J. Nol.
Riol.
(1991)
221,
1453-1460
Mapping Electrostatic Interactions Macromolecular Associations
in
Karla K. Rodgers and Stephen G. SligarjDepartments of Chemistry, Biochemistry, Physiology and Biophysics The Beckman Institute for Advanced Science and Technology University of Illinois at Urbana-Champaign Urbana, IL 61801, U.S.A. (Received
4 February
1991; accepted 25 April
and
1991)
In the association of electron transfer proteins, electrostatics has been proposed to play a role in maintaining the stability and specificity of the biomolecular complexes formed. An excellent model system is the interaction between mammalian cytochrome b5 and cytochrome c, in which the X-ray structures of the individual components reveal a complementary asymmetry of charges surrounding their respective redox centers. Determining the exact extent of the electrostatic interactions and identifying the specific residues involved in the formation of the electron transfer complex has proved more elusive. We report herein the utilization of high-pressure techniques, together with site-directed mutagenesis, to provide a map of the interaction domains in biomolecular complex formation. The application of high pressure disrupts macromolecular associations since dissociation of the complex results in a decreased volume of the system due to the solvation of charges that had been previously sequestered in the interface region and force solvation of hydrophobic surfaces. Site-directed mutagenesis of a totally synthetic gene for rat liver cytochrome b5, which expresses this mammalian protein in Escherichia coli as a hemecontaining soluble component, was used to selectively alter negatively charged residues of cytochrome b5 to neutral amide side-chains. We have demonstrated that the interaction domain of cytochrome b, with cytochrome c can be mapped’ from a comparison of dissociation volumes of these modified cytochrome b,-cytochrome c complexes with the native complex. Using these techniques we can specifically investigate the role of particular residues in the equilibrium association of these two electron transfer proteins. Single-point mutations in the interaction domain give nearly identical effects on the measured dissociation volumes, yet removal of acidic residues outside the recognition surface yield volumes similar to wild-type protein. Multiple mutations in the proposed protein-protein interaction site are found to allow greater solvent-accessibility of the interface as reflected in a diminution in the volume changes on subsequent charge removal. This is indicative that the interprotein salt-bridges in this complex provide a mechanism for a greater exclusion of solvent from the interfacial domain of the complex, resulting in a more stable association.
Keywords: electrostatics;
pressure;
cytochrome
1. Introduction Multiprotein associations play a central role in a multitude of biochemical and biological processes. For example, biological electron transport systems typically involve the transport of reducing equivalents through a series of distinct protein-protein complexes. The role of electrostatics in initiating and providing the stabilizing factors in proteint Author addressed.
to whom
all correspondence
should
be
c; cytochrome
b,; volume
changes
protein complexes is considered an integral and : important part of these interactions. With the advent of X-ray crystallography and utilization of chemical modification techniques, the importance of salt-bridge contacts within several protein-protein complexes has been implicated (Kang et al., 1978; Tamburini et al., 1985). These research efforts have prompted the formulation of computer docking models for several specific complexes based primarily on the ion-pairs that could be formed between the heterologous partners. These proposals have included complexes between cytochrome c-
1453 0022-~83H/91/201453-08
$03.00/0
0
1991
Academic
Press
Limited
1454
K. K. Rodgers and S. G. Sligar
cytochrome c peroxidase (Poulos & Kraut, 1989), cytochrome c-cytochrome b, (Salemme, 1976; Mauk et al., 1986), cytochrome b,-hemoglobin (Poulos & Mauk, 1983), cytochrome c-flavodoxin (Matthew et al., 1983), cytochrome P-450,,,-cytochrome b, (Stayton et al., 1989) and putidaredoxinputidaredoxin reductase (Geren et al., 1986). A diversity of physical techniques have been used to investigate the interactions beween one of the most extensively studied systems, cytochrome c and cytochrome b, (Mauk et al., 1982), including nuclear magnetic resonance (Eley & Moore, 1983), infrared spectroscopy (Holloway & Mantsch, 1988) and fluorescence spectroscopy (Stayton et al., 1988). Although these two proteins do not normally meet physiologically, rapid in vitro electron transfer from cytochrome b, to cytochrome c has been reported (McLendon & Miller, 1985; &in et al., 1991). The crystal structures of cytochrome c and cytochrome 6, reveal an asymmetry of surface charges that surrounds the exposed heme groups of these proteins (Takano et aI., 1973; Mathews et al., 1972) and ionic strength dependencies of the protein-protein association are indicative of a major role for electrostatic interactions. In addition, the complementarity of cytochrome b, acidic residues with the basic residues of cytochrome c provides a model for the diprotein complex based primarily on the electrostatic stabilization that these potential saltbridges could provide (Salemme, 1976). Further modeling of this complex by Wendoloski et al. (1987) using molecular dynamics simulations resulted in a configuration of the complex more favorable for electron transfer with a reduced Fe-Fe distance as compared with the static “docked” model. Chemical derivatization of the lysine residues of cytochrome c proposed to be important in complex formation resulted in reduced reaction rates with cytochrome b, (Ng et al., 1977), thereby substantiating the structure of the proposed complex. However, the method of inferring proteinprotein contact sites from altered electron transfer activities of derivatized proteins is an indirect technique, in that the rate rather than the association of the proteins is used to determine residues involved in the interface. Also, chemical modification of proteins with bulky substituents raises the question of steric effects on association and electron transfer activit,y. Thus, a technique for providing direct’ the particular residues evidence identifying involved in recognition and formation of a proteinprotein complex is necessary. With a combination of high-pressure techniques and site-directed mutagenesis, we have developed a method to directly investigate and identify charged residues involved in protein-protein recognition, thereby mapping the interfacial domain formed during the association. Utilization of high-pressure techniques in the investigation of protein-protein interactions has proven to be a particularly powerful method (Weber & Drickamer, 1983; Fisher et al., 1986; Kornblatt et al., 1988). From Le Chatelier’s principle, increasing pressure induces an equilibrium displacement that
results in a net decrease in volume for the system. Electrostatic contributions to interaction free energies are particularly well suited for high-pressure investigation, since rncreases in pressure will tend to disrupt charge-pair interactions due to increased solvation. The exposure of the charged groups to solvent results in an electrostriction of the solvent,, contributing to a corresponding decrease in the overall volume of the system. Tn this investigation. high-pressure techniques are used to quantify the actual thermodynamic volume change for the association/dissociation of the evtochrome (‘.cytochrome b, complex. These experiments were coupled with site-directed mutagenesis of selected surface charge amino acids of rytochrome b, to identify residues involved in potential salt-bridge formation in the association with cytochrome c. Preliminary data of this work from our laboratory has been reported (Rodgers et al., 1988).
2. Materials
and Methods
Horse heart cytochrome c (type VI) was obtained from Sigma and used without further purification. The gene synthesis. bacterial expression and purification of rat liver cytochrome b, has been described (Beck von Rodman et al., 1986). Site-directed mutagenesis of cytochrome b, was conducted as reported by Rodgers et al. (1988). The reconstitution of cytochrome b, was accomplished by t,he method described by Reid et al. (1984). (a)
Measurement
vf dixsociation
constants
Determination of the dissociation constant of the cytochrome c-cytochrome b, complex (in 2 miw-Tris.HCl, pH 7.4) was achieved by analyzing the perturbation of the optical spectra upon formation of the diprotein aggregat’e using the documented increase in Soret absorbance at 416 nm (Mauk et al.. 1982). Spectrophotometric measurements were conducted with a Cary 219 spectrophotometer. Difference spectra were obtained and spectrophotometric titrations carried out as described by Erman & Vitello (1980) using matched tandem cuvettes. The dissociation constants were determined using the method described by Eisenthal & Cornish-Bowden (1974). (b) High-preuuuru
techniques
The volume change of the diprotein dissociation was quantified from a variation in the dissociation constant with pressure by subjecting a preformed complex of cytochrome b, and cytochrome c to increasing hydrostatic pressure in a steel optical cell with quartz windows (Palladini t Weber, 1981) installed in a Varian DMS 100 spectrophotometer. Hydrostatic pressure was generated by a single stage pump (High Pressure Equipment) using absolute ethanol as the pressurizing fluid. Increasing dissociation of the diprotein complex is represented by a decrease in absorbance at 416nm. Optical spectra were corrected for the spectral changes of isolated cytochrome b, and cytochrome c that occur due to solvent compression. The optical changes occurring due to the dissociation of the cytochrome b,--cytochrome c complex were used to calculate the dissociation constants of the complex at each pressure (Fig. I ). Our methods for the c*alculation of net volume change upon dissociation of a protein--protein complex has been described (Fisher et al.. 1986).
Mapping
Electrostatic
Interactions
1455
were used, since the crystal structure of rat cytochrome b, has not yet been determined. The appropriate salt-bridges proposed by Salemme (1976) were docked with the carboxylate oxygen atoms 3 d (1 a = @l nm) from the amino nitrogen atoms of the lysine residues of cytochrome c. The proposed interaction domain was checked to assure there was no overlap of van der Waals’ radii. The solventaccessibility of the proteins were calculated using the program ACCESS (v. 2), originally developed by Lee &
Richards (1971). A standard probe radius of 1.4A for
-0.01
water was used in the calculations. -0.02 -0.03
I 350
370
390
,v 410 Wavelength
, 430
450
470
The
Figure 1. Representative
optical absorbance changes demonstrating the effect of pressure on the optical difference spectra of the cytochrome b,-cytochrome c complex. Increasing hydrostatic pressure results in complex dissociation shown by the decrease in the spectra at 416nm. (--) 2OObar; (-) 400bar; (.*.) 6OObar; (----) 800 bar; (-) 1000 bar. Inset: a linear plot of the negative log of the dissociation constant versus pressure yields a slope of -AVJRT.
