111)-adenosine 5‘-triphosphate complex» with yeast hexokinase. - Biochemistry 19: 3131-3137. Wang. C. S. & Smith. R. L. 1975. Lowry determination of protein in the presence of Triton-X-IOO. - Anal. Biochem. 63: 414-417. Wang. J. H. & Waisman. D. M. 1979. Calmodulin and its role in the second messenger system. - In Current Topics in Cellular Regulation (B. L. Horeckcr and E. R. Stadtman. eds). Vol. 15. pp. 47-107. Academic Press. New York. lSBN 0-12-152815-4. Womack. F. C. & Colowick. S. P. 1979. Proton-dependent inhibition of yeast hexokinase by aluminum in ATP pre- parations. — Proc. Natl. Acad. Sci. USA 76: 5080-5084. 291 Chapter 5 A Thermodynamic and Electron Paramagnetic Resonance Study of Structural Changes in Calmodulin Induced by Aluminum Binding 30 Vol. 115. No. 2, 1983 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS September 15, 1983 Pages 512-517 A THERMODYNAMIC AND ELECTRON PARAMAGNETIC RESONANCE STUDY OF STRUCTURAL CHANGES IN CALMODULIN INDUCED BY ALUMINUM BIN DING Neal Siegel,“2 Richard Coughlin3 and Alfred Haugz Department of Botany and Plant Pathology and 2Pesticide Research Center, ichigan State University, E. Lansing, Michigan, and the Division of Artherosclerosis and Lipoprotein Research, Baylor College of Medicine, The Methodist Hospital, Houston, Texas 1 Received August 2, 1983 Bovine brain calmodulig, binds gamol aluminum per mol protein with dissociation constants in the range of 109’ to 10 molar. EPR spectra of spin—labelled calmodulin provide data indicating that aluminum binding causes decreased probe immobilization as compared to the effects of calcium binding. This result of aluminum binding indicates that —calmodulin is a more random, Open polypeptide relative to the structure of Ca -ca1modu1in. Calorimetric measurements of aluminum binding provide data showing that the first m_ql of aluminum bound is accompanied by the largest enthalpic change (-3. 9 kcal mol , whereas binding of the second and third mol of aluminum are each entropically driven. Calmodulin is an important Ca2+-dependent regulating protein in almost all eukaryotic tissues and organs (1-4). Recently it was suggested that aluminum ions, which are toxic to plants and animals in low concentration, may exert their toxic properties by interacting with calmodulin (5). This protein loses its structural integrity upon the stoichiometric binding of aluminum and ceases to retain the 2+-ealmodulin dependent phosphodiesterase (5) or a Ca”— capacity to regulate Ca calmodulin dependent ATPase in the barley root plasma membrane (6). To further investigate the changes in cal modulin induced by aluminum binding we have analyzed the thermodynamic properties of this process using colorimetric methods and equilibrium dialysis. Correlation times for covalently attached spin probes on the protein were calculated from EPR spectra to assess relative changes in protein structure in response to metal binding. Data are presented showing that three mol of aluminum bind specifically to each mol of calmodulin; binding of the first mol of aluminum bound is enthalpically driven in contrast to the second and third mol Address for correspondence: Dr. A. Haug, Pesticide Research Center, Michigan State University, E. Lansing, MI 48824, U. S.A. 0006-291 X/83 $1. 50 ( ‘unvrtghl " 2 I983 by Academic Press. Inc. All rights of rt'prfllllu‘llvll in any form reserved. 512 31 Vol. US, No. 2. I983 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS bound which are entropically d'iven. When bound by aluminum the protein apparently takes on an open, more random conformation as compared to the effects of calcium binding. MATERIALS AND METHODS Sources: Bovine brain acetone powder, Tris, from Sigma Chemical Co. (St. Louis, MOP, Affigel-Phenothiazine from Bio-Rad Laboratories (Richmond, CA); AlCl , CaCl from Mallinckrodt Science Products (St. Louis, MO). 3-[a-iodoacetamidof- 2,2,5,%-tetramethyl-l-pyrrolidinyl oxyl from Syva Corp. (Palo Alto, CA). All other chemicals were of the highest purity that were commercially available. Methods: Calmodulin was isolated from bovine brain acetone powder and preparm metal-free as previously described (5). Protein concentrations were determined spectrophotometrically (8). Spin labelling of calmodulin was accomplished using the method of Hewgley and Puett (9). Microcalorimetric data were collected using a LKB 210 batch microcalorimeter equipped with a pair of gold mixing cells and were corrected for heats of mixing as described (11). The experimental temperature was maintained at 23.85 i .0100. Equilibrium dialysis experiments were conducted and aluminum analyzed as previously described (5). Analysis of the EPR signal from a Varian X-band EPR spectrometer E-112 using a Varian 620/L-100 computer showed that 1.34 spin labels were bound per protein molecule. Correlation times (Tc) were calculated from EPR spectra using the following relationship (10): 1 - - 0 1/2 We - 6.5 x 10 wo [(hO/h_1) -1] see where we is the midline width (gauss), and h and h are the peak heights of the mid- and high-field