2 PhysicaI Chemist ry of Proteins By A. GAFNI Department of Chemical Physics The Weizmann Institute of Science Rehovot Israel 1 Introduction In this chapter recent developments in several areas of the physical chemistry of proteins will be reviewed with special emphasis on optical spectroscopic techniques. These techniques and their applications continue to develop very rapidly; hence only a small fraction of the total work done can fit within the scope of this review. The topics to be covered deal with (a)folding processes in proteins (b) the application of optical activity techniques in studies of protein structure and association (c) the environment of aromatic amino-acids in proteins and (d) the determination of distances in proteins by energy-transfer measurements.2 Protein Folding Unfolding and Denaturation The hypothesis that all the information needed to form the three-dimensional structure of a protein molecule is contained in the sequence of its amino-acids has been suggested many years ago.”* The mechanism by which the folding process takes place still is the goal of many studies. Based on both equilibrium and kinetic studies it has been suggested that the reversible denaturation of proteins involves essentially two major stable conformational states native and denatured; i.e. to be a highly co-operative pro~ess.~ However an ever increasing number of studies have shown that many protein-folding and -unfolding reactions do not follow a simple two-state mechani~m,~ and stable intermediate states were identified along the reaction pathway.’ In some cases the unfolding mechanism was found to depend on the denaturing conditions.Thus the heat denaturation of cy -1actalbumin at neutral pH is described as a two-state transition6 while the unfolding by guanidine-HC1 M. Sela F. H. White and C. B. Anfinsen Science 1957 125 691. C.B. Anfinsen Science 1973,181,223. (a)C. Tanford Adv. Protein Chem. 1968 23 121; (b) J. F.Brandts in ‘Structure and Stability of Biological Macromolecules’ ed. G. Fasman and S.Timasheff Marcel Dekker N.Y. 1969,p. 213;(c)C. N.Pace C.R.C. Crit. Rev. Biochem. 1975,3,1;(d)J. Bandekar Internat.J. Peptide Protein Res. 1978 11 191; (e) F.Ahmad and P. McPhie Biochemistry 1978 17 241. (a)R.L. Baldwin Ann. Rev. Biochem. 1975,44,453;(6)T.M.Li J. W. Hook III. H. G. Drickamer and G. Weber Biochemisrry,1976,15,5571; (c) T. Y.Tsong ibid.,p. 5467;(d)P. J. Hagerman and P. L. Baldwin ibid. p. 1462. (a)D.A. Chignell A. Azhir and W. B. Gratzer EuropeanJ. Biochem. 1972,26,37;(6)E.A. Carrey and R. H. Pain Biochim. Biophys. Acta 1978,533,12; (c) K. P.Wong and C. Tanford J. Biol. Chem. 1973,248,8518. K. Kuwajima and S. Sugai Biophys. Chem. 1978,8 247. 5 A. Gafni involves an intermediate form (which has the same amount of helical structure as the native form) and cannot be explained by a two-state mechanism.’ In many proteins groups of amino-acid residues along the polypeptide chain are found to cluster into relatively rigid domains the interactions inside domains being stronger than inter-domain interactions.A multi-domain model has recently been proposed for peni~illinase,~~ in which the domains can separate without appreciable change of secondary structure. The presence of structural domains distributed along the flexible polypeptide chain of cell surface protein has also been suggested from denaturation studies by surfactants and denaturants.’ The existence of these domains in proteins is believed to play a major role in the folding mechanism since it suggests that parts of a continuous polypeptide chain can fold independently to form nucleation sites for protein folding. The nucleation process in which locally ordered segments of the polypeptide chain are formed is the initial stage of folding and the experimental and theoretical evidence for this view has been reviewed by Anfinsen and Scheraga.’ A method for predicting nucleation sites for protein folding based on hydrophobic interactions was introduced.lo The nucleation sites are predicted by searching the amino-acid sequence of the protein for pockets of non-polar residues whose net free energy of interaction is negative. The predictions of the model were tested for several proteins. Kanehisa and Tsong” have treated the mechanisms of folding and unfolding of globular proteins also assuming locally ordered regions of the polypeptide chain (which they termed ‘clusters’) at an early stage of the folding process. The presence of clusters of amino-acid residues in the denatured state of proteins was verified by evaluating the heat capacity and volume changes which accompany protein denaturation.l2 The published values of these parameters are in accord with those estimated for a denatured state in which hydrophobic side-chains remain clustered and out of contact with solvent water. A statistical-mechanical treatment of protein folding has been pre~ented’~ based on the theory of a one-dimensional lattice gas with long-range many-body inter- actions. The model was applied to three typical proteins and their folding pathways were traced. The roles of both short- and long-range interactions in protein folding were studied by a Monte-Carlo simulation method within the framework of a lattice m0de1.l~ The protein molecule was represented by a self-avoiding chain polymer consisting of N units located at the lattice points of the square lattice and connected by linear bonds whose length is the same as the lattice constant.Two types of forces were considered to operate; short-range and long-range interactions. It was concluded that while the short-range interactions accelerate the folding and unfold- ing transitions the highly co-operative stabilization of the native structure is achieved by the long-range interactions. In a second study” a three-dimensional lattice model designed to assimilate the native conformation of lysozyme was tested ’M. Nozaka K. Kuwajima K. Nitta and S. Sugai Biochemistry 1978 17 3753. G. Colonna S. S. Alexander Jr. K. M. Yamada I. Pastan and H. Edelhoch J. Biol. Chem. 1978,253 7787. C.B. Anfinsen and H. A. Scheraga Adv. Protein Chem. 1975,29,205. lo R. R. Matheson Jr. and H. A. Scheraga Macromolecules 1978,11,819. M. I. Kanehisa and T. Y. Tsong J. Mol. Biol. 1978 124 177. J. Bello J. Phys. Chern. 1978,82 1607. l3 H.Wako and N. Saito J. Phys. SOC.Japan 1978,44,1939. l4 N.Go and H. Taketomi Proc. Nut. Acad. Sci. U.S.A.,1978,75 559. Is Y. Ueda H. Taketomi and N.Go Biopolymers 1978 17 1531. Physical Chemistry of Proteins 7 by computer simulation; the process of unfolding as well as the denatured states of the model were discussed. The effect of structural heterogeneity inside globular proteins on their structure and folding has been analysed. l6 This heterogeneity arises from the presence (in the interior of proteins) of clusters of polar residues which separate domains of non-polar residues.From the amino-acid sequence the residues occurring in non-polar clusters were predicted satisfactorily for lysozyme. It was suggested that formation of large non-polar domains in the protein interior is preceded by the formation of small clusters by local interactions. The independent re-folding of domains was studied experimentally in elastase.” By limited proteoly- sis using trypsin a fragment of elastase containing residues 126-245 was prepared and its complete denaturation was achieved by 6M guanidine-HCl in the presence of mercaptoethanol. Upon removal of denaturants spontaneous re-folding of the fragment to its native form took place and the ability to bind elastin was restored.Ribonuclease A.-Several of the studies published in 1978dealt with the re-folding reactions of ribonuclease A the first protein shown to be able to resume its native structure after denaturation and complete reduction of its disulphide bonds.18 The extended two-state model for the re-folding of ribonuclease A has been tested. 