Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Structural influences on reduction potential, electron transfer and stability in cytochrome c Lo, Terence Pui-Kwan 1995

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_1995-983159.pdf [ 5.49MB ]
Metadata
JSON: 831-1.0088865.json
JSON-LD: 831-1.0088865-ld.json
RDF/XML (Pretty): 831-1.0088865-rdf.xml
RDF/JSON: 831-1.0088865-rdf.json
Turtle: 831-1.0088865-turtle.txt
N-Triples: 831-1.0088865-rdf-ntriples.txt
Original Record: 831-1.0088865-source.json
Full Text
831-1.0088865-fulltext.txt
Citation
831-1.0088865.ris

Full Text

STRUCTURAL INFLUENCES ON REDUCTION POTENTIAL,ELECTRON TRANSFER AND STABILITY IN CYTOCHROME CByTerence Pui-Kwan LoB.Sc., The University of British Columbia, 1988A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFDOCTOR OF PHILOSOPHYinTHE FACULTY OF GRADUATE STUDIESTHE DEPARTMENT OF BIOCHEMISTRY AND MOLECULAR BIOLOGYWe accept this thesis as conformingto the required standardTHE UNIVERSITY OF BRITISH COLUMBIAApril 1995© Terence Piii-Kwan Lo, 1995In presenting this thesis in partial fulfilment of the requirements for an advanced degree at theUniversity of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarlypurposes may be granted by the head of my department or by his or her representatives. Itis understood that copying or publication of this thesis for financial gain shall not be allowedwithout my written permission.The Department of Biochemistry and Molecular BiologyThe University of British Columbia2146 Health Sciences MallVancouver, British ColumbiaV6T 1Z3Date:1 2)AbstractThe work described in this thesis was directed toward understanding the contributions of Phe82,Leu85 and associated residues to the structure and function of cytochrome c. One focus ofthese studies concerned the roles of these residues at the interactive face formed during electrontransfer reactions involving cytochrome c and its redox partners. Phe82 was found to makean important contribution to this process, while the adjacent Leu85 does not have a significant impact. However, both residues indirectly affect complex formation by influencing theplacement of the side chain of Argl3. In related studies, Leu85 was found to contribute tothe formation of the heme binding pocket as a participant in a cluster of conserved leucineresidues. This leucine cluster, as well as an adjacent internal hydrophobic cavity, were shownto provide a measure of conformational flexibility when mutations are made in cytochrome cand may facilitate the structural transition of this protein between oxidation states. It was alsodetermined that replacement of Leu94, which is also a member of this leucine cluster, did notalter the helix-helix interactions of the N- and C-terminal helices of cytochrome c as had beensuggested by earlier studies. Also apparent from further studies was the important functionalrole played by Phe82 in the regulation of heme reduction potential which could be attributedto its contribution to the nonpolar environment of the heme pocket. Related to this, it wasfound that the effects of polar groups on reduction potential were mitigated when these groupswere shielded from the heme macrocycle by intervening heme substituents. An additional factorin the regulation of reduction potential was characterized as part of the study of Phe82 anddistal residues involved in interactions with heme propionate A. Here it was shown that themechanisms by which Arg3S and Asn52 influence reduction potential differ from that of Phe82.Furthermore, for Arg38 and Asn52, these effects overlap and synergistic effects are observedwhen they are replaced in concert. In contrast, the structural implications of either single orIImultiple substitutions at Arg38, Asn52 and Phe82 were determined to be independent as wasthe impact on structural stability as a whole. Overall it is clear from these studies that Phe82,Leu85, and likely many other residues in cytochrome c, have multiple structural and functionalroles and that the interplay between the roles of these residues are correspondingly complexand difficult to predict.111Table of ContentsAbstract iiTable of Contents ivList of Tables ixList of Figures xiiList of Abbreviations xvAcknowledgments xvii1 Introduction 11.1 Eukaryotic Cytochromes c 11.2 The Structure of Cytochrome c 41.3 The Reduction Potential of Cytochrome c 81.3.1 Factors which regulate reduction potential 121.3.2 Effect of heme pocket polarity on reduction potential 121.3.3 Effect of electrostatic groups on reduction potential 131.4 Cytochrome c in Electron Transfer Complexes 141.4.1 The cytochrome c: cytochrome b5 complex 151.4.2 The cytochrome c: cytochrome c peroxidase complex 171.5 Mutants of Cytochrome c 171.5.1 The role of phenylalanine 82 191.5.2 The role of leucine 85 211.5.3 The roles of related residues 22iv1.6 Research Objectives• 232 Experimental Methods 262.1 General Experimental Approach 262.2 Design of Mutant Cytochromes c 282.3 The Crystalline State 302.3.1 The hanging drop vapour diffusion method 312.3.2 The free interface diffusion method 322.3.3 Conditions for the crystaffization of mutant proteins 332.4 Theoretical Aspects of X-ray Diffraction 372.4.1 Diffraction of X-rays by protein crystals 382.4.2 Calculation of electron density from structure factors 382.4.3 Mutant protein structure determination 392.4.3.1 Isomorphous crystal forms 392.4.3.2 Non-.isomorphous crystal forms 412.5 Practical Aspects of Diffraction Data Collection 452.5.1 Mounting crystals for data collection 452.5.2 Diffractometer data collection and processing 462.5.3 Area detector data collection and processing 472.5.4 Determination of absolute scale 492.6 Refinement of Structural Models 512.6.1 Stereochemically restrained refinement 512.6.2 Simulated annealing refinement 522.6.3 General considerations in refinement of mutant proteins 532.6.3.1 X-ray diffraction data sets 532.6.3.2 Difference electron density maps 562.6.3.3 Course of a typical refinement 572.6.3.4 Calculation of atomic coordinate errors 59V3 Roles of Residues 82 and 85 in Cytochrome c 613.1 Introduction 613.2 Experimental Procedures 633.3 Results 663.3.1 Structure of F82Y cytochrome c 753.3.2 Structure of L85A cytochrome c 783.3.3 Structure of F82Y/L85A cytochrome c 803.4 Discussion 823.4.1 Structural consequences of residue 82 and 85 mutations 823.4.2 Impact of mutations on reduction potential 843.4.3 Electron transfer in mutant proteins 874 Replacement of Conserved Leucines in Cytochrome c 894.1 Introduction 894.2 Experimental Procedures 924.3 Results 974.3.1 Structural comparison of mutant and wild-type cytochromes c 974.3.2 Structure of L85C cytochrome c 1014.3.3 Structure of L8SF cytochrome c 1054.3.4 Structure of L85M cytochrome c 1074.3.5 Structure of L94S cytochrome c 1094.4 Discussion 1094.4.1 Structural consequences of residue 85 and 94 mutations 1094.4.2 Effects on reduction potential 1124.4.3 Functional alterations 1144.4.4 Hydrophobic internal cavity fluctuations 1155 Aliphatic Replacements of Phe82 in Cytochrome c 120vi5.1 Introduction5.2 Experimental Procedures5.2.1 Crystallization and data collection5.2.2 Structure solution for the F82L mutant protein using molecular replacement methods5.2.3 Refinement of mutant protein structures .5.2.4 Direct electrochemistry5.3 Results5.3.1 Structural comparison of mutant and wild-type5.3.2 Structure of F821 cytochrome c5.3.3 Structure of F82L cytochrome c5.3.4 Structure of F82M cytochrome c5.4 Discussion5.4.1 Structural implications5.4.2 Functional effects5.4.3 Effects on reduction potential . .6 Multiple Distal Mutations in Cytochrome c6.1 Introduction6.2 Experimental Procedures6.3 Results6.3.1 Structural