(c) Graphic
modeling and solvent-accessibility
3. Results
490
(nm)
calculations
The cytochrome b,-cytochrome c complex was visualized using Insight II (Biosym Technologies) running on a Silicon Graphics Personal IRIS. The co-ordinates of bovine cytochrome b, (Brookhaven Data Bank 2B5C)
positions
of the various
acidic
sites
on the
surface of cytochrome b, subjected to mutation and their relation to the cytochrome b, heme prosthetic group are represented in Figure 2. The orientation of the protein-protein complex proposed by Salemme (1976) is shown in Figure 3. This modeling was accomplished utilizing the co-ordinates for tuna cytochrome c and bovine cytochrome b,. The amino acid sequences of tuna versus horse heart cytochrome c and bovine versus rat liver cytochrome 6, are highly homogeneous, particularly in the regions corresponding to the proposed interface of the protein-protein complex. Table 1 lists the AG and AV values obtained for the interaction of cytochrome c with cytochrome 6, proteins at pH 7.4 from titration and high-pressure experiments as described in Materials and Methods. (a) Wild-type cytochrome
bccytochrome c complex
The relationship of AG and AV for the complexes of cytochrome c with the various surface-charge altered cytochrome b, proteins is illustrated in
Figure 2. The alpha carbon backbone of cytochrome b, is shown together with the heme group. The side-chains for the sites of mutation are also shown and labeled. The majority of the mutations consisted of the replacement of the acidic side-chains with their respective amide analogs, with the exception of the mutation of aspartate 66 to a serine residue. The heme group was also replaced with a dimethylester heme to investigate the salt-bridge formed by the propionate group. These changes are designated in the text as DME. The co-ordinates of bovine cytochrome b, were used to produce this illustration (Mathews et al., 1972).
Figure 3. The orientation of the complex of bovine cytochrome b, with tuna cytochrome c proposed by Salemme (1976). The alpha carbon backbone for the 2 proteins is shown along with the side-chains for the residues involved in interprotein salt-bridges. The saltbridges modeled are as follows: E48-K13, E44-K27, DBO-K72, propionate-K79. The distance between the carboxylate oxygen atoms of cytochrome b, with the amino lysine residues of cytochrome c in the modeled complex is 3 d. The heme groups are coplanar with an 8 A distance from heme-edge to heme-edge as in the Salemme proposal.
K. K. Rodgers
and S. G. Sligar
Table 1 Dissociation
of cytochrome b, and cytochrome pH 7.4 in 2 m&f-Tris. HCl
ASBt
Protein
0 0 0 0 0 1 1 1 1 1 2 2 3 3 4
Wild type D66S$ E37Q E56Q Q13E Wild-type-DME§ D60N E48Q E44Q E43Q, E44Q E48Q, DME E44&, E48Q E44&, E48Q, E44&, E48&, E44Q, E48&,
D60N DME D60N,
c at
AG (kcal/mol)ll
-AV (ml/mol)ll
8.69 8.68 861 8.47 8e&? X.12 815 X28 %12 8.09 8.00 7.85 7.55 7.60 7.50
122 117 115 117 127 80 77 87 90 85 60 68 46 52 40
DME
I 5.
301
7.00
that a heme propionate is proposed to form with a lysine residue to be
kO.08
kcal/mol
for
A0
/
/ E440,
E48Q,
I 7.50
DME
DGON,DME
I 8cx3
I 8.50
9.00
AG (kcol/mol)
t ASB is the number of salt-bridges that have been removed in the interface of the proposed protein-protein complex. 1 The nomenclature for the charge modified proteins is as follows: the first letter is the one-letter abbreviation for the original amino acid residue in the native protein at the position specified by the following number and the last letter is the abbreviation for the residue introduced by site-directed mutagenesis. 5 Cytochrome 6, was reconstituted with a ferriprotoporphyrin IX dimethylester heme (DME) to selectively remove a salt-bridge of cytochrome c. 11 Errors are estimated +5 ml/mol for AV.
t
0 E480,
/ E449, E480,DME. E44Q. E48Q, ““5 f
and
Figure 4. The wild-type cytochrome b,-cytochrome c complex was found to form with a favorable free energy of association of - 8.69 kcal (1 cal = 4.184 J)/ mol (in 2 mM-Tris* HCl, pH 7.4), a value that is in good agreement with previous investigations (Mauk et al., 1982; Stayton et al., 1988). The volume change for the dissociation of the protein-protein complex under hydrostatic pressurization was found to be - 122 ml/mol and is itself independent of pressure. Effect on the tertiary structure of the individual proteins is minimal at the pressures utilized (Fisher et al., 1986). Protein denaturation typically occurs at pressures greater than 4 kbar (1 bar = 105 Pa) (Weber & Drickamer, 1983), whereas in this investigation only pressures below 2 kbar were utilized. Weber (1987) has attributed three main reasons for the dissociation of complexes or oligomeric proteins under pressure: (1) the existence of dead volume in the protein-protein interface, (2) the interaction of water dipoles with non-polar groups previously present in the interface resulting in shorter intermolecular distances, and (3) the electrostriction of solvent around newly exposed charge groups after dissociation. We believe that the latter reason may contribute significantly to the dissociation of the cytochrome c-cytochrome 6, complex since this interaction involves several salt-bridges that would be sequestered from solvent in the interface. Complete salt-bridge solvation has been previously estimated to result in a volume change of
Figure 4. Cytochrome h,-eytochrome r complex dissociation. The thermodynamic values of A\(-: and A V for the different complexes of the various charge mutants and wild-type cytochrome 6, with rytochrome c. measured at pH 7.4, are plotted relative t’o each other. The A0 values were determined at atmospheric pressure using optical difference spectroscopy and the AV values were obtained as described in the text,.
roughly -30 ml/mol (Heremans, 1982). If the volume change were due entirely to the solvation of salt linkages, this would be indicative of rough]> four inter-protein salt-bridges in the wild-type cytochrome b,-cytochrome c complex, consistent with the proposed model described by Salemme (1976). The mutation of several acidic side-chains of cyt,ochrome b, to neutral residues that are not positioned in the proposed interaction site (E37Q, E56Q and D66S), resulted in nearly identical A0 and A 1’ values upon interaction with cytochromr c. Thus. these charged groups apparently do not reside in the protein-protein interface or contribute significant,l> to complex formation. (b) Single
site mutations
In order to further characterize the interaction domain of cytochromes 6, and c. we have individually altered the four negatively charged groups in cvtochrome b, implicated in salt-bridge formation with cytochrome c. Single mutations of D60, E48 and E44 to their respective amide analogs were produced. In addition, wild-type cytochrome b, was reconstitut’ed with a dimethylester heme (DMEt) to eliminate the charge of the propionate group proposed to contribute one ion-pair in the complex. The mutations t,o amide groups are nearly isosteric replacements, since the volumes for the acidic and amide side-chains are within 5% for glutamate to glutamine and aspartate to asparagine replacements (Zamyatnin, 1972). These altered side-chains can replace the inter-protein salt-bridges involved with a polar-ionic hydrogen-bond donor/acceptor pair. Thus, this set of mutation substitutions should probe the role of charge interactions while largely
t Abbreviation
used: I)ME.
dimethylester
hrmr
Mapping
Electrostatic
preserving the orientation of the protein-protein complex relative to the wild-type association. In all four mutations of single amino acid residues in the proposed interface between the two proteins, the aflinity was decreased relative to the wild-type complex by approximately 65 kcal/mol. Also, the volume change upon dissociation was suggestive of diminished total electrostriction of solvent around the ionic residues uncovered during dissociation. The volume changes for these single-point mutations averaged 38 ml/mol less than that observed for the wild-type complex. These results are in agreement with the proposal that these four residues form salt-bridges in the protein-protein interface. It has also been proposed that glutamate 43 (E43) participates in a salt linkage in the docked complex of cytochrome b, and cytochrome c (Ng et al., 1977). To examine the role of this residue we generated the double mutation {E43&, E44Q). However, this mutation does not differ significantly from E44Q in t)erms of the observed AG and AV values for the interaction with horse heart cytochrome c, suggesting that there is no major contribution of K43 to the stability of the complex.
(c) Multiple
mutations
The two double mutations, {E48&, DME) and {E44Q, E48&}, resulted in yet further decreases in AC and AV from the wild-type complex. These differences were not thermodynamically additive in relation to the single mutations. For example, the difference in AG and AV for {E44&, E48Q) relative t)o wilcl-type are less than the sum of the differences of these quantities of E44Q and E48Q from wildtype (0.84 versus 0.98 kcal/mol and 54 versus 67 ml/ mol, respectively). Nevertheless, the absolute value of AV within each mutant class is within experimental error, suggesting that a unique proteinprotein complex site exists. A similar case is seen with the triple mutations, {E44Q, E48&, DME} and {E44&, E48Q, D60N). These mutations have a smaller AG and AV value as compared with t.he wild-type than the double mutations. Again, the overall differences in thermodynamic parameters are non-additive. Finally, removal of all four proposed salt-bridges through the quadruple mutation {E44&, E48&, D6ON, DME} results, within the limits of error, in a similar AG and Al’ value as observed for the triple mutants. Elimination of all charge-charge interactions in the interface gives an overall decrease in AG of 1.19 kcal/mol from wild-type, assuming the same complex orientation as the wild-type cytochrome b,-cytochrome c association. This would imply that) the electrostatic interactions contribute to only about 14(& of the free energy of association of cytochrome b, and cytochrome c. In this context, the term ele&rostatic interactions refers to ion-pairs formed between fully charged side-chains. The difference in volume change of 82 ml/mol between t’he wild-type complex and the quadruple mutant
Interactions
1457
suggest that electrostriction accounts thirds of AV upon complex dissociation cytochrome b, and cytochrome c.
for about two of wild-type
(d) Solvent-accessiblecalculations of the complex A model of the cytochrome b,-cytochrome c complex using the proposed salt-bridges is shown in Figure 3. This orientation of the complex was used to calculate a solvent-accessible surface for the docked pair of proteins. The solvent-accessible area of the individual proteins, embedded as in the modeled complex, was also calculated as described in Materials and Methods. The difference between the area for individual proteins and that of the complex is the area sequestered from solvent in the protein-protein interface. The solvent-accessibility of every atom in the molecule is also obtainable and it is possible to determine the amount that each residue was sequestered from solvent in the proposed interface domain. The overall area of the interfacial surface obtained by these calculations using a probe radius of 1.4 A was 800 A2.