19*20 This model assumes that the only species to be found during the folding process of proteins are the native structure and multiple forms of the completely unfolded protein.21 Multiple denatured forms of lysozyme have recently been found by both laser Raman22 and ultrasonic absorption23 studies. Two unfolded species of ribo-nuclease A (U and U,) have indeed been and it has been proposed that these differ only in the cis-trans isomers of the four proline residues.21 One of these two species (U,) the fast-re-folding species is postulated to contain all the proline residues in the same conformations as in the native enzyme while U1 consists of several species each containing one or more proline residues in a conformation different from that found in the native enzyme.The slow folding of U1is rate-limited by the interconversion to U2 which upon formation folds rapidly. Nall et a1.l’ found that the activation enthalpy of the U1eU2interconversion is much too small to be attributed to proline isomerization and that the reaction rate changes sharply with guanidineVHC1 concentration (while the rate of proline isomerization in model compounds does not). Thus proline isomerization may not be the rate-limiting step in the U1 eU2 interconversion.The authors also found a rapid formation of partial folding in re-folding conditions thus casting doubt on the validity of the extended two-state model for re-folding of ribonuclease A. A kinetic study of the folding of the two unfolded forms of ribonuclease A as a function of solvent viscosity involved adding various concentrations of glycerol or sucrose.2o No dependence on solvent viscosity was found for either the fast folding reaction or for the U1 eU2 intercon- J. Crampin B. H. Nicholson and B. Robson Nature 1978,272 558. l7 C. Ghelis M. Tempete-Gaillourdet and J. M. Yon Biochem. Biophys. Res. Comm. 1978,84,31. C. B. Anfinsen E. Haber M. Sela andF. H. White Jr. Proc. Nut. Acad. Sci. U.S.A.,1961,47 1309.l9 B. T.Nall J. R. Garel and R. L. Baldwin J. Mol. Biol. 1978,118,317. *’ T. Y.Tsong and R. L.Baldwin Biopolymers 1978,17 1669. 21 J. F. Brandts H. R. Halvorson and M. Brennan Biochemistry 1975 14,4953. 22 R. S. Porubcan K. L. Watters and J. T. McFarland Arch. Biochem. Biophys. 1978 186 255. 23 K.Yamanaka H. Nakajima and Y. Wada Biopolymers 1978 17 2159. 24 J. R. Garel and R. L. Baldwin Proc. Nut. Acad. Sci.,U.S.A.,1973,70 3347. 25 J. R.Garel B. T. Nall and R. L. Baldwin Proc. Nut. Acad. Sci. U.S.A. 1976,73 1853. 8 A. Gafni version; hence neither reaction is rate-limited by external frictional effects (due to the solvent) on segmental motion. From the effect of various denaturants on the rate of the fast folding reaction it was concluded that a structural intermediate is involved in the rate-limiting step of this reaction.Early formation of a folding intermediate in ribonuclease A has also been shown by 'H n.m.r. spectroscopy which revealed that the slow-folding species U1 rapidly becomes partly folded upon initiation of re-folding.26 The appearance of enzymatic activity of ribonuclease A was also used to monitor re-folding in a study devoted to the early steps of this process.27 A low level of activity (about 0.04% that of the native enzyme) was found at very early stages of the folding and was shown to be due to partially (or completely) reduced molecules. The findings were explained as the result of an equilibrium between native and unfolded conformations in reduced ribonuclease A.The presence of an intact folded region in the reduced enzyme was also shown in measurements of fluorescence quantum yield and lifetime of the tyrosine residues.28 This folded region was found to be hydrophobic and to contain one tyrosine residue. Oligomeric Proteins.-In oligomeric proteins composed of two or more identical subunits it is usually found that re-folding of the individual subunits precedes their re-association and is the rate-limiting step in the generation of native ~tructure.~~ However in some cases the re-association step is found to be the major barrier in renaturation. This was found to be the case in bovine seminal ribon~clease.~' The two subunits of this dimeric enzyme are similar in structure to ribonuclease A and are connected by two disulphide bonds.Fully reduced and denatured seminal ribo- nuclease was regenerated using the glutathione redox system and the major product was found to be the monomer which was twice as active towards yeast RNA as the native enzyme. The circular dichroism of the monomer differed from that of the native enzyme and resembled the c.d. of ribonuclease A. The inability of the subunits to associate is the result of blocking of the cysteine residues which form the disulphide bonds by glutathione. Active subunits have also been observed upon dissociation of yeast tran~ketolase~~ In the latter case and superoxide di~mutase.~~ only dissociation occurred in 8M urea with no further denaturation suggesting that dissociation precedes disruption of the tertiary structure.The re-folding and re-activation of horse liver alcohol dehydrogenase (LADH) was studied after denaturation of the enzyme in 6M guanidine~HC1.~~ Dissociation of the dimeric enzyme was monitored by its enzymatic activity by circular dichroism and by fluorescence. Re-folding and re-activation were achieved by diluting the solutions thereby reducing the concentration of denaturant. The rate of re-activation of LADH was found to depend strongly on the addition of but was 26 A. D. Blum S. H. Smallcombe and R. L. Baldwin J. Mol. Biol. 1978,118 305. 27 J. R. Garel J. Mol. Biol. 1978 118 331. N. Barboy and J. Feitelson Photochem. Photobiol. 1977,26 561. 29 (a) I. Bjork and C. Tanford Biochemistry 1971 10 1289; (b) L. Bornmann B. Hess and H.Zimmermann-Telschow Roc. Nut. Acad. Sci. U.S.A. 1974,71,1525; (c)H. Tennenbaum-Bayer and A. Levitzki Biochim. Biophys. Acta 1976 445 261. 30 G. K. Smith G. D'Alessio and S. W. Schaffer Biochemistry 1978,17 2633. 31 G. A. Kochetov and 0.N. Solovieva Biochem. Biophys. Res. Comm. 1978 84 515. 32 J. V. Bannister A. Anastasi and W. H. Bannister Biochem. Biophys. Res. Comm. 1978,81 469. 33 J. Gerschitz R. Rudolph and R. Jaenicke European J. Biochem. 1978 87 591. 34 R. Rudolph J. Gerschitz and R. Jaenicke European J. Biochem. 1978.87 601. Physical Chemistry of Proteins 9 not affected by the presence of coenzyme NAD' (coenzyme concentration did however affect the reaction yield by promoting the formation of inactive aggregation products).The rate of re-activation showed strong dependence on enzyme concen- tration; hence an association reaction must be involved in the process. The irrever- sible heat denaturation of LADH was used to study its interaction with the reduced coenzyme NADH.35 Binding of NADH in the presence of the substrate analogue isobutyramide protects the dimeric enzyme against heat denaturation at 75 "C. The denaturation was followed at various degrees of occupation of the binding sites by NADH and it was concluded that the two subunits bind the coenzyme indepen- dently. Binding of NADH to one subunit protects both subunits against heat denaturation probably by shifting the equilibrium that exists between intact enzyme and the heat-labile dissociated subunits towards the former.That the dissociation of LADH to its subunits is the first step in the heat denaturation process is consistent with the experimental observations. Heat denaturation was followed in Micrococcus lysodeikticus adenosine triphosphatase (ATpa~e).~~ Upon heating the multi- subunit enzyme undergoes a major unfolding transition. The denaturation was irreversible; however no evidence for dissociation into subunits was found after cooling to room temperature and the products retained a large degree of secondary structure. Unfolding of the subunits thus seems to take place without dissociation pointing to a very compact quaternary structure apparently stronger than the tertiary structure of the subunits. ATP CaC12 or high concentrations of ATPase increased the stability of enzyme against denaturation.Coenzyme binding was found to protect pig heart lactate dehydrogenase against inactivation by y-ray~.~' The formation of ternary complexes of enzyme :coenzyme :substrate analogue provided maximal protection apparently by stabilizing the enzyme conformation. Stabilization of dihydrofolate reductase against heat denaturation by the coenzyme NADPH has been In this case too the formation of ternary complexes with coenzyme and inhibitors enhanced the stabilization effect. Several enzymes which form covalent intermediates with their substrates during catalysis were stabilized against pH-induced denaturation in the presence of substrates owing to the stability of these intermediates3' Yeast D-glyceraldehyde 3-phosphate dehydrogenase a tetrameric enzyme was found to be reversibly inactivated and dissociated to its subunits by lowering the temperature from 25 to 0°C in the presence of ATP and mer~aptoethanol.