comparison of mutant and wild-type6.3.2 R38A mutation site6.3.3 N521 mutation site6.3.4 F82S mutation site6.3.5 The conserved internal water, Wat1666.4 Discussion6.4.1 Structural effectscytochromes c157157159164164166172172177179179120121121125129133134cytochromes c 134135135147150150152153vii6.4.2 Protein stability.1806.4.3 Reduction potential effects 182Summary 187Bibliography 190Addendum 204viiiList of Tables1.1 Alignment of the sequences of yeast iso-i, yeast iso-2, tuna, horse and rice cytochromes c 51.2 Secondary structural elements present in yeast iso-i cytochrome c 62.3 Growth conditions and maximum size of crystals formed by mutant yeast iso-icytochromes C 352.4 Compilation of stereochemistry observed in refined yeast iso-i cytochrome c mutant structures 603.5 Data collection parameters for F82Y, L85A and F82Y/L85A yeast iso-i cytochromes c 643.6 Refinement results and stereochemistry for the F82Y, L85A and F82Y/L85Ayeast iso-i cytochrome c mutant structures 673.7 Overall average positional deviations (A) between wild-type yeast iso-i cytochrome c and the F82Y, L85A and F82Y/L85A mutant proteins 693.8 Heme geometry of F82Y, L85A and F82Y/L85A yeast iso-i cytochromes c . . . 733.9 Heme solvent accessibility in F82Y, L85A, F82Y/L85A and wild-type yeast iso-icytochromes c 743.10 Average thermal factors (A2)of residues 80 through 85 in F82Y, L85A, FS2Y/L85Aand wild-type yeast iso-i cytochromes c 773.11 Solvent accessibility of residues 82 and 85 in F82Y, L85A, F82Y/L85A and wildtype yeast iso-i cytochromes c 783.12 Summary of positional differences observed between the F82Y, L85A and F82Y/L85Amutant yeast iso-i cytochromes c and the wild-type protein 83ix3.13 Reduction potentials for F82Y, L85A, F82Y/L85A and wild-type yeast iso-icytochromes c 864.i4 Data collection parameters for the L85C, L85F, L85M and L94S mutant yeastiso-i cytochromes c 924.15 Refinement results and stereochemistry for the L85C, L85F, L85M and L94Syeast iso-i cytochrome c mutant structures 954.i6 Overall average positional deviations (A) between wild-type yeast iso-i cytochrome c and the L85C, L85F, L85M and L94S mutant proteins 974.17 Heme geometry of L85C, L85F, L85M and L94S yeast iso-i cytochromes c . . . . 1024.18 Heme solvent accessibility in L85C, L85F, L85M, L94S and wild-type yeast iso-icytochromes c 1034.i9 Summary of positional differences observed between the L85C, L85F, L85M andL94S mutant yeast iso-i cytochromes c and the wild-type protein lii4.20 Reduction potentials for L85C, L85F, L85M, L94S and wild-type yeast iso-icytochromes c 1i34.21 Surface area and volume of the hydrophobic internal cavity in mutant and wild-type yeast iso-i cytochromes c 1175.22 Data collection parameters for F821, F82L and F82M yeast iso-i cytochromes c . 1225.23 Refinement results and stereochemistry for the F821, F82L and F82M yeast iso-icytochrome c mutant structures 1325.24 Overall average positional deviations (A) between wild-type yeast iso-i cytochrome c and the F821, F82L and F82M mutant proteins 1345.25 Heme geometry of F821, F82L, F82M and wild-type yeast iso-i cytochromes c. . i405.26 Heme solvent accessibility in F821, F82L, F82M and wild-type yeast iso-i cytochromes c i4i5.27 Reductioii potentials for F821, F82L and F82M yeast iso-i cytochromes c . . . . i55x6.28 Data collection parameters for yeast iso-i cytochromes c with multiple distalmutations i596.29 Refinement results and stereochemistry for the structures of yeast iso-i cytochromes c with multiple distal mutations i626.30 Overall average positional deviations (A) between yeast iso-i cytochromes c withmultiple distal mutations and the wild-type protein i646.3i Heme geometry of yeast iso-i cytochromes c with multiple distal mutations . . . i696.32 Heme propionate hydrogen bond interactions in yeast iso-i cytochromes c withR38A, N521 and F82S replacements i706.33 Wati66 hydrogen bond interactions in yeast iso-i cytochromes c with R38A,N521 and F82S replacements 1736.34 Heme solvent accessibility in yeast iso-i cytochromes c with R38A, N521 andF82S replacements 1766.35 Unfolding of mutant and wild-type yeast iso-i cytochromes c by guanidine hydrochioride i8i6.36 Reduction potentials for yeast iso-i cytochromes c with R38A, N521 and F82Sreplacements i83xiList of Figures1.1 The polypeptide fold of yeast iso-i cytochrome c . 2i.2 Schematic of heme showing atom labeling convention 31.3 Stereo diagrams of the structure of wild-type yeast iso-i cytochrome c 7i.4 The exposed heme edge of yeast iso-i cytochrome c 9i.5 Hydrogen bonds involving internal water molecules and heme propionate A inyeast iso-i cytochrome c 101.6 A conserved leucine cluster in yeast iso-i cytochrome c 11i.7 The cytochrome C: cytochrome b5 complex 16i.8 The cytochrome C: cytochrome c peroxidase complex i81.9 The location of yeast iso-i cytochrome c residues examined by a structure-function approach in the present study 202.10 General scheme used to study structure-function relationships in yeast iso-i cytochrome c 272.11 Schematic diagram of the hanging drop vapour diffusion method 322.i2 Schematic diagram of the free interface diffusion method 332.13 Difference Fourier synthesis of F82Y, L85A and F82Y/L85A yeast iso-i cytochrome c mutant structures 422.14 Wilson plot for the R38A/N521/F82S yeast iso-i cytochrome c data set 502.15 Course of structural refinement of N521/F82S yeast iso-i cytochrome c 583.16 Luzzati plot of the F82Y, L85A and F82Y/L85A yeast iso-i cytochrome c mutantstructures 68xli3.17 Stereo diagrams of the o-carbon backbones of the wild-type, F82Y, L85A andF82Y/L85A iso-i cytochrome c structures 703.i8 Average positional deviations from the wild-type iso-i cytochrome c structurefor the F82Y, L85A and F82Y/L85A mutant proteins 713.i9 Stereo diagrams of the region about Tyr82 in F82Y iso-i cytochrome c 763.20 Stereo diagrams of the region about A1a85 in L85A iso-i cytochrome c 793.21 Stereo diagrams of the region about the mutated residues in F82Y/L85A iso-icytochrome c 814.22 Luzzati plot of the L85C, L85F, L85M and L94S yeast iso-i cytochrome c mutantstructures 964.23 Stereo diagrams of the o-carbon backbones of the wild-type, L85C, L85F, L85Mand L94S iso-i cytochrome c structures 984.24 Average positional deviations from the wild-type iso-i cytochrome c structurefor the L85C, L85F, L85M and L94S mutant proteins 994.25 Stereo diagrams of the region about Cys85 in L85C iso-i cytochrome c 1044.26 Stereo diagrams of the region about Phe85 in L85F iso-i cytochrome c i064.27 Stereo diagrams of the region about Met85 in L85M iso-i cytoclirome c 1084.28 Stereo diagrams of the region about Ser94 in L94S iso-i cytochrome c 1104.29 A stereo diagram of the internal hydrophobic cavity in iso-i cytochrome c . . . . 