4. Discussion One major goal of the research endeavor presented in this paper is to define the area of a protein surface that is involved in the recognition of another macromolecule and to determine the thermodynamic contributions of electrostatic interactions in macromolecular associations. As an ideal model system we have examined the redox transfer partners cytochromes c and b,. The method of coupling high pressure with site-directed mutagenesis allows us to map experimentally the specific charged residues involved in complex formation, as well as to define the overall contribution of electrostatics to the association. We have selectively altered surface charge residues of cytochrome 6, by replacement of aspartic and glutamic acids with their respective amide analogs, resulting in mutations that are nearly isosteric to the native acidic side-chains. Furthermore, the amide groups so introduced can serve as potential hydrogen-bond donors/acceptors with the lysine residues or other polar residues of cytochrome c, avoiding a residual bare charge in the interface and potentially maintaining a similar orientation of the complex as in wild-type cytochrome b,. High-pressure methods can give an indication of the degree of electrostatic: and non-polar interactions involved with complex formation by quantifying the volume changes of dissociation. Combined with site-directed mutagenesis the specific charge groups participating in surface recognition events can be identified, since the polar but neutral amide groups that have replaced the charged carboxylate groups will not lead to large volume changes due to solvation. Other measurements of the dissociation of a bovine cytochrome b,-porphyrin cytochrome c complex under high pressures obtained a AV value of -50 ml/mol (Kornblatt et aZ., 1988), approxi-
1458
K. K. Rodgers
mately 50% of the A V value reported in our current and previously published studies (Rodgers et al., 1988). Their studies, however, used a cytochrome c protein containing a metal-free porphyrin and displaying a dissociation constant differing tenfold from that measured with native proteins. There is thus the distinct possibility that this discrepancy arises from species differences or that a different interfacial interaction or interprotein solvation exists in their complexes. The advantages of site-directed mutagenesis using replacements with nearly isosteric amide functionalities is evident. At pH 7.4, our results demonstrate that four residues of wild-type cytochrome b, are involved in salt-bridges with cytochrome c (E44. E48, D60 and a heme-propionate), consistent with Salemme’s (1976) model. This is readily apparent in Figure 4, since altering the charges of these specific residues resulted in a 30 to 45 ml/mol decrease in the volume of dissociation of the complex, indicative that the wild-type charge groups are sequestered from solvent in the formed diprotein complex. Thus, these single alterations together with the negative control mutations allowed us to identify the major acidic residues of cytochrome b, that provide for the specificity in recognition of cytochrome c. Multiple mutations of surface acid functionalities have provided insight into the overall contributions of electrostatics in complex formation. Removal of all four proposed salt-bridges through generation of the mutant protein {E44&, E48&, D60N, DME) decreased the free energy of association by only 14% (assuming the relative complex orientation has remained unchanged). Clearly, electrostatics do not provide the main stabilizing factors in the overall association of this protein-protein complex. Tnstead, it is likely that hydrogen bonds formed in the interface, as well as van der Waals’ interactions resulting from desolvation due to the complementarity of protein surfaces. contribute to the majority of the association free energy. The conclusion that essentially only four residues participate in salt-bridges with cytochrome c at pH 7.4, demonstrates the specific nature of the association between cytochrome c and cytochrome b,. Tt is apparent that the major interaction sit’e of cytochrome b, with cytochrome c is near the eiectrostatically favorable exposed heme edge. It has been previously suggested that electrostatics provides long-range steering for the itiitial association of protein-protein interactions (Matthew et al., 1983), as well as fine-tuning the orientation of the complex for maximal electron transfer activity (Salemme. 1976). Since the acidic residues of cytochrome 0,. experimentally shown by us to interact in t,he interface of the complex, are positioned close t’o the exposed heme redox center of cytochrome b,. t,his partner could be oriented in a favorable manner in the complex for efficient electron transfer. In contrast to the overall .free energy of association, electrostatic factors are the main contribution to the thermodynamic volume changes of the system. The four salt-bridges combined contribute
and S. G. Sligar to two thirds of A V for protein-protein dissociation. The effectiveness of hydrostatic high-pressure techniques in investigating electrostatic interact,ions during multi-component assembly is evident. Our mutagenesis of the acidic residues on t’he cytochrome b, surface did not greatly alter the geometry or orientation of the cytochrome b,-~ cytochrome c complex as determined by measurement of the first-order electron-transfer rat.es of the wild-type cytochrome b,-cytochrome c complex with respect to the single mutations. E48Q and D60N (&in et al.. 1991). The rates for the two mutant protein complexes were within experimental error with a measured electron-transfer rate of the wild-type complex, 1.7 x 103 s- ’ (pH 7.3). If singlesurface mutation events drastically altered t,hc orientation of the complex. a different distance of the redox centers. or overall orientation factors. might be expected to yield a distinctly different electron-t)ransfer rate. Additionally. if mutation at a single site on the cytochrome b, surface resulted in a shift in the orientation of the complex. the alteration in solvent-accessibility would yield large dif%&ences in the absolute AV obtained for complex dissociation relative t’o the other single-site mutants. As t’he experiment)al errors in the absolutItb AV values within the separate class of single. double and triple mutat’ions are within experimental error, our results suggest, a relatively stable protrill protein interfacial domain. From Figure 4 it. is apparent, there is a significant degree of non-additivity of volume and Gibbs free of complex formation with multiplr energies mutations in the proposed interface domains. Since t’he four salt-bridges in the complex appear t,o define t’he periphery of the protein-protein interaction site, removal of these salt-bridges by mutation ma? allow greater solvation in these areas with respect to t,he wild-t,ype complex. The resulting volumt~ changes t,herefore would be smaller since thta remaining salt)-bridges may already be partially solvated in the mut,ant) complexes. This effect would be increased a,s more salt’-bridges were removed as is evidenced by the data presented in Figure 1. This would also contribute t’o smaller changes in AG as more charges were removed from the intrrfacar allowing greater solvent exposurr. The nhsolutr AC value for the disruption of a salt-bridge depends OII t,he dielectric of the surrounding medium. IYpon greater salvation of the interface, the AG value for t,he dissociation of the remaining salt-bridges therrfore would he signiticantlv less. A similar effect ot solvent on the AV and A(: changes upon intarea,sing mutations would also explain t’hr striking lint)ar c~orrelat~ion obvious in Figure 4. This c:orrelat,ioll reflects t)hr role of &c%rostatics in contributing to changes in both A V and AG and the similar effecsts upon solva,tinp the int,erfac!ial domain of t hc complex. The relationship is also suggest,ire that the complex is in the same relative orientJatiorr with all of t,he various caomplexea with no mtljor st,eric* cahanges upon mutat~ion. Tt also implies another rok for salt-bridges in the csomplex irl that, w-itjh
Mapping
Electrostatic
increasing electrostatic interactions, additional solvent exclusion occurs at the interface resulting in a more stable association. The effect of non-additivity with multiple mutations may also be accounted for by significant changes in the structure of the complex. Upon elimination of three or four charges in the proposed interface, possible formation of multiple configurations of the complex with the interaction of alternate ion-pairs may occur. This would result in a larger volume change than would be expected if the complex orientation remained unchanged. A similar proposal that an alternate complex between bovine cytochrome h, and horse heart cytochrome c may occur at increased pH values has been reported (Mauk et al., 1986). However, it is not likely that the linear correlation between AG and AV would exist’ if significant, changes in complex orientation were to occur upon removal of multiple charges in the proposed interface. Thus, we favor the interpretation that- either an alterat.ion of local dielectric constant due to solvent access or electrostatic interactions between interfacial residues can account for the observed non-additivity of AV and AG with the multiple mutations. However, it is likely that upon elimination of multiple charges in the interface. the cbomplex would be less restrained than in the wildtvpe associat’ion resulting in a greater flexibility &thin the interface. The volume change upon dissociation of the quadruple mutant {E44&, E48&, D6ON, DME} with cytochrome c reflects remaining volume change upon removal of all salt-bridges in the interface. However, our experimental results appear to demonstrate that no major shifts in the configuration of the complex occur upon mutagenesis. Alternative contributions to the remaining AV of the quadruple charge mutations could involve electrostrict,ion of cytochrome c lysine residues buried in the interface. As the interprotein ion-pairs of the proposed complex define the boundaries of the protein-protein interface, mutagenesis of the corresponding residues of cytochrome b, could allow solvat’ion of each respective ion-pair. Thus, the decrease in AV upon subsequent removal of each negatively charged residue proposed to form an ionpair would be due to the solvation of that respective pair including the lysine residues of cytochrome c. A more likely origin for the residual AV observed for the quadruple mutant is the interaction of solvent with non-polar residues that had been buried in the int.erface of the complex. The area of the proteinprot’eih interface sequestered from solvent may be calculated from the solvent-accessible area of the individual proteins versus the modeled cytochrome b,-cyt’ochrome c interaction. Using the method described by Lee & Richards (1971), the area of the in t.erface was calculated to be 800 A’. Approximately half of this value is due to the area of the buried charged side-chains from both proteins which form the salt-bridges previously discussed. Therefore, the remaining 400 A2 of buried residues is due mostly t,o the non-polar residues in the inter-
1459
Interactions
face. It has been estimated that for the interaction of I A2 of non-polar residues with water, a decrease of @l ml/mol in volume is observed (Silva et aZ., 1986). These considerations would predict roughly 40 ml/mol residual volume contribution to the Gibbs free energy of association of the cytochrome b,-cytochrome c complex, which is exactly that observed. In summary, high-pressure techniques coupled with site-directed mutagenesis has permitted us to map the surface-interaction domain of cytochrome b, in its association with cytochrome c. This surface site is in agreement with Salemme’s proposed model for the complex of these electron-transfer partners. We conclude that there is one major orientation of the complex at neutral pH, without a major sampling of multiple orientations on the surface. If exhibited significant t)wothe two proteins dimensional diffusion as suggested by Northrup et al. (1988) for cytochrome c, cytochrome c peroxidase solvation of the surface carboxylate groups of cytochrome hS would tend to equalize any differences in dissociation volumes for various surface charge mutations. Since we observe clear definition of single buried sites in the association of rat liver cytochrome b, and horse heart cytochrome c, we favor a model of a defined surface-interaction domain with only rapid nutational dynamics that leaves the surfaces mainly sequestered from solvent access. Since complete removal of the four carboxylate groups on the cytochrome h, surface proposed to be in the interfacial domain reduces the AC: value of association by only 14%. electrostatics do not appear t)o contribute significantly to the stabi!ity of complex formation. The major role of electrostatics in the formation of the protein-protein complex is t,o provide a mechanism for controlling solvent access to the interface, perhaps providing for a more restrained orientation of the complex, and the resulting configuration of the donor and acceptor prosthetic groups. Clearly, high-pressure techniques in concert with site-directed mutagenesis can be used to define other protein-protein interaction sit,es where electrostatics are proposed to play a role in the association. This research was supported by grants Pu’ational Institutes of Health, PHS (;M33775 GM31756, the Kational Science Foundation Research Laboratory, and the Biotechnology and Development Corporation. We thank Weber for many useful discussions.