~' The inactive monomer was capable of binding one NAD' molecule which indicates a high degree of residual secondary and tertiary structure in it.The dissociation reaction at 0 "C followed first-order kinetics with a T1/2 of 180minutes. The re-activation at low enzyme concentration showed a higher reaction order with a limiting value of 2. Association of the inactive monomers is therefore a prerequisite for enzymatic activity. The kinetic parameters resemble the ones obtained for re-activation after full denaturation in 6M guanidine~HC1.~~ Thus a common intermediate is suggested " A.Gafni and L. Brand Biochim. Biophys. Acra 1978 537,446. 36 J. A. Ayala and M. Nieto Biochem. J. 1978.169 371. 37 M.Saito Internat. J. Radiation Biol. 1978 34 95. B. Bielaski-Kitchell and R. W. Henkens Biochim. Biophys. Acra 1978,534,89. 39 M. Volini and S. F. Wang. Arch. Biochem. Biophys. 1978,187 163. 40 P.Bartholmes and R. Jaenicke European J. Biochem. 1978,87,563. 41 R. Rudolph I. Heider and R. Jaenicke European J. Biochem. 1977,81 563. 10 A. Gafni in re-activation processes following dissociation of the enzyme under widely differing conditions. Tetrameric Escherichia coli aspartase dissociates and denatures upon heating to 55"C and does not recover its activity upon Treatment of the denatured enzyme with 6M guanidineqHC1 followed by dilution of the denaturant was found to restore a significant fraction of the original enzymatic activity as well as the quaternary structure.A similar re-activation was found for acid-denatured aspartase. Extensive unfolding with a potent denaturant of the entangled poly- peptide chains formed by irreversible denaturation reactions may thus lead to regeneration of the native structure upon removal of denaturant. Another tetra- meric enzyme rabbit muscle aldolase was dissociated to its subunits at alkaline pH (pH >12).43The dissociation was followed by denaturation both reactions being of first order. The free subunits were considered to show partial activity. The re- activation kinetics were described by sequential interconversion and association reactions with first- and second-order rate constants respectively.Changes in the absorption spectrum and the enhancement of tryptophan fluores- cence were used to follow the denaturation of Euglena cytochrome c-552caused by urea and g~anidine.HC1.~~ The accessibility of the two tryptophanyl residues to iodide was studied by fluorescence quenching and was found to be heterogeneous at low denaturant concentrations (which however were enough to cause significant changes in the protein structure as judged from the fluorescence enhancement). Upon complete denaturation the two tryptophanyl residues were equally exposed. These results indicate that intermediates with partial native structure exist along the denaturation pathway and hence that the unfolding of cytochrome c-552does not follow a two-step mechanism.Conformational changes which precede denaturation by guanidineSHC1 were also reported for rabbit skeletal (Y(Y -tropomyosin using pyrene excimer flu~rescence.~~ Partially unfolded intermediates exist at low concentrations of the denaturant. A complex mechanism of unfolding was also found for human gly~ophorin~~ by circular dichroism viscosity and fluorescence of 1-anilino-8-naphthalenesulphonatebound to the protein. A simple two-state mechanism of denaturation was found for chymotrypsinogen A that was immobil- ized by covalent attachment to derivatized porous glass beads.47 Denaturation was achieved by urea and guanidineSHC1 and first-order kinetics for both unfolding and re-folding were monitored by changes in the enzyme fluorescence.3 Conformations and Interactions Studied by Optical Activity Techniques Circular Dichroism of Intrinsic Chromophores.-Circular dichroism (c.d.) was used to study the secondary structure of lipophilin (a hydrophobic protein purified from human central nervous system myelin) in aqueous and lipid ~olutions.~~ The secondary structure composition was calculated assuming that the contributions to 42 M. Tokushige and G. Eguchi Biochim. Biophys. Acta 1978,522 243. 43 R. Rudolph A. Haselbeck F. Knorr and R. Jaenicke Hoppe-Seyler'sZ. Physiol. Chem. 1978,359,867. 44 I. Aviram and C. Weissmann Biochemistry 1978,17 2020. 45 S. L. Betcher-Lange and S. S. Lehrer J. Biol. Chem. 1978 253 3757. 46 D.M.Byers and J. A. Verpoorte Biochim. Biophys. Acta 1978 533,478. 47 H. E.Swaisgood V. G. Janolino and H. R. Horton Arch. Biochem. Biophys. 1978,191,259. 48 S. A.Cockle R. M. Epand J. M. Boggs and M. A. Moscarello Biochemistry 1978,17 624. Physical Chemistry of Proteins 11 the c.d. spectrum in the peptide region (190-250 nm) from a-helix @-sheet and unordered conformations were additive. When the protein was introduced into phosphatidylcholine vesicles about 75% a-helical conformation was found. A similar content of a-helix was found for the water-soluble form of lipophilin obtained by dialysis from 2-chloroethanol to aqueous solutions devoid of lipid. Other water-soluble preparations obtained by dialysis from phenol-acetic acid-urea had lower helical contents.The amount of @-sheet was minimal for lipophilin incorporated into vesicles. The far-u.v. c.d. spectra of eight lectins have been studied with the aim of resolving contributions from the various structural forms to their secondary All eight proteins appeared to have secondary struc- tures dominated by @-pleated sheet which is generally true for lectins. An attempt to quantitate the three structural components (a-helix @-sheet and unordered) met with difficulties for some of the lectins due among other reasons to the unusual c.d. spectra which show considerable ellipticity above 225 nm a region where no electronic transition of peptides is known to take place. Alteration in the secondary structure of mouse brain tubulin upon interaction with the tranquillizing drug chlorpromazine was studied by ~.d.~' Upon binding of one mole of the drug large changes in the far-u.v.c.d. (200-250 nm) of the protein were observed while no change in the c.d. above 260 nm could be detected. These findings apparently reflect changes in secondary structure (reduction in a-helix content and increase in @-structure) without extensive changes in tertiary structure or in the state of association of the subunit protein. Binding of additional chlorpromazine molecules to tubulin did not induce further changes in secondary structure. The observation of an unperturbed tertiary structure is in accord with the conclusion reached by Lee etal.'l that in calf brain tubulin a disulphide bond maintains a highly stable structural domain containing aromatic chromophores.A large increase in the helical content of troponin C was found to accompany the binding of Ca2'.52 The increase was biphasic about 62% occurring with Ca2' binding to a class of sites with Kca of 2.7 x lo71mol-' and the remaining change with Ca2' binding to a class of sites having Kcaof 3.1 x lo51 mol-'. No effects of salt concentration on the c.d. spectrum of human myeloma immunoglobulin G between 205 and 250nm were detected while difference spectra indicated a change in the degree of exposure of tyrosine residues.53 Increasing salt concentrations thus affect the tertiary (but not the secondary) structure of this immunoglobulin molecule. The contribution of phenylalanine and tyrosine side-chains to the far-u.v.c.d. of peptides and proteins has been calc~lated.'~ The interaction of the aromatic ring with neighbouring peptide bonds generates rotatory strength in its La transition. From the preferred backbone and side-chain conformations it was concluded that the most probable conformations have positive Labands. Thus,a significant positive contribution due to the rotatory strength of aromatic residues in globular proteins is 49 M. S. Herrmann C. E. Richardson L. M. Setzler W. D. Behnke and R. E. Thompson Biopolymers 1978,17,2107. A. G. Appu-Rao D. L. Hare and J. R. Cann Biochemistry 1978,17,4735. '' J. C. Lee D. Corfman R. P. Frigon and S. N. Timasheff Arch. Biochem. Biophys. 1978 185,4. 52 J. D. Johnson and J. D. Potter J. Biol. Chem. 1978,253 3775.53 V. P. Zavyalov S. Yu. Tetin V. M. Abramov and G. V. Troitsky Biochim. Biophys. Acta 1978,533 496. 54 R. W. Woody Biopolyrners. 1978.17 1451. 12 A. Gafni to be expected from nearest-neighbour interactions without invoking stacking interactions among the aromatic side-chains. This was confirmed by calculations for proteins of known conformations at the nearest-neighbour level. The effect is found in several snake venom toxins where side-chain contributions from tyrosine and tryptophanyl residues manifest themselves as positive c.d. bands in the far-u.v. region. Bush et al? studied the c.d. of two groups of cyclic hexapeptides having p turns whose geometry can be firmly established by X-ray crystallography and n.m.r. spectroscopy.They found that no significant contributions to the c.d. spectra in the 195-240 nm region were made by the aromatic phenylalanyl chromophore. Circular Dichroism of Haem-containing Proteins.