1165.30 Luzzati plot of the F821, F82L and F82M yeast iso-i cytochrome c mutant structures 1245.31 Fast rotation function search for F82L iso-i cytochrome c 1275.32 Translation function search for F82L iso-i cytochrome c 1285.33 Stereo diagrams of the a-carbon backbones of wild-type, F821, F82L and F82Miso-i cytochrome c structures 136xm5.34 Average positional deviations from the wild-type iso-i cytochrome c structurefor the F821, F82L and F82M mutant proteins i375.35 Stereo diagrams of the region about 11e82 in F821 iso-i cytochrome c 1395.36 Stereo diagrams of the region about Leu82 in F82L iso-i cytochrome c i425.37 Packing of the eight molecules within the unit cell for the F82L cytochrome ccrystal 1435.38 Differences in thermal factors between wild-type yeast iso-i cytochrome c andthe F821, F82L and F82M mutant proteins 1455.39 Omit difference electron density map of Leu82 in F82L cytochrome c i465.40 Stereo diagrams of the region about Met82 in F82M iso-i cytochrome c 1486.4i Luzzati plot of the structures of yeast iso-i cytochromes c with multiple distalmutations 1636.42 A stereo diagram of the or-carbon backbones of the wild-type and combinatorialmutant iso-i cytochrome c structures 1656.43 Average positional deviations from the wild-type iso-i cytochrome c structurefor the mutant proteins with multiple distal mutations 1676.44 Stereo diagrams of the R,38A mutation site in yeast iso-i cytochrome c 1716.45 Stereo diagrams of the N521 mutation site in yeast iso-i cytochrome c 1746.46 Stereo diagrams of the F82S mutation site in yeast iso-i cytochrome c i756.47 Differences in thermal factors between wild-type and combinatorial mutant yeastiso-i cytochromes c 178xivList of AbbreviationsB isotropic thermal factorCm concentration at midpoint of unfolding transitionDTT dithiothreitolEm midpoint reduction potentialGdn-HC1 guanidine hydrochlorideMW molecular weightNMR nuclear magnetic resonanceVm volume per unit massa, b, c crystallographic unit cell axes, or axis lengthsd distanceaverage positional deviatione.u entropy units; 1 e.u. = 1 cal mol’ K—’r.m.s room mean squarea, , crystallographic unit cell angles; also, Eulerian angles in molecular replacement rotation functionionicstrengthrotational axes which define in part the orientation of the crystalin the diffraction experimentA Angstrom (0.1 nm)The conventions of the IUPAC—IUB Combined Commissions on Biochemical Nomenclatureare followed for both three letter and one letter abbreviations for amino acids [J. Biol. Chein.xv241, 527—533 (1966); J. Biol. Chem. 243, 3557—3559 (1968)1; and for designating atoms anddescribing the conformational torsion angles of the polypeptide chain [J. Biol. Chem. 245,6489—6497 (1970)]. Designations for atoms of the protoheme IX group are according to theBrookhaven National Laboratory Protein Data Bank (Bernstein et al., 1977; see also Figure 1.2).The amino acid numbering scheme used for the yeast iso-i cytochrome c mutant structuresdescribed in this thesis is based on an alignment with the sequences of vertebrate cytochromes c(Table 1.1). The N-terminal residue is numbered -5 and the C-terminal residue is numbered103.xviAcknowledgmentsI would like to acknowledge several people who have contributed to the success of these studies.I would like to thank my supervisor, Gary Brayer, for his guidance throughout the studiesdescribed in this dissertation. I am grateful to Grant Mauk and Ross MacGiffivray for theiradvice as members of my Ph.D. supervisory committee and their careful reading of this thesis. Iwould like to acknowledge Michael Smith for his ongoing contributions to the protein engineeringof yeast iso-i cytochrome c. Assistance in these studies was also provided by past and presentmembers of the laboratories of Michael Smith, Grant Mauk, George McLendon, Fred Shermanand Gary Brayer. I would especially like to thank Guy Guillemette with whom I collaboratedon many of the studies described in this thesis. The Medical Research Council of Canada andthe Protein Engineering Network of Centers of Excellence provided the resources without whichthis work would not have been possible.Finally, I would like to thank all of my friends who supported me during these six years;you know who you are. It’s been fun. Twenty-eight years on my wayxviiChapter 1Introduction1.1 Eukaryotic Cytochromes cEukaryotic cytochromes c have been a major focus of studies to elucidate the structural factorsresponsible for controffing biological electron transfer and as such, these small soluble proteins(MW 13,000 daltons) have been extensively examined (for reviews of this work see Timkovich,1979; Mathews, 1985; Moore & Pettigrew, 1990; Brayer & Murphy, 1994). Each of theseproteins contains a heme functional group which is covalently attached to the polypeptide chainthrough two thioether linkages to cysteine side chains (Figure 1.1). These two cysteine residuesare located within a Cys-X-Y-Cys-His sequence which is characteristic of c-type cytochromes.The histidine side chain in this sequence and that of an additional methionine residue act asthe two axial ligands to the iron atom of the heme prosthetic group (Figure 1.1). It is thisheme moiety (Figure 1.2) which bestows upon cytochrome c the ability to mediate the transferof electrons between proteins through the reversible cycling of this group between the ferric{Fe(III)] and ferrous [Fe(II)] states.Cytochrome c is located within the intermembrane space of mitochondria where it functionsprimarily as an electron shuttle in the respiratory electron transport chain by accepting electrons from the cytochrome c reductase (bc1) complex and donating these to the cytochrome coxidase (aa3) complex (Pettigrew & Moore, 1987). The soluble nature of cytochrome c allowsit to diffuse between these two complexes, both of which are multi-subunit integral membraneproteins. In addition to this primary function, cytochrome c also plays a role in several otherelectron transfer pathways (Pettigrew & Moore, 1987). In animal systems, cytochrome c servesas au acceptor of electrons from both cytochrome b5 (Ito 1980a,b; Lederer et aL, 1983) and1Chapter 1. Introduction 2Figure 1.1: A ribbon representation of the polypeptide fold of yeast iso-i cytochrome c. Aball-and-stick representation is used to show the heme group as well as the two thioetherlinkages to Cysi4 and Cysi7 and the two axial heme ligands to Hisi8 and Met8O.Chapter 1. Introduction 3CBCFigure 1.2: A schematic diagram of iron protoporphyrin IX with all non-hydrogen atoms shownas grey balls. These atoms and the four pyrrole rings of the heme macrocycle are labeledaccording to the conventions of the Protein Data Bank (Bernstein et aL, 1977). In eukaryoticcytochromes c, this porphyrin ring is covalently attached to the protein through two thioetherlinkages (Cysl4 SG - heme CAB; Cysl7 SG - heme CAC) and the iron atom is hexacoordinate,with the four pyrrole nitrogen atoms, Hisl8 NE2 and Met8O SD providing the ligands. In theview presented, the solvent exposed heme edge observed in cytochromes c is located at theright, the histidine ligand is behind the porphyrin ring and the methionine ligand is positionedin front.CABCHCCMDCAAO1A02DChapter 1. Introduction 4sulfite oxidase (McLeod et at., 1961; Cohen & Fridovich, 1971) while in yeast systems, cytochrome c donates electrons to cytochrome c peroxidase (Altschul et at., 1940; Abrams et at.,1942) and accepts electrons from fiavocytochrome b2 (Bach et at., 1942a,b). Thus an importantaspect of the mechanism by which cytochrome c mediates electron transfer is how this proteinis capable of interacting with such a wide variety of electron transfer partners.1.2 The Structure of Cytochrome cThe amino acid sequences of cytochromes c from a variety of eukaryotic organisms have beendetermined and the overall length of the polypeptide chain is found to vary from 103 to 113residues. A high degree of conservation is observed in the alignment of the 94 sequences available(those of Euglena gracilis and Tetrahymena pyriformis were excluded due to gaps in theirsequences), with 27 residues being invariantly conserved and a further 16 amino acids beinghighly conserved (Hampsey et at., 1986, 1988; Moore & Pettigrew, 1990). Table 1.1 containsthe aligned sequences of S cytochromes c of particular relevance to the studies conducted inthis thesis and indicates those residues that are found to be invariant or highly conserved in allsuch proteins. As one might expect, many of these invariant residues are absolutely requiredfor the structural and functional integrity of cytochrome c. Nevertheless, a surprising resultis that amino acid substitutions at some invariant sites have resulted in mutant cytochrome cproteins which are functionally viable (Pielak et at., 1985; Hampsey et at., 1988). This raisesquestions concerning the necessity of having these particular residues in cytochrome c and thecause of their observed invariant conservation.The three dimensional structures of several eukaryotic cytochromes c have been determinedto atomic resolution by X-ray crystallographic methods. These include the two isozymes ofcytochrome c from yeast, iso-i (Louie & Brayer, 1990; Berghuis & Brayer, 1992) and iso-2(Murphy et at., 1992), as well as the cytochromes c from tuna (Takano & Dickerson, 1981a,b),horse (Bushnell et at., 1990), rice (Ochi et at., 1983) and bonito (Tanaka et at., 1975). Morerecently, the structures of yeast iso-i and horse cytochrome c have also been studied by NMRChapter 1. Introduction 5Table 1.1: Alignment of the sequences of yeast iso-i, yeast iso-2, tuna, horse and rice cytochromes c-9 1 10 20. . . 