from the and PHS Materials Research
Dr Gregorio
References Beck
von Bodman, S., Schuler, M. A.. Jollie, D. R. Sligar, S. G. (1986). Synthesis. bacterial expression and mutagenesis of the gene coding for mammalian cytochrome b,. Proc. Nat. Acad. Rci.. 11.8.A.
&
83,
9443-9447. Eisenthal, linear
R. & Cornish-Bowden, plot. Biochem. J. 139.
A. (1974). 715-720.
The
direct
1460
Eley,
K. K. Rodgers and S. G’. Nigur
C. G. S. & Moore. C. R,. (1983). ‘H-n.m.r. Investigation of the interaction between cytochrome c and cytochrome b,. Riochem. .I. 215. 11-21. Errnan. .J. E. & Vitello, I,. R. (1980). The binding of cytochrome c peroxidase and ferricybochrome c. J. Biol. Chem. 255, 6224-6227. Fisher, M. T.. White. R. E. & Sligar, S. (:. (1986). Pressure dissociation of a protein-protein electron transfer complex. J. Amer. Chrm. Sot. 108. 68X-6837. Geren. L.. Tuls. J.. O’Brien. P.. Millett. F. $ Peterson. ,J. A. (1986). The involvement of carboxylate groups of putidaredoxin in the reaction with putidaredoxin reductase. J. RioZ. Chem. 261, 15491L1.5495. Heremans, K. (1982). High pressure effects on proteins and other biomolecules. An,nu. Rw. Riophys. Riorng. 11. l-21. Holloway. P. W. bz Mantsch. H. H. (1988). Infrared spectroscopir analysis of salt bridge formation and c,,vt,ochrome C. cytochrome b, between Biochemistry, 27. 7991-7993. Kang, C”. H., Brautigan. I>. L., Usheroff. K. XL Margolaish. E. (1978). Definition of cytochrome c binding domains by chetnical modification. -1. Hiol. (‘hum. 253, 6502.--6510. Kornblatt), ,J. A., Hui Bon Hoa. C:.. Eltis, I,. & Mauk. A. U. (1988). The effec%s of pressure on porphyrin cmcytochrome b, complex format)ion. ./. Amrr. C’hwu. sot. 110, 5909-5911. Lee. 13. & Rirhards. F. M. (1971). The interpretation of protein structures: estimation of stat,ic accessibility. J. Mol. Biol. 55. 379-400. Mathews. F. S.. Levine, M. & Argos. I’. (1972). Three\caalf liver synthesis of Fourier dimensional cyt,ochrome b, at 2.8 .A resolution. ,/. No/. Biol. 64. 449-464. Matthew. ,J. J.. Weber, P. (!.. Salrmme. F. R. XL Ric~hards. F. M. (1983). Electrostatic orientation during electron transfer between flavodoxin and q%ochromcb 301. 169%171. c. Xature (London). Mauk, M. R.. Reid, T,. 8. & Mauk, A. (:. (1982). Spretrophotometric analysis of the interaction between cytochrome 6, and cytochrome c. Riochwzistry, 21. 1843-1846. Mauk, M. R.. Mauk. A. G.. Il’rber, I’. C. & Matthew. .J. 1s. (1986). Ele&ostatir analysis of t,hr interact,ion of cytochrome c with native and dimrthyl ester hemr substituted caytochrome h,. Biochemistr~y. 25. 70857091. McLendon. (:. & Miller. J. R. (1985). The dependence of’ biological electron transfer rates on rxothermicity: the cytochrome c/cytochromr b, couple. .J. Amw. C’hrm. A’oc. 107, 7811-7816. Ng. S., Smith, M. B.. Smith, H. T. & Millet. F. (1977). Effect, of modification of individual cytochrome c lysines on the reaction with cytochrome b,. Biochemistry, 16. 4975-4978. Northrup, S. H., Boles, ,J. 0. & R,eynolds. ,J. (‘. I,. (1988).
Brownian dynamics of cytoc*hrorne c ant1 c~~~tochrotttr c peroxidase association. Scierbce, 241. BS--70. Paladini. A. A. Ki Wrbrr. (:. (1981). l’ressur,r-itttlrtc,~(l revrrsi ble dissociation of enolase. Riwhrmistry. 20. 2587 -2593. Poulos. T. I,. Hr Kraut. .I. (19X0). A hypottrt~tical trtotlrl of’ the cytochromr r peroxidase. cyt~ochromr r elect rot1 t,ransfer c~omplrx. ./. Hiol. C’hrm. 255. 10322~ 1OX%). Poulos. T. I,. CG !Mauk. A. (G. (1983). Models for tht c~omplrxt~s formed between q%ochromt~ OS and tttcs subunits of’ met~hrmoglobin. ,/. 12irrl. f ‘SPIN. 258. 7369 737:%. Qin. I,., Rodgers. K. K. & Niger. S. (:. (I!)!)1 ). b:lec+rort bransfer between c*ytochromr 11, surfacr tnutants ant1 qtjoc*hromtl C. Nol. (‘rystallogr. Liq. ( ‘rysttrllo~yr. 194. 3 I I 3 I 6. Reid. 1,. S.. Mauk. M. R. & Mauk. A. (:. (19X-C). IColr trf hrtnr propionate groups in c*ytocahrotntL bs rlcc*trotr transfer. ./. .-Zrnc~r. C’h~n/. SW. 106, 2182 AlX.5. Rodgers. K. K.. I’oehapsky. T. C’. CVSligar. S (:. (I!+?++) rI1N~~1.011101~‘(‘IIlat of Probing the tnec*hanisms i’ recognition: tht, c~yt>oc~hrornt~ bs~c.~toc,hrotn~~ c~omplex. Scirr~~r. 240. I657 1659. Salrmme. F. R. (1976). A hypothetical structurt. for, att t+rt rot1 tritnstrt c~omplex intrrtnolr~ular trf caytocahromrs c and h,. .J. Xol. Hiol. 102. 563&.56X. Silva. .I. I,.. Miles. E. \V. k \Veber. (Z. (1!186). I’ressttrt, disso?iatGon and c~otiforrnationr~l drift of thr /&tlinitJr of t,ryptophan synthase. Rioch.umistr!y. 25. 5780 57%. Stayton. I’. S.. Fisher. M. T. & Sligar. 8 ( :. (198X). of c,ytoc.hromt* h, Drtrrniination iissociatiotr rrac+iotls. ./ Uir,/. (‘//PN 263. I4544 I354X. Stagton. 1’. S.. l’otrlos. ‘I’. I,. & Sligar. 6. C:. (I!N!)). I’utidarrdoxin c~ornprtitively inhibitas c~,vtoc.hromr b,c~~to~hromr P-4.50,,, t+ctron-tr’arisfpr c~trnrltlt~s. Bior:hrm istry. 28. x20 I ~8205. Takano. T.. Kallai. 0. K.. Swanson. Ii. & IGkrrsott. IC. b:. (197R). The strucaturr of ferroc~~t’oc.hrornr C’ at A.45 .q resolution. ./. Hiol. (‘hem. 248. X34 :i%T, Tamburini. 1’. I’.. IlThitr. K. E. & Schrnkman. .I. Ii. (1985). (‘htkmic~l (.hat,actrrizat,ion of protein Itrcttein t,etwern ey tochrome P-450 antI intrrac.tions c~ytoc~hromt~ b,. ./. Hiol. C’hrm. 260, 5007 4Ol5. Wrber. (:. (1987). I)issociation of oligomeric~ proteitts I,>, hydrostatic, pressurt’. In High I’re.ss./~w ( ‘h~mistr!/ rrrrtl Niochrnrisfry (van Eldik. R. Ki .lonas. ,I.. r&). pp. 401 420. I). Keidt~l. New l’ork. H. ( 1!18:1). Thtn ~+fA.t 01’ itiph \Vebrr. (:. k I)rickamrr. am1 other I~iotrrolrc~t~l~~s pre‘ssurp rtlton proteins @curt. tfrr. Hiop/ly.v. 16. R-1 1". Wendoloski. .I. .J.. Matthew. .J. 13.. \Yrl,rr. I’. (I S Salernrnr. 1’. R. (1987). Molrc~ular dyriami~s of it c-c,?-tochromr h, t~lrctron tratisf+r caytochromcx c*oml)lt,x. Sciprac~. 238. 7!14&‘i97. Zamyatnin. A. .A. (1972). Protein volumes itt solution. J’roq. Rioph!/.u. Xol. Bid. 24. 107. 113.