-The far-u.v. c.d. spectrum of yeast cytochrome c peroxidase indicates that its secondary structure contains roughly equal amounts of cy -helix p -structure and unordered ~tructure.~~ Removal of the haem group did not affect the secondary structure to any significant extent. The c.d. spectra of the haem group under various conditions suggest that this moiety is located in a region tightly surrounded by amino-acid residues of the protein. Interaction between the two haem groups (haem c and haem dl) which are found in each subunit of the dimeric enzyme cytochrome oxidase was studied by magnetic circular di~hroism.~’ This technique and some of its biological applications have been reviewed during 197EL5*The spectral range of 350-700 nm was studied for the cytochrome oxidase in the oxidized reduced and reduced-carbon monoxide forms.All the spectra were simple sums of contributions from haem c and haem dt Also no effect of ligand binding to ferrous haem dl was observed in the magnetic c.d. spectrum of haem c in contrast to a previous rep~rt.’~ Thus no interaction between the two haem groups in each subunit was detected. The intrinsic c.d. spectrum of ferric cytochrome b562isolated from E. coli reflected the presence of 52% a-helix and 48% random structure while a slightly lower cy -helix content (49%) was found for the reduced form of the pr~tein.~’ In neither case was the presence of p-structure detected.The near-u.v. and visible c.d. spectra (up to 600 nm) of the oxidized and reduced forms of the proteins were also studied and it was concluded that conformational alterations in the haem group occurred upon change in the valency of the metal ion. The oxidation-reduction process was also found to change the environment of the aromatic chromophores in the cyto- chrome molecule. Changes in the tertiary structure of a cytochrome molecule which depend on its redox state were also reported to occur in the cytochrome b-cl complex from yeast.61 In this case changes in the c.d. spectrum of cytochrome c1 were induced by reduction of the cytochrome b molecule indicating conformational interactions between the two cytochrome molecules.This interaction was not influenced by binding of the inhibitor antimycin A to the complex; hence the interaction does not seem to be involved in the mechanism of electron transport through the complex. 55 C. A. Bush S. K. Sarkar and K. D. Kopple Biochemistry 1978,17,4951. 56 G. Sievers Biochim. Biophys. Acta 1978 536 212. ” L.E.Vickery G. Palmer and D. C. Wharton Biochem. Biophys. Res. Comm. 1978,80,458. 58 B. Holmquist and B. L. Vallee in ‘Methods in Enzymology’ ed. C. H. W. Hirs and S. N. Timasheff Academic Press New York 1978 Vol. 49 p. 149. s9 Y. Orii H. Shimada T. Nozawa and M. Hatano Biochem. Biophys. Res. Comm. 1977,76,983. 6o P.A. Bullock and Y. P.Myer Biochemistry 1978,17 3084.61 J. Reed T. A. Reed and B. Hess European J. Biochem. 1978,91,255. Physical Chemistry of Proteins 13 Circular Dichroism and Circular Polarization of Luminescence in Protein-Ligand Interactions.-The c.d. spectra of thionicotinamide adenine dinucleotides (SNAD’ and SNADH) were investigated in solution and in the complexes formed with horse liver alcohol dehydrogenase (LADH).62.63Cleaving experiments of the coenzyme analogue in aqueous solution by phosphodiesterase revealed the existence of stacking interactions between the adenine and thionicotinamide rings in the intact coenzyme. The c.d. spectrum of SNADH in the ternary complex with LADH and isobutyramide had features resembling the spectrum obtained for the cleaved coenzyme suggesting an open conformation for the bound coenzyme.The larger c.d. signal observed for the bound SNADH relative to the spectrum of the free species was ascribed to a higher degree of conformational rigidity of the bound form. A highly immobilized dihydropyridine ring of NADH was also suggested to be present in the ternary complex between coenzyme UDP-galactose-4-epimerase and UMP.64This suggestion is based on the large increase in the c.d. spectrum of the coenzyme that is observed upon binding. Some interactions between proteins and small ligand molecules were studied using the circular polarization of luminescence (c.p.1.) technique. C.p.1. is a measure of the optical activity of fluorophores in their electronically excited state i.e. a c.p.1. spectrum is related to the molecular conformation in the excited state in a way similar to that in which the c.d.spectrum is related to the molecular conformation in the ground state. Both the instrumental aspects of the c.p.1. technique and its biochemi- cal and biophysical applications have been reviewed during 1978.65*66 A comparison between the c.d. and c.p.1. spectra of NADH in aqueous solution showed them to be markedly different.67 However since cleavage of the coenzyme molecule by phos- phodiesterase did not affect the spectra it was concluded that the differences originate in conformational changes of the excited state of the nicotinamide ribose fragment. Significant differences were also found between the c.d. and c.p.1. spectra of the LADH :NADH binary complex as well as between the corresponding spectra of the coenzyme bound to beef heart lactate dehydrogenase (LDH) in binary and ternary complexes (the latter with the inhibitor oxalate).These observations indicate that in all three complexes the nicotinamide ring enjoys some freedom of movement in its electronically excited state and hence is not fully immobilized in this state. In contrast the c.d. and c.p.1. spectra of NADH in the ternary complex with LADH and isobutyramide (IBA)were found to be very similar indicating that in this case the environment of the nicotinamide ring does not change upon excitation. The coenzyme is therefore rigidly bound in the ternary complex. This conclusion is supported by the very high degree of linear polarization of fluorescence observed for NADH in the ternary complex.While the c.p.1. spectra of LADH:NADH and LADH :NADH :IBA differ substantially from one another their c.d. spectra are identical verifying that binding of the substrate analogue immobilizes the nico- tinamide ring without affecting its environment in the binding site. The increased 62 R. Joppich-Kuhn and P. L. Luisi European J. Biochem. 1978,83 587. R. Joppich-Kuhn and P. L. Luisi European J. Biochem. 1978 83,593. 64 S. S.Wong J. Y. Cassim and P. A. Frey Biochemistry 1978,17 516. 65 I. Z. Steinberg in ‘Methods in Enzymology’ ed. C. H. W. Hirs and S. N. Timasheff Academic Press New York 1978,Vol. 49 p. 179. 66 I. Z. Steinberg Ann.Rev. Biophys. Bioeng. 1978 7 113. 67 A. Gafni Biochemistry 1978,17 1301.14 A. Gafni rigidity of the ternary LADH :NADH :IBA complex may explain its high stability against heat denat~ration.~’ The observed similarity in the c.d. spectra of the binary and ternary complexes of NADH with LADH is not a general feature of complexes of the coenzyme. Thus the c.d. spectra of NADH bound to dihydrofolate reductase in binary and ternary complexes (formed with inhibitors) were found to differ markedly,68 reflecting differences in the environment of the bound nicotinamide group. The signs of both c.d. and c.p.1. spectra of NADH bound to LADH were opposite to those of the LDH-bound coenzyme.67 Structural differences between the two nicotinamide- binding sites are thus evident and undoubtedly result from the proximity of these sites to the substrate-binding sites.The latter may be expected to differ being constructed to accommodate very different substrates. The mode of interaction between a series of oligosaccharides dansylated at their reducing end and the homogeneous mouse IgM secreted by MOPC-104E was studied by various technique^.^^ The c.p.1. of all the dansyloligosaccharide-MOPC-104E complexes was within the experimental error of the instrument proving that the dansyl chromophore was not in a chiral environment. The dansyl group appears to protrude out into the solvent and does not interact effectively with the oligosac- charide-binding site. 4Environment of Amino-acid Residues in Proteins Fluorescence Quenching.