lines, respectively. Limitatioss of this equation are noted as described by Melhorn et a1 (10) and the calculated values are applied as standard quantitative measures for comparison of spectra. RESULTS AND DISCUSSION 2+ addition to calmodulin As shown in Fig. l, the calculated values of 1c for Ca increased 6% beyond the value for the metal-free protein and saturated between a ratio of 4 and 5 mol calcium per mol calmodulin. This change in Tc indicates increased immobilization of the spin label and is comistent with compaction of the protein upon binding of calcium (1-4). In contrast, as a function of increasing 1 c calculated for increasing amounts of aluminum decreased the immobilization of the spin label indicating increased randomness of the polypeptide region near the spin probe. These results are consistent with other aluminum-induced structural changes in calmodulin (5) including decreased helical content, increased random ceiling and an increased hydrophobic surface expression. Calcium binding has been shown to contrast these changes; the helical content increases, random coiling decreases and there is a small increase of hydmphobic surface area (1-4). 513 32 Vol. IIS, No. 2. I983 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS P—_ 'I_—“'T 7““ I “I—_ W 1"" ‘T- "‘T "— 5.3 T1 1', (x IO“°)Seconds .5 .5 0' 0' m (D o ' - :5 \I .b as 4.5 1 l L l I o I 2 3 4 5 e [METAL]/[CALMODULIN] ‘Jr— m»— 51.1- Changes in spin-probe mobility of covalently labelled calmodulin due to metal addition. Values for T c (correlation times) were calculated as described in MATERIALS AND METHODS. Changes due to aluminum (0) or calcium (0) addition to II) uM calmodulin at pH 6.5 are shown. Values are shown for at least two separate trials. For the purpose of the present discussion we refer to the various species of aluminum present in solution collectively as aluminum. Under the conditions used in the present study we can state that mononuclear, hydrated aluminum species are present, as opposed to polynuclear species existing at higher pH values and elevated aluminum concentrations. We cannot be certain, however, of the charge on these hydrated aluminum species (7). Thermodynamic functions associated with aluminum binding to calmodulin are summarized in Table 1. Binding constants were calculated from the equilibrium dialysis data presented in Pig. 2. The enthalpic contribution for the first mol of aluminum bound is -3.9 keel/mol, both opposite in sign and greater in magnitude than that for the next two mol aluminum bound. The calculated entropic contribution increased during the binding of the second and third mol aluminum bound; the total entropic contribution is 103.8 e.u.; only 21.9 e.u. are contributed by the binding of the first aluminum mol bound. These results differ from the effect of Ca2+ binding to 514 33 Vol. IIS, No. 2. I983 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS Table l. Thermodynamic parameters associated with the binding of aluminum to bovine brain calmodulin. Aluminum Binding Site(i) KMi” A Gi A 0‘: A if: Asi Ag; Kcal mol-1 cal deg"x mol"1 1 1.2 x10_6 -9.o -1o.4 -3.9 13.9 21.9 2 9.9 x 10'7 -9.2 -1o.s +1.2 31.4 39.4 3 1.3 x 10'7 -9.4 -11.9 +0.9 34.5 42.5 2 130i was calculated from the relationship: I. Gi = -RTanA1. = if; - msi where KAI. entrbpy, were calculated from: 115‘: = £3, + 7.98 (11, 16). The unitary free energy chmge, Adi, was then calculated as :0? = If: - TS: for T = 298°K at pH 5.5. " KAI. were calculated from equilibrium dialysis data. 1 is expressed in terms of molar concentrations. ’, $2, the unitary changes in Troponin C, a calcium binding protein similar in nature to calmodulin (11). In that study of mponin C, calcium binding to the four available sites on the protein have enthalpic contributions of -7.7 kcal mol.1 for each site and the entropic contribution [ALUMINUM] BOUND /[CALMODULIN] 7 65 "6.0 5.5 5.0 4.5 4.0 3.5 'loq [ALUMINUM] FREE _F_ig_2. Binding of aluminum to calmodulin. One-ml sample volumes of 10 11M calmodulin were dialyzed against various concentrations of aluminum chloride at pH 6.5. Aliquots were analyzed (5) for aluminum content both inside and outside of the dialysis bag after 24 h. SIS 8‘ -\\ 34 Vol. "5, No. 2, I983 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS decreases from 14.7 e.u. for the first two sites to 8.0 e.u. for the third and fourth sites. In the case of Troponin C, Ca2+ binding is in part an enthalpically driven process. The results of our experiments complement the spectroscopic work which detailed structural alterations of calmodulin induced by aluminum binding (5). Explanation of the olnerved thermodynamic changes during aluminum binding to the protein comes from the unique hydration properties of this metal. Both aluminum and calcium have a primary coordination number of six representing the effective hydration number of the innermost hydration shell (12). It has been discussed that only a single water molecule exists at the high affinity Ca2+-binding site of Troponin C (13) 2"’-binding sites of calmodulin (14). Therefore, the