Edited by P. Wright
Riol.
(1991)
221,
1453-1460
Mapping Electrostatic Interactions Macromolecular Associations
in
Karla K. Rodgers and Stephen G. SligarjDepartments of Chemistry, Biochemistry, Physiology and Biophysics The Beckman Institute for Advanced Science and Technology University of Illinois at Urbana-Champaign Urbana, IL 61801, U.S.A. (Received
4 February
1991; accepted 25 April
and
1991)
In the association of electron transfer proteins, electrostatics has been proposed to play a role in maintaining the stability and specificity of the biomolecular complexes formed. An excellent model system is the interaction between mammalian cytochrome b5 and cytochrome c, in which the X-ray structures of the individual components reveal a complementary asymmetry of charges surrounding their respective redox centers. Determining the exact extent of the electrostatic interactions and identifying the specific residues involved in the formation of the electron transfer complex has proved more elusive. We report herein the utilization of high-pressure techniques, together with site-directed mutagenesis, to provide a map of the interaction domains in biomolecular complex formation. The application of high pressure disrupts macromolecular associations since dissociation of the complex results in a decreased volume of the system due to the solvation of charges that had been previously sequestered in the interface region and force solvation of hydrophobic surfaces. Site-directed mutagenesis of a totally synthetic gene for rat liver cytochrome b5, which expresses this mammalian protein in Escherichia coli as a hemecontaining soluble component, was used to selectively alter negatively charged residues of cytochrome b5 to neutral amide side-chains. We have demonstrated that the interaction domain of cytochrome b, with cytochrome c can be mapped’ from a comparison of dissociation volumes of these modified cytochrome b,-cytochrome c complexes with the native complex. Using these techniques we can specifically investigate the role of particular residues in the equilibrium association of these two electron transfer proteins. Single-point mutations in the interaction domain give nearly identical effects on the measured dissociation volumes, yet removal of acidic residues outside the recognition surface yield volumes similar to wild-type protein. Multiple mutations in the proposed protein-protein interaction site are found to allow greater solvent-accessibility of the interface as reflected in a diminution in the volume changes on subsequent charge removal. This is indicative that the interprotein salt-bridges in this complex provide a mechanism for a greater exclusion of solvent from the interfacial domain of the complex, resulting in a more stable association.
Keywords: electrostatics;
pressure;
cytochrome
1. Introduction Multiprotein associations play a central role in a multitude of biochemical and biological processes. For example, biological electron transport systems typically involve the transport of reducing equivalents through a series of distinct protein-protein complexes. The role of electrostatics in initiating and providing the stabilizing factors in proteint Author addressed.
to whom
all correspondence
should
be
c; cytochrome
b,; volume
changes
protein complexes is considered an integral and : important part of these interactions. With the advent of X-ray crystallography and utilization of chemical modification techniques, the importance of salt-bridge contacts within several protein-protein complexes has been implicated (Kang et al., 1978; Tamburini et al., 1985). These research efforts have prompted the formulation of computer docking models for several specific complexes based primarily on the ion-pairs that could be formed between the heterologous partners. These proposals have included complexes between cytochrome c-
1453 0022-~83H/91/201453-08
$03.00/0
0
1991
Academic
Press
Limited
1454
K. K. Rodgers and S. G. Sligar
cytochrome c peroxidase (Poulos & Kraut, 1989), cytochrome c-cytochrome b, (Salemme, 1976; Mauk et al., 1986), cytochrome b,-hemoglobin (Poulos & Mauk, 1983), cytochrome c-flavodoxin (Matthew et al., 1983), cytochrome P-450,,,-cytochrome b, (Stayton et al., 1989) and putidaredoxinputidaredoxin reductase (Geren et al., 1986). A diversity of physical techniques have been used to investigate the interactions beween one of the most extensively studied systems, cytochrome c and cytochrome b, (Mauk et al., 1982), including nuclear magnetic resonance (Eley & Moore, 1983), infrared spectroscopy (Holloway & Mantsch, 1988) and fluorescence spectroscopy (Stayton et al., 1988). Although these two proteins do not normally meet physiologically, rapid in vitro electron transfer from cytochrome b, to cytochrome c has been reported (McLendon & Miller, 1985; &in et al., 1991). The crystal structures of cytochrome c and cytochrome 6, reveal an asymmetry of surface charges that surrounds the exposed heme groups of these proteins (Takano et aI., 1973; Mathews et al., 1972) and ionic strength dependencies of the protein-protein association are indicative of a major role for electrostatic interactions. In addition, the complementarity of cytochrome b, acidic residues with the basic residues of cytochrome c provides a model for the diprotein complex based primarily on the electrostatic stabilization that these potential saltbridges could provide (Salemme, 1976). Further modeling of this complex by Wendoloski et al. (1987) using molecular dynamics simulations resulted in a configuration of the complex more favorable for electron transfer with a reduced Fe-Fe distance as compared with the static “docked” model. Chemical derivatization of the lysine residues of cytochrome c proposed to be important in complex formation resulted in reduced reaction rates with cytochrome b, (Ng et al., 1977), thereby substantiating the structure of the proposed complex. However, the method of inferring proteinprotein contact sites from altered electron transfer activities of derivatized proteins is an indirect technique, in that the rate rather than the association of the proteins is used to determine residues involved in the interface. Also, chemical modification of proteins with bulky substituents raises the question of steric effects on association and electron transfer activit,y. Thus, a technique for providing direct’ the particular residues evidence identifying involved in recognition and formation of a proteinprotein complex is necessary. With a combination of high-pressure techniques and site-directed mutagenesis, we have developed a method to directly investigate and identify charged residues involved in protein-protein recognition, thereby mapping the interfacial domain formed during the association. Utilization of high-pressure techniques in the investigation of protein-protein interactions has proven to be a particularly powerful method (Weber & Drickamer, 1983; Fisher et al., 1986; Kornblatt et al., 1988). From Le Chatelier’s principle, increasing pressure induces an equilibrium displacement that
results in a net decrease in volume for the system. Electrostatic contributions to interaction free energies are particularly well suited for high-pressure investigation, since rncreases in pressure will tend to disrupt charge-pair interactions due to increased solvation. The exposure of the charged groups to solvent results in an electrostriction of the solvent,, contributing to a corresponding decrease in the overall volume of the system. Tn this investigation. high-pressure techniques are used to quantify the actual thermodynamic volume change for the association/dissociation of the evtochrome (‘.cytochrome b, complex. These experiments were coupled with site-directed mutagenesis of selected surface charge amino acids of rytochrome b, to identify residues involved in potential salt-bridge formation in the association with cytochrome c. Preliminary data of this work from our laboratory has been reported (Rodgers et al., 1988).
2. Materials
and Methods
Horse heart cytochrome c (type VI) was obtained from Sigma and used without further purification. The gene synthesis. bacterial expression and purification of rat liver cytochrome b, has been described (Beck von Rodman et al., 1986). Site-directed mutagenesis of cytochrome b, was conducted as reported by Rodgers et al. (1988). The reconstitution of cytochrome b, was accomplished by t,he method described by Reid et al. (1984). (a)
Measurement
vf dixsociation
constants
Determination of the dissociation constant of the cytochrome c-cytochrome b, complex (in 2 miw-Tris.HCl, pH 7.4) was achieved by analyzing the perturbation of the optical spectra upon formation of the diprotein aggregat’e using the documented increase in Soret absorbance at 416 nm (Mauk et al.. 1982). Spectrophotometric measurements were conducted with a Cary 219 spectrophotometer. Difference spectra were obtained and spectrophotometric titrations carried out as described by Erman & Vitello (1980) using matched tandem cuvettes. The dissociation constants were determined using the method described by Eisenthal & Cornish-Bowden (1974). (b) High-preuuuru
techniques
The volume change of the diprotein dissociation was quantified from a variation in the dissociation constant with pressure by subjecting a preformed complex of cytochrome b, and cytochrome c to increasing hydrostatic pressure in a steel optical cell with quartz windows (Palladini t Weber, 1981) installed in a Varian DMS 100 spectrophotometer. Hydrostatic pressure was generated by a single stage pump (High Pressure Equipment) using absolute ethanol as the pressurizing fluid. Increasing dissociation of the diprotein complex is represented by a decrease in absorbance at 416nm. Optical spectra were corrected for the spectral changes of isolated cytochrome b, and cytochrome c that occur due to solvent compression. The optical changes occurring due to the dissociation of the cytochrome b,--cytochrome c complex were used to calculate the dissociation constants of the complex at each pressure (Fig. I ). Our methods for the c*alculation of net volume change upon dissociation of a protein--protein complex has been described (Fisher et al.. 1986).
Mapping
Electrostatic
Interactions
1455
were used, since the crystal structure of rat cytochrome b, has not yet been determined. The appropriate salt-bridges proposed by Salemme (1976) were docked with the carboxylate oxygen atoms 3 d (1 a = @l nm) from the amino nitrogen atoms of the lysine residues of cytochrome c. The proposed interaction domain was checked to assure there was no overlap of van der Waals’ radii. The solventaccessibility of the proteins were calculated using the program ACCESS (v. 2), originally developed by Lee &
Richards (1971). A standard probe radius of 1.4A for
-0.01
water was used in the calculations. -0.02 -0.03
I 350
370
390
,v 410 Wavelength
, 430
450
470
The
Figure 1. Representative
optical absorbance changes demonstrating the effect of pressure on the optical difference spectra of the cytochrome b,-cytochrome c complex. Increasing hydrostatic pressure results in complex dissociation shown by the decrease in the spectra at 416nm. (--) 2OObar; (-) 400bar; (.*.) 6OObar; (----) 800 bar; (-) 1000 bar. Inset: a linear plot of the negative log of the dissociation constant versus pressure yields a slope of -AVJRT.