-The quenching of intrinsic tryptophanyl fluorescence in proteins by compounds of low molecular weight may serve to distinguish between residues exposed to the solvent (which may be quenched) and those buried inside the protein (which are not accessible to the quenching agent).70 The quenching of exposed tryptophans by iodide ions is dynamic and may be described by the Stern-Volmer equation F,/F = 1+&[I-] where Foand F are the fluorescence intensities in the absence and presence of iodide and KOis a constant.However if a second population of tryptophan residues not accessible to I- is present the plot of Fo/F us. [I-] will deviate from linearity. Lehrer70C proposed a modified Stern-Volmer plot to describe the quenching data in such cases FOIW = (1/CI-Ifa&J +(Wa) (2) here hF =Fo-F; i.e.the change in fluorescence intensity observed at a concen- tration of quencher [I-] while fa is the fraction of accessible trytophan residues whose quenching constant is KO. ‘* A. V. Reddy W. D. Behnke and J. H. Freisheim Biochim. Biophys. Acru 1978,533,415. 69 G. Schepers Y. Blatt J. Himmelspach and 1.Pecht Biochemistry 1978,17 2239. 70 (a)S. S. Lehrer Biochem. Biophys. Res. Comm. 1967,29,767; (6) E. A. Burstein Biofiziku 1968,13 433; (c)S. S. Lehrer Biochemistry 1971,10,3254; (d)R. F. Steiner R. E. Lippoldt H. Edelhoch and V. Frattali Biopolymers 1964 Symp. 1 p. 355; (e) S. S. Lehrer in ‘Biochemical Fluorescence Concepts’ ed. R. F. Chen and H. Edelhoch Marcel Dekker New York 1975 Vol. 2 p. 515; (f)S. S. Lehrer and P. C. Leavis in ‘Methods in Enzymology’ ed.C. H. W. Hirs and S. N. Timasheff Academic Press New York 1978 Vol. 49 p. 222. Physical Chemistry of Proteins 15 The effect of iodide on the fluorescence of the oxygen-carrier protein haemocy- anin from the snail Levantina hierosolima was In the apoprotein which is devoid of the copper ions about half of the tryptophanyl fluorescence was accessible to quenching by I-. No quenching by iodide of the fluorescence of oxyhaemocyanin was detected and it was concluded that the residual fluorescence observed in this case originates exclusively in buried tryptophans. Fluorescence decay measure- ments showed that the introduction of copper into the apoprotein to form deoxy- haemocyanin did not affect the emissive properties of the protein while 55% of the tryptophanyl emission was quenched upon the introduction of oxygen.It appears that the copper-binding site in haemocyanin is near the exterior of the protein close to the tryptophans which are accessible to the solvent. Iodide was used as a quencher of exposed tryptophans in antithrombin a single-chain glycoprotein of mol. wt. 65 000 which is a protease inhibitor in the plasma. The quenching curves deviated from the Stern-Volmer equation but followed the modified relation of equation (2). From the data a value of 0.6 was obtained for the fraction of fluorescence due to exposed tryptophan residues in this multi-tryptophan-containing protein. Upon denaturation in 6M guanidine. HCI all the tryptophans were exposed to the solvent. Fluorescence difference spectra revealed that the quenchable tryptophans are located in a relatively polar environ- ment on the protein surface.Upon binding of heparin to antithrombin 111 the pattern of quenching by I- changed drastically. It was concluded that a con- formational change of the protein following heparin binding radically changes the number of exposed tryptophans. The fluorescence of the heparin-antithrombin I11 complex may be dominated by a single tryptophan residue. Horse liver alcohol dehydrogenase contains two tryptophan residues per subunit Trp-15 and Trp-314. From X-ray data it appears that while Trp-314 is buried inside the enzyme Trp-15 is probably close to the surface. 33% of the fluorescence was found to be quenchable by I-.73 The spectrum that was obtained after maximal quenching had an emission maximum at 320nm characteristic of a tryptophan residue in a hydrophobic environment.When this spectrum was subtracted from the enzyme emission in the absence of I- a red-shifted spectrum was obtained with a maximum at 340 nm; hence this emission originates in a tryptophan residue located in a polar environment. Thus only the exposed Trp-15 appears to be accessible to the quencher. The quantum yields of the 'blue' and 'red' tryptophans (i.e.Trp-314 and Trp-15) were evaluated from the data to be 0.37 and 0.19 respectively. In egg-white apo-riboflavin-binding protein no quenching of tryptophanyl fluorescence by I- or Cs' could be detected.74 Both anionic and cationic agents effectively quenched the fluorescence of the denatured protein in 4M guanidineSHC1.All the tryptophan fluorescence in apo-riboflavin-binding protein is therefore due to residues buried inside the protein. Quenching of tryptophan fluorescence of yeast hexokinase isoenzyme B by Cs' I- and glucose has been performed at various pH values.75 The 7' N. Shaklai A. Gafni and E.Daniel Biochemistry 1978 17,4438. 72 R. Einarsson Internat. J. Peptide Protein Res. 1977,10,342. 73 M.A. Abdallah J. F. Diellrnann P. Wiget R. Joppich-Kuhn and P. L. Luisi European J. Biochem. 1978,89,397. 74 Y. Nishina K. Horiike K. Shiga and T. Yarnano J. Biochem. (Japan) 1977,82 1715. D. C.Kramp and I. Feldman Biochirn. Biophys. Actu 1978 537,406. 75 16 A. Gafni data indicated that the four tryptophan residues of the monomer subunit may be classified as (a) one highly accessible surface tryptophan (6) one surface tryp- tophan possibly in a crevice with restricted accessibility (c) one tryptophan residue inside a cleft (partially shielded) and (d)one buried tryptophan in the hydrophobic interior of the protein.An uncharged quenching agent of tryptophan fluorescence in proteins is acryl- amide. This molecule has been used to study the exposure to solvent of tryptophan residues in several immunoglobulin^.^^*^^ Schreiber et al.76studied the quenching in three homogeneous rabbit antipeptidoglycan antibodies in the absence and in the presence of their specific ligands. The fluorescence of the Fc and F(ab‘),fragments is quenched more effectively than that of the intact protein.Binding of the specific haptens to the antibodies has a small effect on the quenching by acrylamide apparently resulting from shielding of tryptophan residues which are located in or near the combining sites. Middaugh and Litman77 studied acrylamide quenching of tryptophan fluorescence in six cold-soluble IgM proteins and in two IgM types which possess cryoglobulin properties (abnormal cold insolubility). The quenching curves deviated from the Stern-Volmer relation and showed an upward curvature. The quenching was accompanied by a blue shift of the emission maximum. The spectra of the IgM molecules in 6M guanidineqHC1 were typical of tryptophans exposed to aqueous environment having a maximum at 350 nm.The quenching data were resolved into static and dynamic components by a modified Stern-Volmer equation. While the static components for quenching of the two cryo-immunoglobulins tested were similar to those of the non-cryo-immunoglobulins some differences between the two classes of IgM were found in the kinetic quenching components. Quenching of tyrosine fluorescence by acrylamide was used to study the effects of an inter-chain disulphide cross-link introduced into rabbit skeletal tropomyosin on the structure of the 52% of the tyrosine fluorescence in the folded state of tropomyosin was accessible to acrylamide quenching while in the unfolded state (in 5M guanidine-HCl) the accessibility was 88%. No increase in accessibility of tryosines to the quencher was found between 0 and 2 moles of denaturant despite the increase in flexibility and partial unfolding of the protein as shown by other techniques.Thus the quenching data suggest that the disulphide cross-link keeps the chains together in the region where most of the tyrosines are located i.e. the C-terminal half of the molecule. Trichloroethanol is an efficient quencher of tryptophanyl fluorescence in pro- tein~,’~ and was used to probe these fluorescent residues in wheat-germ agglutinin.*’ At low quencher concentrations the Stern-Volmer law was followed with a rate constant for quenching of 1.2x lo91 mol-’ s-l. In the presence of trichloroethanol the two fluorescent tryptophan residues in each polypeptide chain of the dimeric protein are photochemically modified.This modification leads to a reduction in the haemagglutinating activity of the protein and in its affinity towards chitin oligomers. It appears from the data that tryptophan residues are located in the binding sites of 76 A. B. Schreiber A. D. Strosberg and I. Pecht Zmmunochemistry 1978,15,207. 77 C. R. Middaugh and G. W. Litman Biochim. Biophys. Acra 1978 535 33. ” S. S. Lehrer J. Mol. Biol. 1978,118 209. 