binding of calcium by or the Ca these proteim is accompanied by the release of five coordinated water molecules and an entropy change of about 15 e.u. (11). The strength of the ion/water coordination bonds is thee-fold higher for aluminum than calcium: i.e., 1144 keel/mol Al3+ vs. 399 kcal/mol Ca2+ (15). Also, the value of the formal charge over ionic radius and the intermediate electronegativity are both a factor of 2 higher for aluminum relative to calcium (14), and the number of water molecules in the outer hydration shells of aluminum is significantly higher than for calcium (12). Together, these physical data support the conclusion that aluminum ions differ appreciably from calcium ions in their interaction with calmodulin. Under our experimental conditions, mononuclear, hydrated aluminum species interact with calmodan and promote an open, solvent- rich, disordered polypeptide region, effects which contrast those of calcium binding to this protein. The observed enthalpy change for the first mol aluminum bomd is composed of two terms, A H for hydrogen bond breakage and A H for solvation. Since the A H for hydrogen bond breakage is positive, the measured enthalpy change must have as a major contribution an enthalpic change associated with increased salvation of the protein upon binding the first mol of aluminum. As discussed above, interaction of calmodulin with the highly hydrated aluminum ions accounts for the observed enthalpy change. 516 {C 35 Vol. IIS, No. 2, I983 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS The results of this study indicate that the aluminum calmodulin interaction seems to rauIt from the unique hydration properties of this metal. This type of interaction may serve to explain the potent, biologically toxic properties of aluminum. Acknowlecggment: We thank Dr. Henry J. Pownall for his helpful comments. Partial funding was from the Welch Foundation Grant Q906. Acknowledgement is also given to the Michigan Agricultural Experiment Station (Journal Article Number 10927). REFERENCES 1. Wang, J. H. and Waisman. D. M. (1979) Curr. Top. Cell Regl. 15, 47-107 2. Klee, C. B., Crouch, T. H. and Richman. P. G. (1980) Annu. Rev. Biochem. 49, 489-515 3. Lin, Y. M. (1982) Mol. and Cell. Biochem. 45, 101-112 4. Cormier, M. J., Anderson, J. M., Charbonneau, H., Jones, H. P. and McCann, R 0. (1980) in Calcium and Cell Function (Cheung, W. Y., ed.) Vol. 1, pp. 201-218, Academic Press, New York 5. Siegel, N. and Haug, A. (1983) Biochim. Biophys. Acta 744, 36-45 6. Siegel, N. and Haug, A. (1983) Physiol. Plant. in press 7. Bees, C. F. and Mesmer, R. E. (1976) in The Hydrolysis of Cations, pp. 112-113, J. Wiley and Sons, New York 8. Kilhoffer, M. C., Demaille, J. G. and Gerard, D. (1980) FEBS Lett. 116, 269-272 9. Hewgley, P. B. and Puett, D. (1980) Ann. New York Acad. Science 356, 20-31 10. Mehlhorn, R., Swanson, M., Packer, L. and Smith, P. (1980) Arch. Biochem. Biophys. 204, 471-476 11. Potter, J. D., Hsu, F. J. and Pownall, H. J. (1977) J. Biol. Chem. 252, 2452-2454 12. Hindman, J. C. and Sullivan, J. C. (1971) in Coordination Chemistry (Martell, A. 13., ed.) Vol 1, pp. 393-426, Van Nostrand Reinhold Co., New York 13. Wang, C. L. A., Leavis, P. C., Harrocks, W. D. W. and Gergeley, J. (1981) Biochemistry 20, 2439-2444 14. Wang, C. L. A., Aquaron, R. R.. Leavis, P. C. and Gergeley, J. (1982) Eur. J. Biochem. 124, 7-12 15. Noller, M. (1982) Acta Chim. Acad. Sci. Hing. 109, 429-448 16. Kauzmann, W. (1959) Adv. Protein Chem. 14, 1-83 5” Chapter 6 Summary 37 The results presented in chapters 2 and 3 show that aluminum ions interact specifically with the calmodulin molecule. Three mol of aluminum are bound per mol of cal- modulin. Thermodynamic data regarding this binding as presented in chapter 5 indicate that the binding is tight with Kd's in the range of 0.1 to I UM. Calorimetry indicates that entropic changes accompany aluminum binding, a result that is complementary to spectrosc0pic data showing the expression of a large hydrophobic surface and a helix- coil transition within the protein that also accompany alum- inum binding. In the most fundamental sense, calmodulin is denatured upon binding aluminum, even with calcium present. The reason that calmodulin loses its structural inte- grity through binding aluminum rests in the unique physica- chemical characteristics of the highly solvated aluminum ion. Both aluminum and calcium have similar binding affinities for calmodulin and both have an inner hydration shell composed of six water molecules (Hindman and Sullivan, 1971). Five of these water molecules coordinated to calcium are released upon its binding to Troponin C, a calcium binding protein that shares sequence homology in the calcium binding loops with calmodulin (Wang et al, 1981). The strength of these coordinated bonds is appreciably higher for aluminum than for calcium; 1144 Kcal/mole as compared with 399 Kcal/mole, respectively (Matheja and Degens, 1971). Additionally, the ratio of the formal charge to ionic radius and the intermediate electronegativity are both a