(c) Graphic
modeling and solvent-accessibility
3. Results
490
(nm)
calculations
The cytochrome b,-cytochrome c complex was visualized using Insight II (Biosym Technologies) running on a Silicon Graphics Personal IRIS. The co-ordinates of bovine cytochrome b, (Brookhaven Data Bank 2B5C)
positions
of the various
acidic
sites
on the
surface of cytochrome b, subjected to mutation and their relation to the cytochrome b, heme prosthetic group are represented in Figure 2. The orientation of the protein-protein complex proposed by Salemme (1976) is shown in Figure 3. This modeling was accomplished utilizing the co-ordinates for tuna cytochrome c and bovine cytochrome b,. The amino acid sequences of tuna versus horse heart cytochrome c and bovine versus rat liver cytochrome 6, are highly homogeneous, particularly in the regions corresponding to the proposed interface of the protein-protein complex. Table 1 lists the AG and AV values obtained for the interaction of cytochrome c with cytochrome 6, proteins at pH 7.4 from titration and high-pressure experiments as described in Materials and Methods. (a) Wild-type cytochrome
bccytochrome c complex
The relationship of AG and AV for the complexes of cytochrome c with the various surface-charge altered cytochrome b, proteins is illustrated in
Figure 2. The alpha carbon backbone of cytochrome b, is shown together with the heme group. The side-chains for the sites of mutation are also shown and labeled. The majority of the mutations consisted of the replacement of the acidic side-chains with their respective amide analogs, with the exception of the mutation of aspartate 66 to a serine residue. The heme group was also replaced with a dimethylester heme to investigate the salt-bridge formed by the propionate group. These changes are designated in the text as DME. The co-ordinates of bovine cytochrome b, were used to produce this illustration (Mathews et al., 1972).
Figure 3. The orientation of the complex of bovine cytochrome b, with tuna cytochrome c proposed by Salemme (1976). The alpha carbon backbone for the 2 proteins is shown along with the side-chains for the residues involved in interprotein salt-bridges. The saltbridges modeled are as follows: E48-K13, E44-K27, DBO-K72, propionate-K79. The distance between the carboxylate oxygen atoms of cytochrome b, with the amino lysine residues of cytochrome c in the modeled complex is 3 d. The heme groups are coplanar with an 8 A distance from heme-edge to heme-edge as in the Salemme proposal.
K. K. Rodgers
and S. G. Sligar
Table 1 Dissociation
of cytochrome b, and cytochrome pH 7.4 in 2 m&f-Tris. HCl
ASBt
Protein
0 0 0 0 0 1 1 1 1 1 2 2 3 3 4
Wild type D66S$ E37Q E56Q Q13E Wild-type-DME§ D60N E48Q E44Q E43Q, E44Q E48Q, DME E44&, E48Q E44&, E48Q, E44&, E48&, E44Q, E48&,
D60N DME D60N,
c at
AG (kcal/mol)ll
-AV (ml/mol)ll
8.69 8.68 861 8.47 8e&? X.12 815 X28 %12 8.09 8.00 7.85 7.55 7.60 7.50
122 117 115 117 127 80 77 87 90 85 60 68 46 52 40
DME
I 5.
301
7.00
that a heme propionate is proposed to form with a lysine residue to be
kO.08
kcal/mol
for
A0
/
/ E440,
E48Q,
I 7.50
DME
DGON,DME
I 8cx3
I 8.50
9.00
AG (kcol/mol)
t ASB is the number of salt-bridges that have been removed in the interface of the proposed protein-protein complex. 1 The nomenclature for the charge modified proteins is as follows: the first letter is the one-letter abbreviation for the original amino acid residue in the native protein at the position specified by the following number and the last letter is the abbreviation for the residue introduced by site-directed mutagenesis. 5 Cytochrome 6, was reconstituted with a ferriprotoporphyrin IX dimethylester heme (DME) to selectively remove a salt-bridge of cytochrome c. 11 Errors are estimated +5 ml/mol for AV.
t
0 E480,
/ E449, E480,DME. E44Q. E48Q, ““5 f
and
Figure 4. The wild-type cytochrome b,-cytochrome c complex was found to form with a favorable free energy of association of - 8.69 kcal (1 cal = 4.184 J)/ mol (in 2 mM-Tris* HCl, pH 7.4), a value that is in good agreement with previous investigations (Mauk et al., 1982; Stayton et al., 1988). The volume change for the dissociation of the protein-protein complex under hydrostatic pressurization was found to be - 122 ml/mol and is itself independent of pressure. Effect on the tertiary structure of the individual proteins is minimal at the pressures utilized (Fisher et al., 1986). Protein denaturation typically occurs at pressures greater than 4 kbar (1 bar = 105 Pa) (Weber & Drickamer, 1983), whereas in this investigation only pressures below 2 kbar were utilized. Weber (1987) has attributed three main reasons for the dissociation of complexes or oligomeric proteins under pressure: (1) the existence of dead volume in the protein-protein interface, (2) the interaction of water dipoles with non-polar groups previously present in the interface resulting in shorter intermolecular distances, and (3) the electrostriction of solvent around newly exposed charge groups after dissociation. We believe that the latter reason may contribute significantly to the dissociation of the cytochrome c-cytochrome 6, complex since this interaction involves several salt-bridges that would be sequestered from solvent in the interface. Complete salt-bridge solvation has been previously estimated to result in a volume change of
Figure 4. Cytochrome h,-eytochrome r complex dissociation. The thermodynamic values of A\(-: and A V for the different complexes of the various charge mutants and wild-type cytochrome 6, with rytochrome c. measured at pH 7.4, are plotted relative t’o each other. The A0 values were determined at atmospheric pressure using optical difference spectroscopy and the AV values were obtained as described in the text,.
roughly -30 ml/mol (Heremans, 1982). If the volume change were due entirely to the solvation of salt linkages, this would be indicative of rough]> four inter-protein salt-bridges in the wild-type cytochrome b,-cytochrome c complex, consistent with the proposed model described by Salemme (1976). The mutation of several acidic side-chains of cyt,ochrome b, to neutral residues that are not positioned in the proposed interaction site (E37Q, E56Q and D66S), resulted in nearly identical A0 and A 1’ values upon interaction with cytochromr c. Thus. these charged groups apparently do not reside in the protein-protein interface or contribute significant,l> to complex formation. (b) Single
site mutations
In order to further characterize the interaction domain of cytochromes 6, and c. we have individually altered the four negatively charged groups in cvtochrome b, implicated in salt-bridge formation with cytochrome c. Single mutations of D60, E48 and E44 to their respective amide analogs were produced. In addition, wild-type cytochrome b, was reconstitut’ed with a dimethylester heme (DMEt) to eliminate the charge of the propionate group proposed to contribute one ion-pair in the complex. The mutations t,o amide groups are nearly isosteric replacements, since the volumes for the acidic and amide side-chains are within 5% for glutamate to glutamine and aspartate to asparagine replacements (Zamyatnin, 1972). These altered side-chains can replace the inter-protein salt-bridges involved with a polar-ionic hydrogen-bond donor/acceptor pair. Thus, this set of mutation substitutions should probe the role of charge interactions while largely
t Abbreviation
used: I)ME.
dimethylester
hrmr
Mapping
Electrostatic
preserving the orientation of the protein-protein complex relative to the wild-type association. In all four mutations of single amino acid residues in the proposed interface between the two proteins, the aflinity was decreased relative to the wild-type complex by approximately 65 kcal/mol. Also, the volume change upon dissociation was suggestive of diminished total electrostriction of solvent around the ionic residues uncovered during dissociation. The volume changes for these single-point mutations averaged 38 ml/mol less than that observed for the wild-type complex. These results are in agreement with the proposal that these four residues form salt-bridges in the protein-protein interface. It has also been proposed that glutamate 43 (E43) participates in a salt linkage in the docked complex of cytochrome b, and cytochrome c (Ng et al., 1977). To examine the role of this residue we generated the double mutation {E43&, E44Q). However, this mutation does not differ significantly from E44Q in t)erms of the observed AG and AV values for the interaction with horse heart cytochrome c, suggesting that there is no major contribution of K43 to the stability of the complex.
(c) Multiple
mutations
The two double mutations, {E48&, DME) and {E44Q, E48&}, resulted in yet further decreases in AC and AV from the wild-type complex. These differences were not thermodynamically additive in relation to the single mutations. For example, the difference in AG and AV for {E44&, E48Q) relative t)o wilcl-type are less than the sum of the differences of these quantities of E44Q and E48Q from wildtype (0.84 versus 0.98 kcal/mol and 54 versus 67 ml/ mol, respectively). Nevertheless, the absolute value of AV within each mutant class is within experimental error, suggesting that a unique proteinprotein complex site exists. A similar case is seen with the triple mutations, {E44Q, E48&, DME} and {E44&, E48Q, D60N). These mutations have a smaller AG and AV value as compared with t.he wild-type than the double mutations. Again, the overall differences in thermodynamic parameters are non-additive. Finally, removal of all four proposed salt-bridges through the quadruple mutation {E44&, E48&, D6ON, DME} results, within the limits of error, in a similar AG and Al’ value as observed for the triple mutants. Elimination of all charge-charge interactions in the interface gives an overall decrease in AG of 1.19 kcal/mol from wild-type, assuming the same complex orientation as the wild-type cytochrome b,-cytochrome c association. This would imply that) the electrostatic interactions contribute to only about 14(& of the free energy of association of cytochrome b, and cytochrome c. In this context, the term ele&rostatic interactions refers to ion-pairs formed between fully charged side-chains. The difference in volume change of 82 ml/mol between t’he wild-type complex and the quadruple mutant
Interactions
1457
suggest that electrostriction accounts thirds of AV upon complex dissociation cytochrome b, and cytochrome c.
for about two of wild-type
(d) Solvent-accessiblecalculations of the complex A model of the cytochrome b,-cytochrome c complex using the proposed salt-bridges is shown in Figure 3. This orientation of the complex was used to calculate a solvent-accessible surface for the docked pair of proteins. The solvent-accessible area of the individual proteins, embedded as in the modeled complex, was also calculated as described in Materials and Methods. The difference between the area for individual proteins and that of the complex is the area sequestered from solvent in the protein-protein interface. The solvent-accessibility of every atom in the molecule is also obtainable and it is possible to determine the amount that each residue was sequestered from solvent in the proposed interface domain. The overall area of the interfacial surface obtained by these calculations using a probe radius of 1.4 A was 800 A2.