0 •••o 0 0 •Iso-iIso-2 AKESTGFKPGSAKKGATLFKTRCQQCHTIEEGGPNKVGTuna --GDVAKGKKTFVQKCAQCHTVENGGKHKVGHorse GDVEKGKKIFVQKCAQCHTvEKGGKHKTGRice -ASFSEAPPGNPKAGEKIFKTKCAQCHTVDKGAGHKQG30 40 50 60•o. • o. 0 o.Iso-i PNLHGIFGRHSGQAEGYSYTDANIKKNvLWDENNMs EIso-2 PNLHGIFGRHSGQVKGySYTDANINKNVKWDEDSMS ETuna PNLWGLFGRKTGQAEGYsYTIJANKS KG I VWNNDTL MEHorse P N L H G L F G R K T G Q A P G F T Y T D A N K N K G I T W K E E T L M ERice PNLNGLFGRQSGTTPGYSYSTANKDMAVIWEENTLyD70 80 90 100•• O...oo..o.. • • • 00 00Iso-i YLTNPJ KYIPGTKMAFGGLKKEKDRNDL I TYL KKACE -Iso-2 YLTNPJ KYIPGTKMAFAGLKKEKDRNDL I TYMTKAAKTuna YLENPKKYIPGTKM I FAGIKKKGERQDLVAYL KSATS -HorseYLENPKKYIPGTKM I FAGIKKKTEREDLI AYLKKATNERice YLLNPJKYIPGTKMVFPGLKJPQERADLI SYLKEATS -The sequences of yeast iso-i (Smith et at., 1979), yeast iso-2 (Montgomery et at., 1980), tuna(Kreil, 1965), horse (Margoliash et at., 1961) and rice (Mon & Morita, 1980) cytochromes chave been aligned so as to maximize the structural homology present. The single letter codeis used to identify amino acids and the residue numbering is based on the sequence of tunacytochrome c. Residues 52 and 54 in rice cytochrome c are listed as asparagine and aspartate,respectively, as opposed to the aspartate at position 52 and the asparagirie at position 54reported by Mon and Morita (1980). The present assignments for these residues fit the currentchemical data and are more consistent with other related sequences (Moore & Pettigrew, 1990).Note the single letter code J is used to denote e-N-trimethyl lysines. Those amino acid residuesidentical in all five protein sequences are enclosed in boxes. Residues which are invariant in 94eukaryotic sequences are marked with a filled dot (.) and those which are conserved in at least90 sequences are marked with an open circle (o).Chapter 1. Introduction 6Table 1.2: Secondary structural elements present in yeast iso-i cytochrome cSecondary structural Residueselement involveda-helix 2—14/3-turn (type II) 21—247-turn 27—29/3-turn (type II) 32—35,i3-turn (type II) 35—38/3-turn (type II)t 43—46a-helix 49—55a-helix 60—70a-helix 70—75/3-turn (type II) 75—78a-helix 87—102tmediated through a water molecule.techniques (Moench & Satterlee, 1989; Busse et at., 1990; Gao et at., 1990, 1991; Qi, 1994a,b).All these cytochromes c share a common polypeptide fold, with significant conformationaldifferences occurring primarily in flexible surface ioops (Bushnell et at., 1990). In the followingdiscussion, yeast iso-i cytochrome c is used as an example to describe the general structuralfeatures of these proteins since it has not only been determined to the highest resolution (1.2 A;Louie & Brayer, 1990), but it also forms the basis of the structural studies described herein.The global polypeptide fold of yeast iso-i cytochrome c is ifiustrated in Figure 1.1 andthedetails of the positioning of all main and side chain atoms are shown in Figure 1.3. Thisproteinhas a high a-helical content; a list of residues involved in these and other secondary structuralfeatures is given in Table 1.2. The heme prosthetic group of cytochrome c is found buriedwithin the hydrophobic core of the protein, with only one edge of this group being exposed toChapter 1. Introduction 7ABFigure 1.3: Stereo diagrams of yeast iso-i cytochrome c showing the positions of all main chain(thick lines) and side chain (thin lines) atoms in (a) the standard heme edge-on view and (b) aview rotated by 900 that looks down on the face of the heme plane. The heme grouphas alsobeen drawn with thick lines, along with the two heme ligand bonds to Hisi8 and Met8O andthe two covalent thioether linkages to Cysl4 and Cysl7. Every fifth a-carbon atom has beenlabeled with its one letter amino acid designation and sequence number.Chapter 1. Introduction 8the external solvent environment (Figure 1.4). Also sequestered within the interior of the proteinare the two propionate groups of the heme moiety which participate in an extensive networkof hydrogen bonds with polar side chains and buried water molecules. Two of these watermolecules, Wat166 and Watl2l (Figure 1.5), are conserved among all eukaryotic cytochrome cstructures determined to date (Bushnell t al., 1990; Louie & Brayer, 1990; Murphy et al.,1992; Qi et al., 1994a). Changes in the internal hydrogen bonding network about the hemepropionates are found to accompany the transition of cytochrome c between the reduced andoxidized states (Takano & Dickerson, 1981a; Berghuis & Brayer, 1992; Qi et al., 1994b).Two other notable structural features are present in the hydrophobic core of yeast iso-icytochrome c. The first of these is an internal cavity which is defined by one edge of the hememacrocycle and the side chains of Leu32, 11e35, Met64 and Leu98 (Louie & Brayer, 1990).Although this cavity is large enough to accommodate a water molecule, none is observed, likelydue to a lack of hydrogen bonding partners available in this region of the protein. The secondnotable feature is a cluster of conserved leucine residues, consisting of Leu9, Leu68, Leu85,Leu94 and Leu98, whose side chains pack against each other. This leucine cluster is locatedadjacent to the heme moiety, the internal heme pocket hydrophobic cavity and the intersectionof the N and C-terminal cr-helices of yeast iso-i cytochrome c (Figure 1.6). These latter ahelices pack against each other in a roughly perpendicular manner and are notable in that theyare thought to be involved in the initiation of the folding of this protein (Roder et aL, 1988).1.3 The Reduction Potential of Cytochrome cThe difference in the midpoint reduction potentials of two proteins represents the driving forcefor the transfer of an electron between them. Among eukaryotic cytochromes c, the value of themidpoint reduction potential is highly conserved (r.270 + 20 mV; Mathews, 1985; Pettigrew& Moore, 1987), likely as a functional requirement for optimal transfer of electrons both toand from physiological redox partners. A significant alteration of the reduction potential ofcytochrome c would result in a change in the driving force of its electron transfer reactions andChapter 1. Introduction 9Figure 1.4: A space filling representation of yeast iso-i cytochrome c showing the exposure ofthe edge of the heme prosthetic group (in black) to the external environment. Shown in greyare the side chains of lysine and arginine residues thought to be important in the formation ofelectron transfer complexes between cytochrome c and its redox partner proteins.Chapter 1. Introduction 10Figure 1.5: A diagram showing the hydrogen bonds formed (dotted lines) between three internal water molecules, heme propionate A