79 M. R. Eftink and C. A. Ghiron J. Phys. Chem. 1976,80,486. J. P. Privat and M. Charlier European J. Biochem. 1978.84,79. Physical Chemistry of Proteins wheat-germ agglutinin. This proposition is supported by experiments in which mercuriated oligosaccharides were bound to the protein and the heavy-atom effect on the phosphorescence of tryptophan residues was studied." The mercuriated glycoside derivatives bind to wheat-germ agglutinin in the same way as the cor- responding sugars.The binding results in a drastic quenching of tryptophan fluores- cence with a concomitant increase in the intensity of phosphorescence at 77K. These changes are due to the heavy-atom effect which leads to enhanced inter- system-crossing rates in excited chromophores whose distances from the heavy atom are of the order of van der Waals radii.82 It is therefore evident that a tryptophan residue in the binding site is in close contact with the ligand. Dynamic quenching effects induced in a fluorescent pyrene conjugate of acetyl- cholinesterase by iodide nitromethane and thallous ion were studied and The derivative pyrenebutyl methylphosphonofluoridate (l),conjugates (1) with acetylcholinesterase with a stoicheiometry of one fluorescent label per subunit.The quenching constants of the conjugated pyrene chromophore by all three quenchers are much smaller than those of the pyrene moiety in solution proving the pyrene-binding site to be partially shielded from the solvent. However the decrease in the rate constants for quenching was found to be much smaller for T1' than for I- indicating that the pyrene is located in an anionic environment. Upon binding of (l) the intrinsic tryptophanyl fluorescence of acetylcholinesterase is largely quenched whereas when propidium (a peripheral-site ligand) binds to the protein 80-90% of the pyrene fluorescence is lost.These two quenching processes are considered to result from non-radiative energy transfer in which the pyrene serves as acceptor or donor of energy. Thermal quenching of fluorescence in a series of proteins was studied over the temperature range 5-75 "C and an attempt was made to correlate the quenching with intramolecular structural mobility.84 For proteins containing one or two fluorescent tryptophan residues the following relation was found to hold l/q=a+b. T/q (3) where q is the quantum yield a and b are temperature-independent constants T is the temperature and q is the viscosity of the solvent (water). This dependence of q on T/q (whose value is proportional to the diffusion coefficient in water) was found both in proteins whose tryptophans are exposed to the solvent and those in which " M.Monsigny F. Delmotte and C. Helene Proc. Nut. Acad. Sci. U.S.A.,1978,75 1324. 82 W.C.Galley and R. M. Purkey Proc. Nut. Acad. Sci. U.S.A. 1972,69,2198. *' H.A.Berman and P. Taylor Biochemistry 1978,17 1704. 84 T.L.Bushueva. E. P. Busel and E. A. Burstein Biochim. Biophys. Acta 1978,534,141. 18 A. Gafni these residues are shielded. Deviations from linearity of the l/q us. T/q plots occur only in the temperature ranges in which denaturation of the proteins takes place. The results indicate that the mobility of protein structures is controlled by the diffusion characteristics of the solvent and that the intramolecular collisions between excited chromophores and quenching groups occur during fluctuations of the protein’s structure.Different values were obtained for the constants [a and b of equation (3)] in different proteins reflecting differences in their structures. Solvent Perturbation.-The environment of tryptophan residues in proteins may be studied from perturbations of their absorption spectra induced by solvent mole- cule~.~~ This technique was used in a comparative study of bovine human and guinea-pig a-lactalbumins.86 The perturbing solvents used were ethylene glycol glycerol and dimethyl sulphoxide (20% in each case). The number of exposed tryptophans was calculated from the ratio of A&for the proteins to A& of N-acetyl-L- tryptophan ethyl ester (A& is the difference betwe.cn extinction coefficients in the region 291-294 nm in the presence and absence of perturbing solvent).The results showed almost the same degree of exposure of the tryptophan residues in all three a-lactalbumins both at 25 “C and at 2 “C. Since Trp-26 and Trp-60 of bovine a-lactalbumin are replaced by other amino-acids in the human and guinea-pig protein species these residues cannot be the exposed ones and they must be inaccessible to solvent. This conclusion is based on the assumption that the three lactalbumins have similar conformations and evidence for this was found in the very similar far- and near-u.v. c.d. spectra. A cationic detergent was also used to perturb the tryptophan absorption. From the similar difference spectra obtained for a-lactalbumin and lysozyme it was suggested that the former has a cleft-like region similar to the active-site cleft of lysozyme.The difference spectra indicated that two tryptophan residues are located in this region of the protein. Exposure of tyrosine residues to the solvent in ovomucoid was studied by difference spectroscopy following solvent perturbation by glycerol ethylene glycol and dimethyl ~ulphoxide.~’ The difference spectra in 3.5M guanidineaHC1 showed the same number of tyrosine residues to be exposed as in the absence of denaturant; hence although the protein was significantly unfolded it retained a degree of native structure. Upon increasing the concentration of guanidineSHC1 to 6M complete loss of native structure occurred and all six tyrosine residues were exposed. 5 Determination of Distances by Non-radiative Energy Transfer The quantum-mechanical theory of resonance energy transfer by dipole-dipole interaction between a pair of donor and acceptor chromophores has been developed by Forster.88 The possible use of energy-transfer efficiency as a ‘spectroscopic ruler’ for the determination of distances (of up to several nm) among chromophores in macromolecules was discussed by Stryer and Ha~gland.~~ In the past few years 85 T.T. Herskovits in ‘Methods in Enzymology’ ed. C. H. W. Hirs Academic Press New York 1967 vol. 11 p. 748. 86 K. Takase R. Niki and S. Arima J. Biochem. (Japan) 1978 83 371. M. A. Baig and A. Salahuddin Biochem. J. 1978 171 89. Th. Forster in ‘Modern Quantum Chemistry’ ed. 0.Sinanoglu Academic Press New York 1965 Part 111 p.93. 8y L. Stryer and R. P. Haugland Proc. Nar. Acad. Sci. U.S.A..1967 58 719. Physical Chemistry of Proteins 19 energy transfer has been used in an ever-increasing number of studies and several review articles have been published dealing with various aspects of this technique.” The energy-transfer process by depopulating the electronically excited state of the donor directly competes with light emission and with non-radiative de-excitation processes. The transfer efficiency T,in terms of the donor’s fluorescence intensity is given by T = 1-FIFO= 1/[1+(r/Ro)6] (4) where F and Foare the fluorescence intensities of donor in the presence and absence of the acceptor r is the distance between donor and acceptor and R is that distance between the donor-acceptor pair at which the energy-transfer efficiency is 0.5.According to Forster,88 Ro (in Angstroms) is given by here 4 is the fluorescence quantum yield of the donor in the absence of acceptor K’ is the orientational factor given by K~ = (cos 8DA -3 cos 8A cos 8D)2 where 8DA is the angle between the dipoles of donor and acceptor 8D and 8A are the angles between the dipoles of the donor and acceptor respectively and the line joining them n is the index of refraction of the medium while JDA the overlap integral is given by JDA=([om&A(A) ‘fD(A) ’ A4dA)/IOmfD(n)an (6) where E~(A)is the extinction coefficient of the acceptor and f&) is the relative (corrected) emission intensity of the donor per unit wavelength interval.In order to obtain the distance r between donor and acceptor one has to determine the values of F F, and R,. In most cases it is the determination of R which limits the accuracy of distances determined from energy-transfer measurements mainly due to ~~~*~ uncertainty in the value to be used for K~. Theoretically the orientational factor can range between 0 and 4; however in many cases spectroscopic or structural information regarding the participating chromophores enables one to reduce considerably the range of possible values thereby increasing the accuracy of the determined distance. Haas treated the effect of mixed polarizations in the electronic transitions of donor and acceptor on the accuracy of distances determined from energy-transfer measurements.Mixed polarizations occur when the absorption or emission spectra are characterized by two or more incoherent transition dipole moments. This situation is not uncommon and may result from the presence of several electronic transitions in the pertinent spectral range of absorption of the acceptor or from the partially forbidden character of the electronic transitions of the donor or acceptor. In the latter case the various vibronic transitions of an electronic band may have (a) I. Z. Steinberg Ann.Rev. Biochem. 