4. Discussion One major goal of the research endeavor presented in this paper is to define the area of a protein surface that is involved in the recognition of another macromolecule and to determine the thermodynamic contributions of electrostatic interactions in macromolecular associations. As an ideal model system we have examined the redox transfer partners cytochromes c and b,. The method of coupling high pressure with site-directed mutagenesis allows us to map experimentally the specific charged residues involved in complex formation, as well as to define the overall contribution of electrostatics to the association. We have selectively altered surface charge residues of cytochrome 6, by replacement of aspartic and glutamic acids with their respective amide analogs, resulting in mutations that are nearly isosteric to the native acidic side-chains. Furthermore, the amide groups so introduced can serve as potential hydrogen-bond donors/acceptors with the lysine residues or other polar residues of cytochrome c, avoiding a residual bare charge in the interface and potentially maintaining a similar orientation of the complex as in wild-type cytochrome b,. High-pressure methods can give an indication of the degree of electrostatic: and non-polar interactions involved with complex formation by quantifying the volume changes of dissociation. Combined with site-directed mutagenesis the specific charge groups participating in surface recognition events can be identified, since the polar but neutral amide groups that have replaced the charged carboxylate groups will not lead to large volume changes due to solvation. Other measurements of the dissociation of a bovine cytochrome b,-porphyrin cytochrome c complex under high pressures obtained a AV value of -50 ml/mol (Kornblatt et aZ., 1988), approxi-
1458
K. K. Rodgers
mately 50% of the A V value reported in our current and previously published studies (Rodgers et al., 1988). Their studies, however, used a cytochrome c protein containing a metal-free porphyrin and displaying a dissociation constant differing tenfold from that measured with native proteins. There is thus the distinct possibility that this discrepancy arises from species differences or that a different interfacial interaction or interprotein solvation exists in their complexes. The advantages of site-directed mutagenesis using replacements with nearly isosteric amide functionalities is evident. At pH 7.4, our results demonstrate that four residues of wild-type cytochrome b, are involved in salt-bridges with cytochrome c (E44. E48, D60 and a heme-propionate), consistent with Salemme’s (1976) model. This is readily apparent in Figure 4, since altering the charges of these specific residues resulted in a 30 to 45 ml/mol decrease in the volume of dissociation of the complex, indicative that the wild-type charge groups are sequestered from solvent in the formed diprotein complex. Thus, these single alterations together with the negative control mutations allowed us to identify the major acidic residues of cytochrome b, that provide for the specificity in recognition of cytochrome c. Multiple mutations of surface acid functionalities have provided insight into the overall contributions of electrostatics in complex formation. Removal of all four proposed salt-bridges through generation of the mutant protein {E44&, E48&, D60N, DME) decreased the free energy of association by only 14% (assuming the relative complex orientation has remained unchanged). Clearly, electrostatics do not provide the main stabilizing factors in the overall association of this protein-protein complex. Tnstead, it is likely that hydrogen bonds formed in the interface, as well as van der Waals’ interactions resulting from desolvation due to the complementarity of protein surfaces. contribute to the majority of the association free energy. The conclusion that essentially only four residues participate in salt-bridges with cytochrome c at pH 7.4, demonstrates the specific nature of the association between cytochrome c and cytochrome b,. Tt is apparent that the major interaction sit’e of cytochrome b, with cytochrome c is near the eiectrostatically favorable exposed heme edge. It has been previously suggested that electrostatics provides long-range steering for the itiitial association of protein-protein interactions (Matthew et al., 1983), as well as fine-tuning the orientation of the complex for maximal electron transfer activity (Salemme. 1976). Since the acidic residues of cytochrome 0,. experimentally shown by us to interact in t,he interface of the complex, are positioned close t’o the exposed heme redox center of cytochrome b,. t,his partner could be oriented in a favorable manner in the complex for efficient electron transfer. In contrast to the overall .free energy of association, electrostatic factors are the main contribution to the thermodynamic volume changes of the system. The four salt-bridges combined contribute
and S. G. Sligar to two thirds of A V for protein-protein dissociation. The effectiveness of hydrostatic high-pressure techniques in investigating electrostatic interact,ions during multi-component assembly is evident. Our mutagenesis of the acidic residues on t’he cytochrome b, surface did not greatly alter the geometry or orientation of the cytochrome b,-~ cytochrome c complex as determined by measurement of the first-order electron-transfer rat.es of the wild-type cytochrome b,-cytochrome c complex with respect to the single mutations. E48Q and D60N (&in et al.. 1991). The rates for the two mutant protein complexes were within experimental error with a measured electron-transfer rate of the wild-type complex, 1.7 x 103 s- ’ (pH 7.3). If singlesurface mutation events drastically altered t,hc orientation of the complex. a different distance of the redox centers. or overall orientation factors. might be expected to yield a distinctly different electron-t)ransfer rate. Additionally. if mutation at a single site on the cytochrome b, surface resulted in a shift in the orientation of the complex. the alteration in solvent-accessibility would yield large dif%&ences in the absolute AV obtained for complex dissociation relative t’o the other single-site mutants. As t’he experiment)al errors in the absolutItb AV values within the separate class of single. double and triple mutat’ions are within experimental error, our results suggest, a relatively stable protrill protein interfacial domain. From Figure 4 it. is apparent, there is a significant degree of non-additivity of volume and Gibbs free of complex formation with multiplr energies mutations in the proposed interface domains. Since t’he four salt-bridges in the complex appear t,o define t’he periphery of the protein-protein interaction site, removal of these salt-bridges by mutation ma? allow greater solvation in these areas with respect to t,he wild-t,ype complex. The resulting volumt~ changes t,herefore would be smaller since thta remaining salt)-bridges may already be partially solvated in the mut,ant) complexes. This effect would be increased a,s more salt’-bridges were removed as is evidenced by the data presented in Figure 1. This would also contribute t’o smaller changes in AG as more charges were removed from the intrrfacar allowing greater solvent exposurr. The nhsolutr AC value for the disruption of a salt-bridge depends OII t,he dielectric of the surrounding medium. IYpon greater salvation of the interface, the AG value for t,he dissociation of the remaining salt-bridges therrfore would he signiticantlv less. A similar effect ot solvent on the AV and A(: changes upon intarea,sing mutations would also explain t’hr striking lint)ar c~orrelat~ion obvious in Figure 4. This c:orrelat,ioll reflects t)hr role of &c%rostatics in contributing to changes in both A V and AG and the similar effecsts upon solva,tinp the int,erfac!ial domain of t hc complex. The relationship is also suggest,ire that the complex is in the same relative orientJatiorr with all of t,he various caomplexea with no mtljor st,eric* cahanges upon mutat~ion. Tt also implies another rok for salt-bridges in the csomplex irl that, w-itjh
Mapping
Electrostatic
increasing electrostatic interactions, additional solvent exclusion occurs at the interface resulting in a more stable association. The effect of non-additivity with multiple mutations may also be accounted for by significant changes in the structure of the complex. Upon elimination of three or four charges in the proposed interface, possible formation of multiple configurations of the complex with the interaction of alternate ion-pairs may occur. This would result in a larger volume change than would be expected if the complex orientation remained unchanged. A similar proposal that an alternate complex between bovine cytochrome h, and horse heart cytochrome c may occur at increased pH values has been reported (Mauk et al., 1986). However, it is not likely that the linear correlation between AG and AV would exist’ if significant, changes in complex orientation were to occur upon removal of multiple charges in the proposed interface. Thus, we favor the interpretation that- either an alterat.ion of local dielectric constant due to solvent access or electrostatic interactions between interfacial residues can account for the observed non-additivity of AV and AG with the multiple mutations. However, it is likely that upon elimination of multiple charges in the interface. the cbomplex would be less restrained than in the wildtvpe associat’ion resulting in a greater flexibility &thin the interface. The volume change upon dissociation of the quadruple mutant {E44&, E48&, D6ON, DME} with cytochrome c reflects remaining volume change upon removal of all salt-bridges in the interface. However, our experimental results appear to demonstrate that no major shifts in the configuration of the complex occur upon mutagenesis. Alternative contributions to the remaining AV of the quadruple charge mutations could involve electrostrict,ion of cytochrome c lysine residues buried in the interface. As the interprotein ion-pairs of the proposed complex define the boundaries of the protein-protein interface, mutagenesis of the corresponding residues of cytochrome b, could allow solvat’ion of each respective ion-pair. Thus, the decrease in AV upon subsequent removal of each negatively charged residue proposed to form an ionpair would be due to the solvation of that respective pair including the lysine residues of cytochrome c. A more likely origin for the residual AV observed for the quadruple mutant is the interaction of solvent with non-polar residues that had been buried in the int.erface of the complex. The area of the proteinprot’eih interface sequestered from solvent may be calculated from the solvent-accessible area of the individual proteins versus the modeled cytochrome b,-cyt’ochrome c interaction. Using the method described by Lee & Richards (1971), the area of the in t.erface was calculated to be 800 A’. Approximately half of this value is due to the area of the buried charged side-chains from both proteins which form the salt-bridges previously discussed. Therefore, the remaining 400 A2 of buried residues is due mostly t,o the non-polar residues in the inter-
1459
Interactions
face. It has been estimated that for the interaction of I A2 of non-polar residues with water, a decrease of @l ml/mol in volume is observed (Silva et aZ., 1986). These considerations would predict roughly 40 ml/mol residual volume contribution to the Gibbs free energy of association of the cytochrome b,-cytochrome c complex, which is exactly that observed. In summary, high-pressure techniques coupled with site-directed mutagenesis has permitted us to map the surface-interaction domain of cytochrome b, in its association with cytochrome c. This surface site is in agreement with Salemme’s proposed model for the complex of these electron-transfer partners. We conclude that there is one major orientation of the complex at neutral pH, without a major sampling of multiple orientations on the surface. If exhibited significant t)wothe two proteins dimensional diffusion as suggested by Northrup et al. (1988) for cytochrome c, cytochrome c peroxidase solvation of the surface carboxylate groups of cytochrome hS would tend to equalize any differences in dissociation volumes for various surface charge mutations. Since we observe clear definition of single buried sites in the association of rat liver cytochrome b, and horse heart cytochrome c, we favor a model of a defined surface-interaction domain with only rapid nutational dynamics that leaves the surfaces mainly sequestered from solvent access. Since complete removal of the four carboxylate groups on the cytochrome h, surface proposed to be in the interfacial domain reduces the AC: value of association by only 14%. electrostatics do not appear t)o contribute significantly to the stabi!ity of complex formation. The major role of electrostatics in the formation of the protein-protein complex is t,o provide a mechanism for controlling solvent access to the interface, perhaps providing for a more restrained orientation of the complex, and the resulting configuration of the donor and acceptor prosthetic groups. Clearly, high-pressure techniques in concert with site-directed mutagenesis can be used to define other protein-protein interaction sit,es where electrostatics are proposed to play a role in the association. This research was supported by grants Pu’ational Institutes of Health, PHS (;M33775 GM31756, the Kational Science Foundation Research Laboratory, and the Biotechnology and Development Corporation. We thank Weber for many useful discussions.