factor of 38 two higher for aluminum than for calcium (Noller, 1982). These physical data lead to the expectation that aluminum should interact differently with calmodulin than should calcium, and this conclusion is supported by data presented in chapters 2,3 and 5 showing that aluminum promotes dissintegration of the stabilizing forces involved with maintaining the structural integrity of the protein. The specific amino acid residues to which aluminum becomes bound are presently not known. As far as calcium is concerned, four specific binding 100ps exist in the protein and this metal appears to be coordinated to oxygen ligands with its mobility becoming restricted when bound (Andersson et al, 1982, Krebs, 1981). Aluminum also forms stable com- plexes wtih electronegative ions such as oxygen and nitro- gen. In contrast with calcium, the higher electronegativity of aluminum would result in an increased covalent character of the coordinated bond established, a conclusion which is borne out by the data indicating that the binding affinity for calnodulin is 5 to 10 fold higher for aluminum than for calcium as presented in chapters 2, 3, and 5. The effects exerted by aluminum on calmodulin must extend to the intramolecular bonds that maintain this protein's structure in order to cause the observed physical alterations. Similarly, the interaction of calcium with calmodulin must affect intramolecular bonds with effects Opposite to those caused by the binding of aluminum. It is thought that upon binding calcium, negative charges on the 39 protein are neutralized. In turn, weakened constraining forces result permitting the formation of additional helical elements. Ultimately the calcium-calmodulin complex becomes a more compact structure than the metal-free calmodulin (Seamon, I980). The interaction of aluminum is more complex particularly because of the highly hydrated character of this ion as previously discussed. The results presented in chapters 2 and 3 show that aluminum binding leads to a helix-coil transition that is accompanied by a strong enhancement of the protein's hydr0phobic surface domains. Simply stated, the hydrated aluminum ions promote open, solvent-rich, disordered polypeptide regions contrasting the effects of calcium. The biological consequences of aluminum-induced alter- ations in calmodulin's structure are described in chapters 3 and 4. Calcium-calmodulin stimulation was lost in the presence of aluminum for phosphodiesterase activity and for the electrogenic ATPase activity in barley root plasma membrane vesicles. As a result of the loss in calmodulin's flexibility when aluminum is bound, its ability to correctly interface with and regulate various target proteins are impaired. Thus, aluminum-calmodulin complex represents a primary biochemical lesion in aluminum toxicity. Con- sidering the pivotal role of calmodulin in calcium regu- lation, aluminum interference is expected to result in severe imbalances of cellular processes of plant cells such as maintenace of the membrane potential, cell growth, root 4O elongation and chloroplast function. This is consistent with observations that aluminum-induced toxic responses of plants resemble, in part, those occuring as a result of calcium deficiency. The data presented in this dissertation represent the first reports of a specific biochemical target for aluminum under conditions of physiological relevence. This is not to say that aluminum interaction with calmodulin is the only toxic interaction that involves aluminum. However, this study does serve to elucidate a model on which other toxic aluminum interactions within the living organism may be based. The recent report that metal toxicity might be correlated with calmodulin inhibition (Cox and Harrison Jr., I983) adds support to this model. The path leading to the discovery that aluminum attach- es strongly to calmodulin and thus may exert toxic effects in the cell was less than straightforward. At first, inhi- bition of the barley root plasma membrane ATPase via com- plexation of ATP by aluminum was investigated. There was a precedent for this in that yeast hexokinase is inhibited by Al-ATP (Viola et al, 1980). Under the conditions in which the ATPase activity of the isolated vesicles was assayed (Chapter 4), aluminum interfered with the colorimetric assay for released phosphate which necessitated a better way for monitoring the enzymatic activity. The approach used assumed that the ATPase activity was electrogenic, that is, the ATPase acts as an ionic pump at the plasma membrane. 