1971,40,83; (6)R. E. Dale and J. Eisinger in ‘Biochemical p. 115; (c) P. W. Schiller ibid. p. 285; (d) R. F. Fairdough and C. R. Cantor in ‘Methods in Fluorescence Concepts’ ed. R. F. Chen and H. Edelhoch Marcel Dekker New York 1975 Vol.1 Enzymology’ ed. C. H. W. Hirs and S. N. Timasheff Academic Press New York 1978 Vol. 48 p. 347; (e)L. Stryer Ann.Rev. Biochem. 1978,47,819. 91 E. Haas E. Katchalski-Katzir. and I. Z. Steinberg Biochemistry 1978 17 5064. 20 A. Gafni different polarization proper tie^.^' In a chromophore whose electronic transitions have mixed polarizations there is no unique orientation of a transition dipole moment within the chromophore for a given electronic transition. Formally this situation is equivalent to that of a fast but restricted Brownian rotatory motion during the lifetime of the excited state; this has been discussed by Dale and Ei~inger.~" In the extreme case when both donor and acceptor have three perpen- dicular transition moments of equal intensities the value of K will be independent of the orientation of the chromophores (as it is in the case of a fast unrestricted Brownian rotatory motion).A value of f is to be used for K~ in this case. Haas et d9'determined the transition dipole moments of naphthalene and dansyl chromo- phores from their polarization spectra in the spectral range involved in energy transfer between them. The results were applied to evaluate the distribution of end-to-end distances in a series of oligopeptides in viscous solutions using the energy transfer between the two chromophores which were attached to the molecu- lar ends. Due to the mixed polarizations of the naphthalene and dansyl groups the calculated end-to-end distribution function is only very slightly affected by the orientational dependence of the efficiency of energy transfer.In a second paper by Haas et~zl.,~~ the kinetics of fluorescence decay of the donor in the homologous series of oligopeptides labelled at their ends by donor and acceptor was studied in mixtures of glycerol and trifluoroethanol of various viscosities. A disturbance of the equilibrium distribution of end-to-end distances in the population of excited molecules is to be expected due to the energy-transfer process whose efficiency decreases rapidly with distance. This efficiency is enhanced when rear- rangement by diffusion of the molecular ends relative to one another follows excitation. Indeed the fluorescence decay rate of the donor was found to increase with decreasing solvent viscosity.The diffusion coefficients for the Brownian motion of the molecular ends were evaluated from the fluorescence decay data and their values were found to be considerably smaller than those expected for the free chromophores in solvents of the same viscosity. It thus appears that the polymeric chains possess appreciable internal friction which slows down the rate of the Brownian motion. The diffusion coefficients of the end-to-end motion were found to increase with chain length reflecting smaller internal friction in longer chains. A theoretical study of the unperturbed chains Tyr(Ala) Tyr with n =4 or 9 revealed no correlation between end-to-end distances and the relative orientations of the chromophores at the chains ends.94 The dipole moments of the tyrosine residues are very nearly randomly oriented relative to the vector connecting the peptide ends.In a second a Monte-Carlo method was used to generate oligopeptide chains composed of 4,9,or 14repeating units in the random-coil state labelled with tyrosine or tryptophan. Interactions with the solvent (water) were taken into account and the chains represented oligopeptides composed of hydrophobic or hydrophilic amino-acid residues. For all the chains considered the values of K~ were not far from 2. 92 I. Z. Steinberg in 'Biochemical Fluorescence Concepts' ed. R. F. Chen and H. Edelhoch Marcel Dekker New York 1975,Vol. 1 p. 79. 93 E. Haas E. Katchalski-Katzir and I. Z. Steinberg Biopolymers 1978 17 11. 94 A.Englert and M.Leclerc Proc. Nat. Acad. Sci. U.S.A. 1978,75 1050. " M. Leclerc. S. Premilat and A. Englert Biopolymers 1978.17.2459. Physical Chemistry of Proteins 21 Energy transfer was used to evaluate the distances between the AMP- and Mn2'-binding sites in E. coli glutamine ~ynthetase.~~ The fluorescent derivative 1,N6-etheno-AMP(EAMP) was used as donor of energy while the Mn2' ions were replaced by Co2+ which served as the acceptor. The efficiency of energy transfer was followed as a function of Co2' concentration. A biphasic titration curve was obtained and analysed in terms of two Co2'-binding sites. The dissociation constants of Co2' from these sites were calculated from the data. The distances from the E adenine ring to the two bound Co2' ions were evaluated to be 11 and 13 A.The value to be used for Rowas computed assuming a value of $for u2. The absorption band of Co2' in the region of overlap with the ethenoadenine emission spectrum is composed of three orthogonal almost degenerate transition moments; hence the acceptor's dipole moment is effectively isotr~pic.~~*~~ This limits the possible values of K* to a range between f and $(the first value applies when the donor's emission dipole is perpendicular to the line joining the donor and acceptor while the latter value holds when the emission dipole is along this line). The authors concluded that the uncertainty in the distance determination due to the possible range of K values was not more than ~12%.This seems to be a conservative estimate since the electronic transition involved in emission from the E adenine chromophore has mixed polarizations and is not associated with a single dipole moment.99 Co2' was also used as an energy acceptor when substituting for the Zn2' ion of thermolysin molecules in which a luminescent Tb3' ion replaced Ca2' ions in two sites and served as the energy donor.100 A value of $was used for u2,based on the isotropy of the electronic transitions of the two ions in the spectral region pertinent to energy transfer.In an extension of previous studie~,~~,~~ the distance between Tb3' and Co2' was determined as a function of temperature and was found to increase from 14 A at 25 "C to about 21 A at 80 "C. Gradual changes in the structure of the protein thus occur upon heating leading to the increase in distance between the metal-binding sites.These proposed structural changes were not detected by tryptophan fluorescence enzyme activity or optical rotation. The ethenoadenine chromophore was also used as the energy donor when bound (as EATP) to the unique nucleotide-binding site of G-actin."' The acceptor was a non-fluorescent dinitrophenyl derivative bound to the sulphydryl group of Cys-373. While the E ATP molecule was found to be immobilized in its binding site in the time scale of the donor's excited state lifetime the acceptor seemed to enjoy partial rotational freedom. Based on a possible range of K~ values between 5 and $ the distance between the nucleotide-binding site and Cys-373 was calculated to be between 26 and 33 A.Energy transfer was used to study the distance between specific sites on bacterial luciferase."' This enzyme is a heterodimer designated ap where cy is the catalytic subunit. The individual subunits are inactive and in the intact enzyme they are bound with a high affinity. Bacterial luciferase was labelled by allowing an essential 96 J. J. Villafranca S. G. Rhee and P. B. Chock Proc. Nut. Acud. Sci. U.S.A. 1978,75 1255. 97 V. G. Berner D. W. Darnall and E. R. Birnbaum Biochem. Biophys. Res. Comm. 1975,66,763. 98 W. D. Horrocks Jr. B. Holmquist and B. L. Vallee Proc. Nut. Acud. Sci. U.S.A.,1975,72,4764. ''A. Gafni. J. Schlessinger and I. Z. Steinberg J. Amer. Chem. SOC. 1979 101 463. loo S. M. Khan E. R. Birnbaum and D. W. Darnall Biochemistry 1978,17,4669.lo' M. Miki and K. Mihashi Biochim. Biophys. Actu 1978,533 163. '02 S. C. Tu C. W. Wu and J. W. Hastings Biochemistry 1978 17 987. 