and several conserved amino acids in yeast iso-icytochrome c. The overall polypeptide chain is represented as a ribbon, whereas a ball-and-stickrepresentation is used to illustrate the heme group (light shading) and the side chains of Arg38,Asn52, Tyr67, Thr78 and Phe82 (dark shading). The three internal water molecules are shownas larger single spheres.Chapter 1. Introduction iiFigure 1.6: A diagram of a conserved leucine cluster found in the hydrophobic core of yeastiso-i cytochrome c as well as residues forming the outer boundaries of a large internal cavityfound at the back of the heme pocket. Also drawn is the conformation of the central hemegroup. The side chains of leucines 9, 68, 85, 94 and 98 make up the leucine cluster while theside chains of Leu32, 11e35, Met64 and Leu98 and one edge of the heme group define the outerlimits of the internal hydrophobic cavity.Chapter 1. Introduction 12therefore, this would be expected to adversely impact the control and efficiency of such reactions. However, studies of mutant yeast cytochromes c having significantly decreased reductionpotentials have found that this does not inhibit the growth of yeast cells (McLendon et al.,1991; Komar-Panicucci et aL, 1992). Thus these studies have raised questions concerning thefunctional requirement for the maintenance of the reduction potential of cytochrome c.1.3.1 Factors which regulate reduction potentialDue to the biological importance of electron transfer reactions, there has been much interestin understanding the factors which determine the midpoint reduction potentials of proteinsinvolved in these reactions (Moore & Williams, 1977). Towards this end, many studies havefocused on the members of the cytochrome family, which have midpoint reduction potentialsspanning a range of nearly 800 mV (Churg & Warshel, 1986). From this range of reductionpotentials, the identity of the axial ligands to the heme iron atom account for a variation ofi15O mV (Harbury et aL, 1965), while an additional variation of the same magnitude can beattributed to substitutions to the porphyrin ring, such as the thioether linkages in c-type cytochromes (Marchon et al., 1982). However, reduction potentials can vary by nearly ‘500 mVamong cytochromes having identical heme iron ligands and porphyrin ring substituents (Cusanovich et al., 1988). Thus this large variation must arise from the differences between theenvironments provided by the polypeptide matrices of these proteins.1.3.2 Effect of heme pocket polarity on reduction potentialMeasurements of the midpoint reduction potential of iron porphyrin ring groups have beenshown to be dependent on the dielectric constant of the solution in which the experiment isperformed (Kassner, 1972). Thus the polarity of the heme environment likely plays a largerole in determining the reduction potential of a heme protein (Kassner, 1972, 1973; Churg &Warshel, 1986). Nonetheless, proteins having heme environments of similar polarity have beenobserved to have reduction potentials which vary by as much as 300 mV. This variation ofChapter 1. Introduction 13reduction potential has been correlated with the exposure of the heme moiety to the externalsolvent environment (Stellwagen, 1978). Although this would seem to indicate that reductionpotential is dependent on solvent exposure rather than the polarity of the heme binding pocket,it must be noted that the heme groups of the proteins examined had varying axial ligands andporphyrin ring substituents. Furthermore, the notion that reduction potential is also dependenton heme solvent exposure is consistent with the idea that the polarity of the heme environmentas a whole is the primary determinant of reduction potential since an increase in heme solventexposure in a protein would result in a corresponding increase in the polarity of the hemeenvironment.1.3.3 Effect of electrostatic groups on reduction potentialStudies have shown that there is a relationship between the presence of charged protein groupsand observed heme reduction potential. It has been proposed that this contribution is straightforward, based on the observation of a direct correlation between the global net charge of aprotein and its reduction potential (Rees, 1985). However, this simple correlation has beenshown to be overly simplistic and cannot consistently account for the influence of electrostaticeffects on protein reduction potential (Moore et al., 1986). Despite the apparent complexityof the relationship between electrostatics and reduction potential, it has been shown that particular charged groups can have a specific and appreciable effect on reduction potential. Forexample, it has been demonstrated that the reduction potential of cytochrome c can be alteredin a predictable manner through the chemical modification of lysine residues (Schejter et aL,1982). In another example, the substitution of Arg38 in cytochrome c by uncharged residuesresulted in a decrease in the reduction potential (Cutler et at., 1989). Also a factor in this lattercase are differences in heme propionate charge which have been proposed to lead to midpointreduction potential variations (Moore et at., 1984). Not only does Arg38 have a charged sidechain but it also interacts with the heme propionate A of cytochrome c (Figure 1.5). While itis evident that electrostatic contributions to heme reduction potential can be substantial, it isChapter 1. Introduction 14also clear that much further work needs to be done to elucidate the precise factors involved inthis process.1.4 Cytochrome c in Electron Transfer ComplexesCytochrome c function is also dependent on its ability to form complexes with its electrontransfer partners. The contact surface between two electron transfer proteins must allow forthe formation and dissolution of a viable complex while providing a suitable medium throughwhich electrons can travel between the proteins. Extensive studies have been performed oncytochrome c aimed at determining the nature of the protein surface involved in the formationof such complexes (for reviews see Margoliash & Bosshard, 1983; Mathews, 1985; Pettigrew &Moore, 1987). Many of these studies have focused on the lysine and arginine residues whichare located around the exposed heme edge of cytochrome c (Figure 1.4). It appears that thesecharged residues provide much of the impetus for complex formation and determine the specificconfigurations of these complexes through the formation of intermolecular electrostatic interactions (Ferguson-Miller et al., 1979). However, important contributions to the stabifizationof such electron transfer complexes also come from hydrophobic surface residues (Mauk et iii.,1986; Guillemette et al., 1994) and other surface groups capable of forming hydrogen bonds(Mauk et aL, 1991).Computer generated models of complexes formed between cytochrome c and various electrontransfer partners have been used as one way to understand the relative functional importance ofdifferent regions of cytochrome c. The model complexes formed with cytochrome b5 (Salemme,1976; Mauk et gil., 1986) and cytochrome c peroxidase (Poulos & Kraut, 1980) have beenthe most extensively studied since the three-dimensional structures of both electron transferproteins are known. These models are described separately in the following sections, along withthe recent determination of the structure of the complex of cytochrome c with cytochrome cperoxidase (Pelletier & Kraut, 1992). In other studies, the regions of cytochrome c whichinteract with cytochrome c oxidase and cytochrome c reductase have been mapped (PettigrewChapter 1. Introduction 15& Moore, 1987), but models of these complexes