41 Fluorescent, potential-sensitive probes were chosen to assay this system of sealed vesicles because these vesicles were shown in preliminary studies not to be influenced by aluminum or other metal addition. Optimization of the system as reported in chapter 4 involved manipulation of the divalent metal concentrations and included calmodulin, previously shown to enhance the enzymatic activity of the ATPase (Caldwell and Haug, 1980). Addition of aluminum to the optimized system consistently lowered the electrogenic activity, but not below a basal value established in the absence of calmodulin (Chapter 4). This was the first indication that the inter- action of aluminum involved calmodulin. Introductory studies of the effect of aluminum on isolated calmodulion indicated that the protein's hydrophobic surface expression was greater when aluminum interacted with it than when calcium interacted with it. Only as more data were col- lected indicating that calmodulin became denatured upon the specific binding of aluminum did it become clear that the interaction may represent a primary biochemical lesion during aluminum intoxication. Once the specificity and potency of aluminum interaction with calmodulin became established, it was logical to assume that because calmo- dulin is ubiquitous in eukaryotic cells, and because plants and animals including man suffer adverse reaction to aluminum exposure, the common link was calmodulin. 42 It is known in some plants that aluminum tolerance is genetically linked (Foy et al, 1978). Because the aluminum- calmodulin complex represents a primary lesion in the toxic responses of plants to this metal, future research must attempt to discover what protective mechanisms are available to those genetically resistant plant varieties. These defenses might include enhanced levels of aluminum-specific chelators that can sequester aluminum away from any inter- action with calmodulin (Suhayda and Haug, 1983), amino acid substitutions in calmodulin that maintain structural stabil- ity after aluminum is bound or extracellular compounds that prevent aluminum entry into the cell. When the protective processes that prevent aluminum-calmodulin interaction are elucidated, therapeutic methods of relieving this stress should not be far behind. It is hoped that this study is a first step toward that end. BIBLIOGRAPHY 44 Anderson, T., Drakenbsrg, T., Forsen, S., Thulin, E. and Sward, M. 1982 J. Amer. Chem. Soc. 104:576-580 Baes, C.F. and Mesmer, R.E. 1976 The Hydrolysis of Cations. J. Wiley and Sons, New York pp 112-123 ISBN 0-471-03985-3 Beeler, T.J., Farmen, R.H. and Martonosi, A.N. 1981 J. Caldwell, C.R. and Haug, A. 1980 Physiol. Plant. 50: 183-193 Caldwell, C.R. and Haug, A. 1981a Plant Physiol. 67 (Suppl):136 Caldwell, C.R. and Haug, A. 1981b Anal. Biochem. 116: 325-330 Chen, Y.H.. Yang, J.T. and Martinez, H.M. 1972 Biochem- istry 11:4120-4131 Cheung, W.Y. 1970 Biochem. Biophys. Res. Comm. 38: 355-537 Clarkson, D.T., and Hanson, J.B. 1980 Annu. Rev. Plant Physiol. 31:239-298 Cormier, M.J., Anderson, J.M., Charbonneau, H., Jones, H.P. and McCann, R.D. 1980 In. Calcium and Cell Function (W.Y. Choung, ed.), Vol 1. pp 201-218 Academic Press New York ISBN 0-12-171401-2 Cornish-Bowden, A. 1979 Fundamentals of Enzyme Kinetics pp 147-176, Butterworths, London Crapper, D.R., Karlik, S. and DeBoni, U. 1978 In: Alzheimer's Disease: Senile Dementia and Related Disorders (Katzman, R., Terry, R.D. and Bick, K.L., eds.), Aging, Vol 7, pp 471-485, Raven Press, New York Crapper-McLachlan, D.R. 1980 Neurotox. 1:3 Crouch, T.H. and Klee, C.B. 1980 Biochemistry 19:2692- 3698 Dedman, J.R., Potter, J.D., Jackson, R.L., Johnson, J.D. and Means, A.R. 1977 J. Biol. Chem. 252:8415-8422 Dieter, P. and Marme, D. 1980 Proc. Natl. Acad. Sci. USA 77:7311-7314 45 Dieter, P. and Marme, D. 1981 FEBS Lett. 125:245-248 Foy, C.D., Chaney, R.L. and White, M.C. 1978 Annu. Rev. Plant Physiol. 29:511-566 Goldberg, N.D. and Haddox, M.K. 1977 Annu. Rev. Biochem. 46:823-896 Greenfield, N. and Fasman, G.D. 1969 Biochem. 8:4108- 4116 Hampp, R. and Schnabl, H. 1975 Z. Pflanzenphysiol. 76:300-306 Hecht-Buchholz, C. and Foy, C.D. 1981 Plant Soil 63: 93-95 Hewgley, D.B. and Puett, D. 1980 Ann. New York Acad. Science 356:20-31 Hindman, J.C. and Sullivan, J.C. 1971 in Coordination Chemistry (Martell, A.E., ed.), Vol 1., pp 393-426, Van Nostrand Reinhold Co.. New York Hodges, T.K. and Leonard, R.T. 1974 in Methods in Enzymology (Fleischer, S. and Packer, L., eds.), Vol 32B pp 392-406, Academic Press, New York ISBN 0-12-181895-0 Jen, C.J. and Haug, A. 1981 Exp. Cell Res. 131:7987 Karplus, M. and McCammon, J.A. 1981 CRC Crit. Rev. Biochem. 11:293-349 Kauzmann, W. 1959 Adv. Protein Chem. 14:1-63 Kilhoffer, M.C., Demaille, J.C. and Gerard, D. 1980 FEBS Lett. 116:269-272 Klee, C.B. 1980 in Calcium and Cell Function (Cheung, W.Y., ed.), Vol 1., pp 59-77, Academic Press, New York Klee, C.B., Crouch, T.H. and Richman, P.G. 1980 Annu. Rev. Biochem. 49:489-515 Krebs, J. 1981 Cell Calcium 2:295-311 Lambert, J.B., Shurwell, H.P., Verbit, L., Cooks, R.G. and Stout, G.H. 1976 Organic Structural Analysis, p. 341, MacMillan, New York 46 LaPort, D.C., Wierman, B.M. and Storm, D.R. 1980 Biochemistry 19:3814-3819 Lin, Y.M. 1982 Mol. Cell. Biochem. 45:101-112 Lowry, O.H., Rosebrough, N.J., Farr, A.L. and Randall, R.J. 1951 J. Biol. Chem. 193:265-275 Malencik, D.A. and Anderson, S.R. 1982 Biochemistry 21: 3480-3486 Martell, A.E. and Smith, R.M. 1974 Critical Stability Constants, Vol 1., p 271, Plenum Press, New York Matheja, J. and Degens, E.T. 1971 in Structural Molecular Biology of Phosphotes, p 79, Gustav Fischer Vorlag, stuttgart ISBN 3-437-30109-8 Merril, C.R., Goldman, D., Sedman, S.A. and Ebert, M.H. 1981 Science 211:1437-1438 Mehlhorn, R., Swanson, M., Packer, L. and Smith, P. 1980 Arch. Biochem. BiOphys. 