from the and PHS Materials Research
Dr Gregorio
References Beck
von Bodman, S., Schuler, M. A.. Jollie, D. R. Sligar, S. G. (1986). Synthesis. bacterial expression and mutagenesis of the gene coding for mammalian cytochrome b,. Proc. Nat. Acad. Rci.. 11.8.A.
&
83,
9443-9447. Eisenthal, linear
R. & Cornish-Bowden, plot. Biochem. J. 139.
A. (1974). 715-720.
The
direct
1460
Eley,
K. K. Rodgers and S. G’. Nigur
C. G. S. & Moore. C. R,. (1983). ‘H-n.m.r. Investigation of the interaction between cytochrome c and cytochrome b,. Riochem. .I. 215. 11-21. Errnan. .J. E. & Vitello, I,. R. (1980). The binding of cytochrome c peroxidase and ferricybochrome c. J. Biol. Chem. 255, 6224-6227. Fisher, M. T.. White. R. E. & Sligar, S. (:. (1986). Pressure dissociation of a protein-protein electron transfer complex. J. Amer. Chrm. Sot. 108. 68X-6837. Geren. L.. Tuls. J.. O’Brien. P.. Millett. F. $ Peterson. ,J. A. (1986). The involvement of carboxylate groups of putidaredoxin in the reaction with putidaredoxin reductase. J. RioZ. Chem. 261, 15491L1.5495. Heremans, K. (1982). High pressure effects on proteins and other biomolecules. An,nu. Rw. Riophys. Riorng. 11. l-21. Holloway. P. W. bz Mantsch. H. H. (1988). Infrared spectroscopir analysis of salt bridge formation and c,,vt,ochrome C. cytochrome b, between Biochemistry, 27. 7991-7993. Kang, C”. H., Brautigan. I>. L., Usheroff. K. XL Margolaish. E. (1978). Definition of cytochrome c binding domains by chetnical modification. -1. Hiol. (‘hum. 253, 6502.--6510. Kornblatt), ,J. A., Hui Bon Hoa. C:.. Eltis, I,. & Mauk. A. U. (1988). The effec%s of pressure on porphyrin cmcytochrome b, complex format)ion. ./. Amrr. C’hwu. sot. 110, 5909-5911. Lee. 13. & Rirhards. F. M. (1971). The interpretation of protein structures: estimation of stat,ic accessibility. J. Mol. Biol. 55. 379-400. Mathews. F. S.. Levine, M. & Argos. I’. (1972). Three\caalf liver synthesis of Fourier dimensional cyt,ochrome b, at 2.8 .A resolution. ,/. No/. Biol. 64. 449-464. Matthew. ,J. J.. Weber, P. (!.. Salrmme. F. R. XL Ric~hards. F. M. (1983). Electrostatic orientation during electron transfer between flavodoxin and q%ochromcb 301. 169%171. c. Xature (London). Mauk, M. R.. Reid, T,. 8. & Mauk, A. (:. (1982). Spretrophotometric analysis of the interaction between cytochrome 6, and cytochrome c. Riochwzistry, 21. 1843-1846. Mauk, M. R.. Mauk. A. G.. Il’rber, I’. C. & Matthew. .J. 1s. (1986). Ele&ostatir analysis of t,hr interact,ion of cytochrome c with native and dimrthyl ester hemr substituted caytochrome h,. Biochemistr~y. 25. 70857091. McLendon. (:. & Miller. J. R. (1985). The dependence of’ biological electron transfer rates on rxothermicity: the cytochrome c/cytochromr b, couple. .J. Amw. C’hrm. A’oc. 107, 7811-7816. Ng. S., Smith, M. B.. Smith, H. T. & Millet. F. (1977). Effect, of modification of individual cytochrome c lysines on the reaction with cytochrome b,. Biochemistry, 16. 4975-4978. Northrup, S. H., Boles, ,J. 0. & R,eynolds. ,J. (‘. I,. (1988).
Brownian dynamics of cytoc*hrorne c ant1 c~~~tochrotttr c peroxidase association. Scierbce, 241. BS--70. Paladini. A. A. Ki Wrbrr. (:. (1981). l’ressur,r-itttlrtc,~(l revrrsi ble dissociation of enolase. Riwhrmistry. 20. 2587 -2593. Poulos. T. I,. Hr Kraut. .I. (19X0). A hypottrt~tical trtotlrl of’ the cytochromr r peroxidase. cyt~ochromr r elect rot1 t,ransfer c~omplrx. ./. Hiol. C’hrm. 255. 10322~ 1OX%). Poulos. T. I,. CG !Mauk. A. (G. (1983). Models for tht c~omplrxt~s formed between q%ochromt~ OS and tttcs subunits of’ met~hrmoglobin. ,/. 12irrl. f ‘SPIN. 258. 7369 737:%. Qin. I,., Rodgers. K. K. & Niger. S. (:. (I!)!)1 ). b:lec+rort bransfer between c*ytochromr 11, surfacr tnutants ant1 qtjoc*hromtl C. Nol. (‘rystallogr. Liq. ( ‘rysttrllo~yr. 194. 3 I I 3 I 6. Reid. 1,. S.. Mauk. M. R. & Mauk. A. (:. (19X-C). IColr trf hrtnr propionate groups in c*ytocahrotntL bs rlcc*trotr transfer. ./. .-Zrnc~r. C’h~n/. SW. 106, 2182 AlX.5. Rodgers. K. K.. I’oehapsky. T. C’. CVSligar. S (:. (I!+?++) rI1N~~1.011101~‘(‘IIlat of Probing the tnec*hanisms i’ recognition: tht, c~yt>oc~hrornt~ bs~c.~toc,hrotn~~ c~omplex. Scirr~~r. 240. I657 1659. Salrmme. F. R. (1976). A hypothetical structurt. for, att t+rt rot1 tritnstrt c~omplex intrrtnolr~ular trf caytocahromrs c and h,. .J. Xol. Hiol. 102. 563&.56X. Silva. .I. I,.. Miles. E. \V. k \Veber. (Z. (1!186). I’ressttrt, disso?iatGon and c~otiforrnationr~l drift of thr /&tlinitJr of t,ryptophan synthase. Rioch.umistr!y. 25. 5780 57%. Stayton. I’. S.. Fisher. M. T. & Sligar. 8 ( :. (198X). of c,ytoc.hromt* h, Drtrrniination iissociatiotr rrac+iotls. ./ Uir,/. (‘//PN 263. I4544 I354X. Stagton. 1’. S.. l’otrlos. ‘I’. I,. & Sligar. 6. C:. (I!N!)). I’utidarrdoxin c~ornprtitively inhibitas c~,vtoc.hromr b,c~~to~hromr P-4.50,,, t+ctron-tr’arisfpr c~trnrltlt~s. Bior:hrm istry. 28. x20 I ~8205. Takano. T.. Kallai. 0. K.. Swanson. Ii. & IGkrrsott. IC. b:. (197R). The strucaturr of ferroc~~t’oc.hrornr C’ at A.45 .q resolution. ./. Hiol. (‘hem. 248. X34 :i%T, Tamburini. 1’. I’.. IlThitr. K. E. & Schrnkman. .I. Ii. (1985). (‘htkmic~l (.hat,actrrizat,ion of protein Itrcttein t,etwern ey tochrome P-450 antI intrrac.tions c~ytoc~hromt~ b,. ./. Hiol. C’hrm. 260, 5007 4Ol5. Wrber. (:. (1987). I)issociation of oligomeric~ proteitts I,>, hydrostatic, pressurt’. In High I’re.ss./~w ( ‘h~mistr!/ rrrrtl Niochrnrisfry (van Eldik. R. Ki .lonas. ,I.. r&). pp. 401 420. I). Keidt~l. New l’ork. H. ( 1!18:1). Thtn ~+fA.t 01’ itiph \Vebrr. (:. k I)rickamrr. am1 other I~iotrrolrc~t~l~~s pre‘ssurp rtlton proteins @curt. tfrr. Hiop/ly.v. 16. R-1 1". Wendoloski. .I. .J.. Matthew. .J. 13.. \Yrl,rr. I’. (I S Salernrnr. 1’. R. (1987). Molrc~ular dyriami~s of it c-c,?-tochromr h, t~lrctron tratisf+r caytochromcx c*oml)lt,x. Sciprac~. 238. 7!14&‘i97. Zamyatnin. A. .A. (1972). Protein volumes itt solution. J’roq. Rioph!/.u. Xol. Bid. 24. 107. 113.
Edited by P. Wright