have not been developed since the structuresof these latter two proteins are not known. However, it has been shown that the region aroundthe exposed heme edge of cytochrome c, including several charged residues, interacts withthese two redox partners (Rieder & Bosshard, 1980). Other model complexes studied includethose formed between cytochrome c and flavocytochrome b2 (Tegoni et at., 1993), flavodoxin(Matthew et at., 1983) and plastocyanin (Roberts et at., 1991).1.4.1 The cytochrome C: cytochrome b5 complexThe complex formed between cytochrome c and cytochrome b5 has been the subject of numerous studies examining interprotein electron transfer (for example, see Eltis et at., 1988; Qinet at., 1991; Willie et at., 1992, 1993; Meyer et at., 1993). This was the first protein-proteinelectron transfer complex for which a model was proposed based on the structures of the individual proteins (Salemme, 1976). It was built by matching the complementary electrostaticsurfaces of cytochrome c and cytochrome b5 and then docking the structures of these proteins together. This model complex has been further subjected to theoretical analysis usingelectrostatic (Mauk et at., 1986; Eltis et at., 1991), molecular dynamics (Wendoloski et at.,1987), Brownian dynamics (Eltis et at., 1991; Northrup et a!., 1993) and energy minimization(Guillemette et at., 1994) techniques. Figure 1.7 shows one such enhanced model of the cytochrome c-cytochrome b5 complex. Although the models generated by these techniques likelydo not completely duplicate the ensembles of closely related complexes which occur in solution(Hartshorn et a!., 1987; Burch et at., 1990), they nevertheless provide a general outline of theresidues involved in electron transfer between these proteins. Two such residues implicated bymodeling and which are the focus of the studies in this thesis are those at positions 82 and 85in cytochrome c (Wendoloski et at., 1987; Burch et at., 1990).Chapter 1. Introduction 16Figure 1.7: A model of the complex formed between yeast iso-i cytochrome c (bottom) andbovine cytochrome b5 (top). A ribbon representation is used to depict the two polypeptidebackbones while the heme groups of the two proteins are shown in a ball and stick representation,as are the side chains of Phe82 and Leu85 in cytochrome c. This model represents the mostcommon conformation observed in a Brownian Dynamics simulation of the complex (Northrupet al., 1993); energy minimization was used to optimize intermolecular interactions (Gufflemetteet al., 1994).Chapter 1. Introduction 171.4.2 The cytochrome C: cytochrome c peroxidase complexAnother complex intensively studied to understand the nature of the electron transfer processhas been that formed between cytochrome c and cytochrome c peroxidase (for example, seeErman et at., 1991; Everest et al., 1991; Nocek et at., 1991; Beratan et at., 1992; Corin et at.,1993). This complex has also been the subject of a number of modeling studies (Poulos &Kraut, 1980; Poulos & Finzel, 1984; Lum at at., 1987) and more recently, the structure ofthis complex has been determined by crystallographic methods to 2.3 A resolution (Pelletier &Kraut, 1992). The structure determined for this complex, which is ifiustrated in Figure 1.8, hasproven to be a valuable tool in understanding redox partner complexation. It has also revealedsignificant deficiencies in the earlier model complexes that had been constructed, especially withrespect to the identity of the interactive surface region of cytochrome c peroxidase. However,both this complex structure and earlier modeling suggest that the surface of cytoclirome caround the exposed heme edge is located at the site of intermolecular contact. It should benoted that while the structural determination outlines the general aspects of cytochrome ccytochrome a peroxidase complex formation, this does not preclude the existence of alternatedocking geometries for these two proteins such as those suggested by Brownian Dynamics studies(Northrup et at., 1988). In fact, there is strong evidence that two distinct binding regions forcytochrome a exist on the surface of cytochrome a peroxidase (Stemp & Hoffman, 1993; Zhou& Hoffman, 1993; Mauk et at., 1994; Zhou & Hoffman, 1994).1.5 Mutants of Cytochrome cIn the past, classical mutagenesis methods have provided valuable insight into cytochrome cfunction (Hampsey et at., 1988; Sherman, 1990). A more recently developed technique thatallows considerable extension of these studies is site-directed mutagenesis, which allows therational design of specific mutant proteins (Zoller & Smith, 1983; Smith, 1986). Using thismethod, both the roles of individual amino acids in cytochrome a and more global aspectsof function can be probed. Of particular interest to the work in this thesis are the factorsChapter 1. Introduction 18Figure 1.8: The structure of the complex formed between yeast iso-i cytochrome c (bottom)and yeast cytochrome c peroxidase (top) as determined by Pelletier & Kraut (1992). A ribbonrepresentation is used to depict the two polypeptide backbones. The heme groups of the twoproteins are shown in a bail and stick representation, as are the side chains of Phe82 and Leu85in cytochrome c.Chapter 1. Introduction 19involved in the maintenance of heme reduction potential, the nature of the interactive face ofcytochrome c with electron transfer partners, and the roles of internal structural features suchas hydrophobic cavities and clusters of leucine side chains.Two amino acids that appear to have roles in these functional and structural aspects ofcytochrome c are Phe82 and Leu85. These two residues form the focus of the studies discussed in this thesis, wherein specifically designed mutant proteins having single substitutionsor combinations of substitutions at these and other related positions in the polypeptide chainare made, with the goal being to gain a better understanding of how Phe82 and Leu85 influence the function of cytochrome c. The overall placement of these residues in yeast iso-icytochrome c is shown in Figure 1.9. A brief description of the extent of knowledge concerningthese residues which was available at the onset of this work is given in the following sections.A more complete analysis of structure-function relationships found for Phe82 and Leu85 on thebasis of structural studies conducted as part of this thesis is presented in Chapters 3—6.1.5.1 The role of phenylalanine 82This residue is invariantly conserved in the sequences of cytochrome c and it is believed to havemultiple functional roles. Substitution of Phe82 in yeast iso-i cytochrome c with other aminoacids has a significant effect on the ability of this protein to support growth of yeast on non-fermentable carbon sources, although no such mutation was found to completely abolish thisfunction (Hilgen & Pielak, 1991; Inglis et at., 1991). Mutants of this type also have an effecton measured heme reduction potential (Pielak et at., 1985; Rafferty et al., 1990). Structuralstudies have found that the side chain of this residue forms a part of the hydrophobic hemepocket of cytochrome c, with its aromatic ring packing against the face of the heme group ina coplanar fashion (Takano & Dickerson, 198ia,b; Louie et at., 1988a; Bushnell et at., 1990).Subsequent structural studies of the F82S (Louie et at., i988b) and F82G (Louie & Brayer, 1989)mutants of yeast iso-i cytoclirome c have revealed some aspects of the relationship betweenPhe82 and heme reduction potential, indicating the latter is correlated to the degree of hemeChapter 1. Introduction 20Figure 1.9: A ribbon diagram of the polypeptide fold of yeast iso-i cytochrome c showingthe