204:471-476 Muto, S. and Miyachi, S. 1977 Plant Physiology 49:55-60 Nagahashi, G., Leonard, R.T. and Thompson, W.W. 1978 Plant Physiol. 61:993-999 Nightingale, E.R. 1959. J. Phys. Chem. 63:1381-1387 Noller, H. 1982 Acta Chim. Acad. Sci. Hong. 109:429-448 Perlin, D.S. and Spanswick, R.M. 1981 Plant Physiol. 68:521-526 Potter, J.D., Hsu, F.J. and Pownall, H.J. 1977 J. Biol. Chem. 252:2452-2454 Rasmussen, H., Goodman, D.B.P. and Tenenhouse, A. 1972 CRC Crit. Rev. Biochem. 1:95-148 Schulman, H., Huttner, W.B. and Greengard, P. 1980 in Calcium and Cell Function (Cheung, W.Y., ed.), Vol 1. pp 219-250, Academic Press, New York Seamon, K.B. 1980 Biochemistry 19:207-215 Siegel, N., Coughlin, R. and Haug, A. 1983 Biochem. Biophys. Res. Comm. 115:512-517. Siegel, N. and Haug, A. 1983 Biochem. Biophys. Acta 744:36-45 47 Siegel, N. and Haug, A. 1983 Physiol. Plant. 59:285-291. Siegel, N., Suhayda, C. and Haug, A. 1982 Physiol. Chem Phys. 14:165-167 Smith, J.C., Halliday, L. and Topp, M.R. 1981 J. Membr. Biol. 60:173-185 Smith, M.B. Niemeczura, W.P., Murtaugh, T.J., Siegel, F.L. and Gibbons, W.A. 1981 Fed. Proc. 40:1871 Smoake, J.A., Song, S.Y. and Cheung, W.Y. 1974 Biochem. Biophys. Acta 341:402-411 Suhayda, C. and Haug, A. 1983 Plant Physiology 72:136 Viola, R.E., Morrison, J.F. and Cleland, W.W. 1980 Biochemistry 19:3131-3137 Wallace, R.W., Tallant, E.A. Dockter, M.B. and Cheung, W.Y. 1982 J. Biol. Chem. 257:1845-1854 Wang, C.L.A., Aquaron, R.R., Leavis, P.G. and Gergeley, J. 1982 Eur. J. Biochem. 124:7-12 Wang, C.L.A., Leavis, P.G. Harrocks, W.D.W. and Gregeley, J. 1981 Biochemistry 20:2439-2444 Wang, C.S. and Smith, R.L. 1975 Anal. Biochem. 63: 414-417 Wang, J.H. and Waisman, D.M. 1979 in Current Topics in Cellular Regulation (Horecker, B.L., and Stadtman, E.R., eds.), Vol 15., pp 47-107, Academic Press, New York, ESBN 0-12-152815-4 White, H., Tudor, M., Bose, K. and Markley, D. 1982 BiOphys. J. 37:51a Wolff, D.J. and Brostrom, C.O. 1979 Adv. Cyclic Nucleotide Res. 11:28-88 Wolff, D.J., Poirier, P.G., Brostrom, C.O. and Brostrom, M.A. 1977 J. Biol. Chem 252:4108-4118 Womack, F.C. and Colowick, S.P. 1979 Biochemistry 76: 5080-5084 APPENDIX 1 49 STRUCTURAL DISTINCTIONS BETWEEN BARLEY AND BRAIN CALMODULIN: A STUDY OF STRUCTURAL CHANGES ACCOMPANING CALCIUM AND ALUMINUM INTERACTION WITH BARLEY CALMODULIN Neal Siegel1 and Alfred Haug2 1Department of Botany and Plant Pathology and 2Pesticide Research Center Michigan State University, E. Lansing, Michigan Calmodulin isolated from barley, like that from other plant sources contains one tyrosine and one cysteine residue whereas calmodulin isolated from animal sources contains two tyrosine residues and does not contain cysteine. Other dif- ferences in amino acid content are reflected in decreased (whelical content of barley as compared to bovine brain cal- modulin. Application of calcium increases the protein's helical content and aluminum application decreases the heli- cal content, each to a similar extent regardless of protein source. Hydrophobic surface exposure is also enhanced upon metal interaction with calmodulin when monitored with 8-ani- lino-I-naphthalene sulfonic acid. These results indicate that aluminum-calmodulin represents a primary biochemical lesion of aluminum intoxication in both plants and animals. INTRODUCTION Aluminum interaction with calmodulin has recently been focused upon as a primary biochemical lesion that occurs during the toxic response of eukaryotic cells to this metal (1-3). This research has used bovine brain calmodulin as a model system due to the conserved sequence of this ubiqui- tous calcium-dependent regulation protein (4). Differences exist, however, in the amino acid composition between plant calmodulin and animal calmodulin, most notably in the occurence of a cysteine residue in the plant protein that is 50 not found in the animal protein (5). Other differences include a higher amount of proline in the plant protein and one instead of two tyrosine residues as found in animal calmodulin. Because calmodulin is known to undergo struc- tural alterations upon binding metals including aluminum and calcium (1,2,6), it is important to understand how composi- tional differences between plant and animal calmodulin affect the structural response of each to metal binding. Changes in the barley protein's circulr dichroism spectrum in the ultraviolent wavelength region upon metal binding show that calcium induces a coil-helix transition; these results are qualitatively identical with results of a similar study utilizing bovine brain calmodulin (1). More hydrophobic surface exposure occurs on the barley protein when aluminum application is compared with calcium application; these results are also qualitatively similar to results using bovine brain calmodulin. ElectrOphoretic