22 A.Gafni sulphydryl group on the a subunit to react with fluorescent maleimide derivatives. Each of these derivatives was used as donor while 8-anilino-1 -naphthalenesul- phonate (ANS) which binds to the enzyme in a molar ratio of 1:1 served as the acceptor. Energy-transfer efficiencies were determined from the enhancement of ANS fluorescence and by estimating the range of possible values for IC~from fluorescence polarization the distance was evaluated to be between 21 and 37 A. Bound ANS was also used as donor of energy to the natural coenzyme FMN. Since bound FMN is not fluorescent the transfer efficiency was determined from the fluorescence decay rate of the ANS.The distance between the two chromophores was found to be between 30 and 58 A. Energy transfer between chromophores bound to two interacting proteins was found to play a role in Renilla biolumines~ence.'~~ The interaction between luciferase and the green fluorescent protein is highly specific and the energy transfer from enzyme-bound oxyluciferin to the green fluorescent protein is efficient. The distances between individual components of yeast cytochrome c oxidase cytochrome c complex were also evaluated from measurements of energy tran~fer."~ Cytochrome c oxidase is composed of seven distinct subunits and the cytochrome cmolecule is bound to subunit 3. The electron-transferring moiety of the protein contains two haem groups (haem a).Subunit 2 was covalently labelled NHCH,CH,NHCCH,I II 0 with the fluorescent dyes N-(iodoacetamidoethy1)-1-aminonaphthalene-5-sulphonic acid [1,5-AEDANS (2)] or with the 2,6-isomer. These dyes attach to reactive sulphydryl groups. The fast decay of the emission anisotropy of the bound AEDANS indicates that the fluorophore is highly mobile in the complex. A value of $ was assumed for IC~in calculating the distances from AEDANS to the haem a groups or to the haem of cytochrome c. Further justification for using the value for IC'may be found in the fact that the distances measured with the two AEDANS isomers which have quite different molecular symmetries were identical. The distances obtained were subunit 2 to haem a 52A;subunit 2 to cytochrome c (bound to subunit 3) 35A; cytochrome c to haem a 25A.Based on these distances it appears that electron transfer from the haem of cytochrome cto haem a groups of cytochrome coxidase must bridge a considerable distance. Pulse fluorometry was used to study energy transfer from tryptophan residues to NADPH in beef liver glutamate dehydrogenase (GDH).'" Both ternary GDH :NADPH :L-glutamate and quaternary GDH :NADPH :L-glutamate:GTP complexes were studied. Since several tryptophan residues of GDH may serve as energy donors a detailed description of all the transfer processes was not given. The W. W. Ward and M. J. Cormier Photochem. Photobiol. 1978,27 389. M. E.Dockter A. Steinemann and G. Schatz J.Biol. Chem. 1978,253,311. J. C. Brochon Ph. Wahl J. M. Jallon and M. Iwatsubo Biochim. Biophys. Acta 1977,462,759. Physical Chemistry of Proteins 23 tryptophan residues may however be divided into two classes having different energy-transfer rates. Ligand binding induces a conformational change in the enzyme leading to partial quenching of tryptophanyl fluorescence. Energy transfer from tryptophan residues to NADH and NAD' bound to liver alcohol dehy- drogenase was also studied.73 The fluorescence of the two tryptophans in each subunit of the dimeric enzyme is largely quenched upon NADH binding and to a lesser extent when NAD' binds. Energy transfer from tryptophan to NADH accounts for less than half of the observed quenching the rest being apparently caused by conformational changes induced in the enzyme by NADH binding.In spite of the small overlap between tryptophan emission and NAD' absorption it was concluded that energy transfer can contribute to the quenching observed upon NAD' binding. However also in this case a sizeable part of the fluorescence quenching is due to conformational changes which follow coenzyme binding. The fluorescence of tryptophan in proteins decreases with increasing pH,Io6and from this dependence it was suggested that the quenching mechanism is by energy transfer from excited tryptophan to the tyrosinate anions.1o7 Indirect evidence to support this hypothesis was found in p-trypsin from the effects of viscosity and chemical modification of the protein on the fluorescence intensity and lifetime.'0s Quenching by Trp +Tyr energy transfer was the only mechanism consistent with all the experimental results.The minimum distances among the active sites of the four enzyme components of the pyruvate dehydrogenase multi-enzyme complex (from Azotobacter uinlandii) have been estimated from fluorescence energy transfer. lo9 No energy transfer could be detected between thiochrome diphosphate bound to the active site of pyruvate dehydrogenase and the FAD in the active site of lipoamide dehydrogenase. Also no energy transfer was observed from several fluorescent sulphydryl labels bound to the lipoyl moiety of lipoyl transacetylase to either the FAD or to the thiochrome diphosphate. Thus all the active centres of enzymes in the complex are more than 40 A apart at least during some stages of catalysis.The distances between active sites on different catalytic subunits of aspartate trans- carbamoylase and between active sites and sulphydryl groups (which are located near the active sites) on different catalytic subunits were determined by energy transfer.'*' The enzyme of molecular weight 300 000 is composed of two catalytic subunits each being a trimer of mol. wt. 100000 and three regulatory subunits. The energy transfer was measured using hybrid enzyme molecules in which the binding sites of one catalytic subunit were covalently labelled with pyridoxamine phosphate while those of the second catalytic subunit were labelled with pyridoxal phosphate. Energy transfer between these two chromophores and also from pyridoxamine phosphate to 2-mercuri-4-nitrophenol bound to the sulphydryl groups of the second catalytic subunit was measured.In both cases the transfer of energy was efficient and was not affected by the presence of allosteric effectors or of the substrate carbamoyl phosphate. Assuming that the active sites of each catalytic subunit define an R. F. Steiner and H. Edelhoch Nature 1961 192 873. 107 J. W.Longworth in 'Excited States of Proteins and Nucleic Acids' ed. R. F. Steiner and I. Weinryb Plenum Press New York 1971,p. 319. N. Ramachadran and C. A. Ghiron Biochim. Biophys. Acta 1978,532,286. W.H. Scouten A. C. De Graaf-Hess A. Dekok H. J. Grande A. J. W. G. Visser and C. Veeger European J. Biochem.1978,84 17. 110 L.Hsien E.Hahn and G. G. Harnmes Biochemistry,1978.17 2423. 24 A. Gafni equilateral triangle and that the planes of the two triangles are parallel the distance between the binding sites on different catalytic subunits was calculated to be about 30 A. Migration of excitation energy in the highly ordered macromolecular aggregate in phycobilisomes was studied by measuring quantum yields and fluorescence polarization.'" The mean transfer time among the constituent phycobiliproteins was calculated to be about 280 ps. This corresponds to an average of 28 jumps of the excitation energy in the phycoerythrin layer before being captured by phycocyanine. Triplet energy transfer was observed from p-benzophenone that is covalently linked to bovine serum albumin to low-molecular-weight water-soluble quen- The number and locations of the bound benzophenone chromophores were limited by the conditions of preparation and the acceptor molecules were reversibly bound in a protein-quencher complex.Triplet excited acetone is generated in the oxidation of isobutanal catalysed by horseradish peroxidase. 113*114 The triplet species is generated inside the enzyme and is considerably protected from quenching by oxygen. As a result acetone phosphorescence may be observed. Long-range triplet +singlet energy transfer was observed from the excited acetone to the flavin chromophore of FMN or FAD. The sensitized emission of the flavin was indepen- dent of FAD concentration. This emission may therefore arise from a fraction of FAD molecules tightly bound to the enzyme.J. Grabowski and E. Gantt Photochem. Photobiol. 1978 28,47. 'lZ G. I. Glover P. S. Mariano andR. A. Hildreth Photochern. Photobiol. 1978,28 7. 'I3 0.M. M. Faria-Oliveira M. Ham N. Duran P. J. O'Brien C. R. O'Brien E. J. H. Bechara and G. Cilento J. Biol. Chem. 1978 253,4707. '14 M. Ham N. Duran and G. Cilento Bwchem. Biophys. Res. Comm. 1978,81,779.