location of the side chains of Phe82 and Leu85 which form the focus of structural studiesdiscussed in this thesis. Related residues that were also examined include Arg38, Asn52 andLeu94. As part of these studies, all of these amino acids were replaced either singly or in groupsby site-directed mutagenesis techniques.Chapter 1. Introduction 21solvent exposure and the proximity of polar groups.The side chain of Phe82 is also partially exposed to solvent in the surface region of cytochrome c believed to form complexation interactions with electron transfer partners. The importance of this residue in the formation of electron transfer complexes has been suggested bymodeling (Poulos & Kraut, 1980), NMR (Burch et at., 1990) and molecular dynamics methods(Wendoloski et at., 1987), and this has been confirmed by the structure determination of onesuch complex (Pelletier & Kraut, 1992). Figures 1.7 and 1.8 illustrate the putative positioningof Phe82 in complexes of this type. In regard to the functional role of Phe82, mutant proteinswith replacements at this residue show alterations in steady-state electron transfer activity(Pielak et at., 1985; Michel et at., 1989) and intracomplex electron transfer kinetics (Everestet at., 1991; Hazzard et at., 1992).1.5.2 The role of leucine 85Relatively little study has been conducted on. the role of the conserved Leu85 residue, althoughit has been implicated as an element at the interactive face of complexes formed by cytochrome cwith cytochrome c peroxidase (Nocek et at., 1991; Pelletier & Kraut, 1992). On the basis ofmodeling (Salemme, 1976; see Figure 1.7) and NMR studies (Burch et at., 1990), 11e85 in horsecytochrome c appears to play a comparable role in complexes with cytochrome b5. In termsof location, the side chain of Leu85 in yeast iso-i cytochrome c is partially exposed to solvent,packs against the side chain of Phe82, and forms a portion of the hydrophobic heme bindingpocket (see Figure 1.9; Louie & Brayer, 1990). Leu8S is one of a cluster of such residues (Leu9,Leu68, Leu94, Leu98) that is found along the most buried edge of the heme group. Furthermore,this leucine cluster is adjacent to an internal hydrophobic cavity in the heme pocket of yeastiso-i cytochrome c. To this point, no study of the role of this leucine cluster or the nearbycavity has been conducted.Chapter 1. Introduction 221.5.3 The roles of related residuesAs part of studies conducted on Phe82 and Leu85 in this thesis, a number of other nearby orfunctionally related amino acids were examined. One of these is the highly conserved Leu94which is part of the heme pocket leucine cluster that includes Leu85. The side chain of Leu94packs directly against both the side chain of Leu85 and the heme group. Furthermore, thisresidue is located at the intersection point of the N and C-terminal helices of cytochrome c(Figure 1.9), a structural feature believed to be part of an early folding event in this protein(Roder et al., 1988). It has been shown that replacement of Leu94 by alternative amino acidsresults in reduced activity in vivo (Hampsey et al., 1988) and this has been attributed topossible alterations in the interaction between these N and C-terminal helices.In order to study the role of Phe82 in the maintenance of heme reduction potential andits relationship to other residues that are also a factor in this property, proteins with multiplemutations involving Phe82, Asn52 and Arg38 were examined. Arg38 is an invariant residuepositioned adjacent to the buried propionate A group of the heme to which it forms interactions mediated by two water molecules (Figure 1.5). It has been shown that replacement of thisresidue with other amino acids decreases the observed reduction potential of the resultant mutant protein (Cutler et aL, 1989) without causing large conformational shifts in the polypeptidechain (Barker et al., 1991; Thurgood et al., 1991).Mutation of the invariant Asn52 to other residues can also have a significant effect onreduction potential (Burrows et al., 1991; Langen et at., 1992), as well as on the rate of electrontransfer with cytochrome b5 (Whitford et at., 1991) and in vivo function (Mcbendon et a!.,1991). This residue is completely buried in the structure of cytochrome c near the plane of theheme group and the Met8O heme ligand. In yeast iso-i cytochrome c, AsnS2 also forms a partof a hydrogen bonding network to a conserved internal water molecule, Wat166 (Figure 1.5).A particularly novel feature of the mutation of Asn52 to isoleucine is that this acts to suppressthe effects of other mutations that would otherwise disable the function of cytochrome c (Daset at., 1989; Berroteran & Hampsey, 1991). It has been found that this replacement results in aChapter 1. Introduction 23large increase in the thermal stability of cytochrome c (Das et al., 1989; Hickey et aL, 1991) andsubsequent structural analyses have investigated the source of this stabilization (Berghuis et al.,1994a). Both Arg38 and Asn52 are structurally remote from Phe82 and the goal of combiningmutations at these sites is to determine to what extent these residues act, independently or inunison, in maintenance of heme reduction potential and other properties related to the stabilityand function of cytochrome c.1.6 Research ObjectivesThe goals of this thesis were fourfold and were centered on the use of a structure-functionmutagenesis approach in the study of Phe82, Leu85 and associated residues of cytochrome c.The first objective was to investigate the roles of Phe82 and Leu85 as elements at the cornplexation interface formed between cytochrome c and electron transfer partners. Clearly, thisfeature is central to cytochrome c function and there has been considerable interest in definingthe role of these two residues in this process. Further studies were then designed to elucidate thecontributions of Phe82 and Leu85 to other aspects of cytoclirome c function. Thus a second objective was to characterize the nature of an internal hydrophobic leucine cluster to which Leu85belongs, and which forms a substantial portion of the structure of the heme binding pocket.Associated with this work was an analysis of the highly conserved leucine cluster participant,Leu94, which is positioned at the juncture of the interaction of the N and C-terminal helicalregions of cytochrome c. This helix-helix interaction is also an important structural element informing the heme binding pocket. As part of the study of the leucine cluster, the possible roleof an adjacent internal cavity, which is found next to the most buried edge of the heme group,was examined, along with the relationship between alterations in the volume of this cavity andfunctional aspects of cytochrome c.The third objective of this work was to expand our understanding of the factors which governheme reduction potential and particularly the role that Phe82 plays in this regard. Specificallytargeted were aliphatic side chain replacements for Phe82. Also examined in conjunction withChapter 1. Introduction 24these studies were the effects of mutations at Leu85 and Leu94. In a related fourth objective, anattempt was made to place these results in a larger context by looking simultaneously at Phe82and two other distally located residues (Arg38 and Asn52), all of which have a significant role inthe maintenance of heme reduction potential. The goal was to establish the degree of interplaybetween these sites with regards to this property. These studies also provided informationconcerning the contribution of each of these sites to the overall stability of cytochrome c andthe effects on stability of introducing simultaneous mutations at these