analysis using denaturing polyacrylimide gel electrophoresis (PAGE) shows that the molecular weight of barley calmodulin is virtually identical to bovine brain calmodulin whereas analysis using non-denaturing PAGE and protein bound with calcium or aluminum shows that the electrophoretic mobility is dependent on the type of metal bound. MATERIALS Sources: Tris, from Sigma Chemical Co. (St Louis, MO), AffigeI-Phenothiazine and SDS-PAGE Low Molecular Weight 51 Standards from Bio-Rad Laboratories (Richmond, CA), AlCl3 and Ca012 from Mallinckrodt Science Products (St. Louis, MO), sodium salt of ANS from K&K Laboratories (Plainview, N.Y.). All other chemicals were of the highest quality available. Methods: An acetone powder from barley (Hordeum Vulgare var. Conquest) was prepared by excising shoots of plants grown at pH 6.5 in 250 uM Ca304 possessing intact coleoptiles, freeze-drying these and washing the resulting powder with acetone to remove soluble pigments. The acetone powder was stored at 0°C. The procedure for isolating calmodulin from this material was the same as that described for bovine brain acetone powder (1). 33 g of starting material yielded 14 mg of calmodulin which ran as a single band using SDS-PAGE and the silver stain technique (7). Circular dichroism measurements were made using a Jasco spectropolarimeter, model 0RD/UV/CD-5 as previously described (1). Fluorescence measurements were made as previously described using a SLM 4000 spectrofluorimeter; each point was signal-averaged over 10 scans, 10 times to provide data points falling within 5% standard error of the mean value. Amino acid analysis was performed using a modified Dionex Amino Acid Analyzer with a DCSA microcolumn. ElectrOphoresis was done on 15% acrylimide gels; denaturing gels included 0.4% SDS. Abbreviations: ANS, 8-anilino-1-naphthalene sulfonic acid; SDS, sodium dodecyl sulfate; CD, circular dichroism; Tris, Tris-(hydroxymethyl)-aminomethane. 52 RESULTS AND DISCUSSION As shown in Fig. 1, the electrophoretic mobility of bovine brain caomodulin and barley calmodulin under denaturing conditions shows calculation that the molecular weights for each lie between 16,500 and 17,300 agreeing with published reports (4,5). ElectrOphoresis under non-denatur- ing conditions as shown in Fig. 2 indicate that structural changes induced by metal binding with calmodulin result in differing mobilities dependent on whether the protein is metal free, bound with calcium, or bound with aluminum. Barley calmodulin is less electrOphoretically mobile than brain calmodulin under these conditions regardless of pre- treatment indicating that structural differences exist between the two molecules; however, charge differences under these conditions become a major factor and may be responsi- ble to some extent for the different mobilities of the metal-protein complexes. Metal-free bovine braine calmodulin has a helical content of 37% (1) calculated using [6]222 from CD spectra and applying the method of Chen et al. (8). Under identical conditions, metal-free barley calmodulin has a helical content of 26% (Tab. 1, Fig. 3). Application of calcium caused increased helical formation for both barley and brain calmodulin with net helix content increases of 12% and 18%, respectively at saturation. Aluminum application caused helical loss for both barley and brain calmodulin with net helix losses of 42% and 40%, respectively. The physical 53 —14 Fig. I. Electrophoresis of calmodulin on 15% acrylimide gels in the presence of SDS. 10 ng of barley calmodulin (lane 1) and bovine brain calmodulin (lane 2) were run with SDS-PAGE molecular weight standards (lane 3) Numbers next to lane 3 indicate standard molecular weights. Proteins were visualized using the silver stain technique (7). 54 Fig. 2. Electrophoresis of calmodulin on 15% acrylimide gels under non-denaturing conditions. 10 pM stock solutions of calmodulin from bovine brain or barley were prepared in metal free buffer or in buffer containing 100,pM A1013 or CaClz. 10 ng of each type of treated calmodulin were electrophoresed. Bovine brain calmodulin in metal free, calcium or aluminum treated buffers are in lanes 1,3, and 5, respectively. Barley calmodulin in metal free, calcium or aluminum treated buffers are in lanes 2,4 and 6, respec- tively. Proteins were visualized as described in Fig. 1. 55 Table 1. Changes in [6]222 and helical content of calmodu- lin derived from bovine brain and barley induced by metal binding [ijgg (mdeg cm /dmole) % a-Helix Barléy' Brain Barley Brainb -Metal -10,182 -13,571 26 37 Casat -11,283 -16,100 29 (+12) 44 (+18) Alsat - 7,086 -10,280 15 (-42) 22 (-40) a numbers in parentheses indicate percent change from initial helical percent b data from Siegel and Haug (1983a) Results were calculated by the method of Chen et al. (8) 56 % (X-HELIX. I 2 3 4 5 6 7 [METAL] /[CALMODULIN] Fig. 3. Changes in the helical content of 10 uM barley calmodulin titrated with CaClz (e), AlC13 (O) or A1013 in the presence of 70 ”M CaClz (D). All procedures were done in 10 mM Tris, pH 6.5. 57 reasons for these structural changes have been previously discussed for aluminum interaction with bovine brain calmodulin (1), and the statement that highly hydrated, aluminum ions promote Open, solvent-rich, random polypeptide regions in barley calmodulin applies. The structural changes induced by aluminum binding to bovine brain calmodulin was the same in the absence of calcium or in the presence of saturating amounts of calcium (1). Barley calmodulin, however, had more helix remaining if calcium was present at saturating amounts that when it was absent (Fig. 3). There was a net helical loss of 30% when aluminum was applied with calcium present; this effect saturated at a value of 21.3% as compared with a saturating value without calcium present of 15%. Differences in the amino acid composition between the two types of calmodulin are probably involved and are reflected in increased structural stabilization of the barley protein as compared to the animal protein when bound with calcium. Hydrophobic surfaces on the calmodulin molecule are expressed in the presence of calcium and are though to provide the interface for regulating target protein activity (9). Aluminum application greatly enhanced their exposure in bovine brain calmodulin (1,2) and the aluminum induced change was insensitive to the presence of calcium. Fig. 4 shows that aluminum application increased the hydrOphobic surface exposure relative to that of calcium application; exposure of this surface by aluminum was approximately 58 2.6 I- 2J4 1L2 2A) ,4." 1.8 1.6 ‘04 . . . 0 e o 1.2 1.0 3 4 5 «6 7 [METAL] ZJCALMODULI N] Fig. 4. Changes in ANS fluorescence of 10 pM barley calmodulin titrated with CaClz (e), A1Cl3 (O), or AlCl3 in the presence of 70 UM Ca012 (D). Ordinate values are treatment fluorescence value, F, divided by the initial fluorescence value for the individual treatment, Fo.l All procedures were done in 10 mM Pipes buffer, pH 6.5, 2 uM ANS. Excitation and Emission monochrometers were set at 360 1 4mm and 490 : 4nm, respectively. 59 double in the presence of calcium than in the absence of calcium. Similarly to the CD data, structural differences reflecting amino acid differences between the animal and plant proteins are probably involved. Amino acid composition of barley and brain calmodulin are presented in Table 2. In light of the decreased helical content observed for barley calmodulin as compared to brain calmodulin, it is important to note that cysteine and proline whcih destabilize the helical structure of proteins are present in higher amounts in the plant protein than in the animal protein. Cysteine has been shown to occur only in plant calmodulin (5). Thus, the observation that the plant protein has less helical content than the animal protein is due to the presence of these helix destabilizing amino acid residues. The results of this study indicate that the interaction of aluminum with calmodulin is as potent whether the protein source is plant or animal. Barley calmodulin which possesses more helix-destabilizing amino acid residues than bovine brain calmodulin has less a-helical content than bovine brain calmodulin, but undergoes equivalent structural changes in the presence of calcium or aluminum. Structural stabilization of the calcium-calmodulin complex from barley as compared with bovine brain may result from the different amino acid composition. An aluminum-calmodulin complex has been hypothesized to be a key lesion in the toxic responses of plants to this metal. It is known that some species of 60 Table 2. Amino acid content of calmodulin Amino Acid Mole Percent barley, bovine braina ASP 16.0 15.5 THR 6.7 8.1 SER 5.5 2.7 GLU 16.1 18.2 GLY 9.1 7.4 ALA 7.0 7.4 VAL 7.5 4.7 MET 2.4 6.1 ILU 5.0 5.4 LEU 8.0 6.1 TYR 0.6 0.6 PHE 4.8 5.4 HIS 0.8 0.6 LYS 4.8 4.7 ARG 2.0 4.0 PRO 2.3 1.3 CYS 0.6b 0.6 afrom Cormier et al (1980) not directly determined 61 barley posses a genetic tolerance to aluminum intoxication (10). Other amino acid substitutions in the calmodulin molecule may further stabilize the protein against aluminum induced structural changes and thus may be the basis for aluminum tolerance. 10. 62 REFERENCES Siegel, N. and Haug, A. (1983) Biochim. BiOphys. Acta 744: 36-45 Siegel, N. and Haug, A. (1983) Physiol. Plant. 59: 285-291 Siegel, N., Coughlin, R., and Haug, A. BiOphys. Res. Comm. 115: 512-517 Klee, C.B., Crouch, T.H. and Richman, P.G. (1980) Annu. Rev. Biochem. 49: 484-515 Cormier, M.J., Anderson, J.M., Charbonneau, H. Jones, H.P. and McCann, R.0. (1980) in Calcium and Cell Function (Cheung, W.Y., ed.) Vol. 1, pp. 201-218, Academic Press, New York Wolff, D.J., Poirrier, P.G., Brostrom, 0.0. and Brostrom, M.A. (1977) J. Biol. Chem. 252: 4108-4118 Merril, C.R., Goldman, D., Sedman, S.A. and Ebert, M.B. (1981) Science 211: 1437-1438 Chen, Y.H., Yang, J.T. and Martinez, H.M. (1972) Biochemistry 11: 4120-4131 LaPorte, D.C., Wierman, B.M. and Strom, D.R. (1980) Biochemistry 19: 3814-3819 Foy, C.D., Chaney, Chaney, R.L. and White, M.C. (1978) Annu. Rev. Plant Physiol. 29: 511-566 "11111111111111“