sites.Overall, the chapters in this thesis are organized in the order of the objectives just discussed.Beyond Chapter 2, which deals with general aspects of experimental procedure, Chapter 3focuses on the interactive complexation face of cytochrome c, Chapter 4 looks at the conservedleucine cluster in cytochrome c and the associated internal heme pocket cavity, Chapter 5documents the influence of Phe82 on reduction potential, and Chapter 6 examines the individualand cumulative contributions of Arg38, Asn52 and Phe82 to heme reduction potential andoverall protein stability.It must be emphasized that although structural analyses are central to the work completedin this thesis, the interpretation of results of this type are greatly assisted and enhanced byassociated functional analyses. For this reason structural studies were planned as part of acomprehensive program of mutagenesis and functional analyses in collaboration with two othergroups. One of these was based at the University of British Columbia (J. G. Guillemette, A.G. Mauk and M. Smith) and focused on the structural and functional roles of residues at theinteractive face of cytochrome c. The second group of collaborators was based at the Universityof Rochester (S. Komar-Panicucci, F. Sherman, G. McLendon), with the focus being interactionsbetween distal residues in cytochrome c and the investigation of the cumulative and synergisticeffects of introducing multiple substitutions of these residues. The efforts of these collaboratorswere directed toward the site-directed mutagenesis of cytochrome c, as well as the functionalcharacterization of the resultant mutant proteins. My role in these collaborative efforts centeredon the initial modeling and design of amino acid substitutions at the residues of interest, theChapter 1. Introduction 25crystaffization and determination of the three-dimensional structures of mutant proteins, andthe detailed analysis of these protein structures with regard to correlating observed changes inthe functional behavior of mutant proteins with their three-dimensional structures. The resultsof the work presented in this thesis have been published as a number of articles in scientificjournals as summarized in the Addendum at the end of the thesis.Chapter 2Experimental MethodsAn overview of the methodology used to study the structures of mutants of yeast iso-i cytochrome c is described in this chapter. The details of each structural determination are given infollowing chapters.2.1 General Experimental ApproachThe general scheme used in this thesis to investigate structure-function relationships in yeastiso-i cytochrome c is shown in Figure 2.10. This process was initiated by the design of mutations which would provide insight into the roles of Phe82 and LeuS5 in particular, as well asassociated residues that were involved in common structural features or related functionalities.The specific goals in studying each individual mutant protein designed are discussed in detailin the introductory sections of each of Chapters 3—6. Once mutant proteins had been expressedand purified, each was subjected to two streams of study. One of these was a series of functionaltests; the other the elucidation of its three-dimensional structure using X-ray diffraction techniques. In the interpretive phase of this work the collected functional and structural data wereassessed as to the impact of the introduced mutation(s) by comparison to the observed datafor wild-type yeast iso-i cytochrome c. In this assessment, a particular emphasis was placedon defining the role of the amino acid(s) which had been replaced. Yeast iso-i cytochrome cis an ideal candidate for this type of study given the existing extensive knowledge base of thestructural and functional attributes of the wild-type protein, as well as the ability to specificallyalter this protein by site-directed mutagenesis.26Chapter 2. Experimental Methods 27Figure 2.10: General scheme used to study structure-function relationships in yeast iso-i cytochrome c.Chapter 2. Experimental Methods 282.2 Design of Mutant Cytochromes cThe design of each mutant yeast cytochrome c was based on a number of factors. These tookinto consideration not only the available structural and functional data available for yeast cytochrome c, but also those of related cytochromes c. Primary considerations were comparisonsof the known amino acid sequences of eukaryotic cytochromes c, the high resolution structuresof yeast iso-i cytochrome c and the related tuna, horse and yeast iso-2 cytochromes c, andavailable data concerning the manner in which cytochrome c forms complexation interactionswith electron transfer partners such as cytochrome c peroxidase and cytochrome b5.In a procedural sense a number of steps were followed in the development of each mutantprotein. Initially, the amino acid of interest was studied in terms of its impact on functionalityand its location within the protein. An analysis was then made of all the available alternativeamino acids which might be substituted to determine which one(s) might most directly beuseful to assess the functional characteristics of interest. Naturally many substitutions wereinappropriate because they would likely lead to unwanted alterations and therefore be of limiteduse in interpreting or understanding the targeted functional property. An example of this isthe substitution of small side chains with large ones leading to substantial steric conflicts thatdisrupt a whole range of functional properties over and above the one of interest. Also avoidedwere substitutions likely to result in substantial polypeptide chain refolding, which would againproduce too many global shifts to make the interpretation of functional data meaningful forthe targeted residue. The goal of choosing substitutions was to find suitable replacementsthat would only alter one particular facet of cytochrome c function and therefore allow a clearassessment of the role of the residue involved. The detailed rationale for designing each of themutant proteins studied in this thesis is documented in the introductory sections of Chapters3—6.An important element in the determination of suitable substitutions was the use of molecular graphics to examine the structural environment of the residue of interest and to visualizepossible replacement residues. Modeling of possible amino acid replacements was based on theChapter 2. Experimental Methods 29high resolution structure of yeast iso-i cytochrome c (Louie & Brayer, 1990). The primary toolused to accomplish this was the program MMS (Dempsey, 1986) running on either a SiliconGraphics 3130 or 4D/340 workstation. In this way amino acid(s) that might be substituted at agiven position could be inserted and adjusted manually to evaluate the structural consequencesthat might arise. This was particularly useful in assessing the potential for substituted residuesto lead to structural conflicts. Amino acid substitutions were also screened as to their potentialeffects with regards to the electron transfer complexes formed by cytochrome c. This was accomplished by examining modeled mutations in the context of models of the complexes formedbetween cytochrome c and cytochrome c peroxidase (Poulos & Kraut, 1980) and cytochrome b5(Salemme, 1976).Another consideration concerning the design of specific mutants of yeast iso-i cytochrome cand the study of their properties should be noted. Specifically the wild-type form of this proteinhas the tendency to form homodimers through the free cysteine residue at position 102 (Bryantet al., 1985; Moench & Satterlee, 1989). This interferes with both functional and structuralanalyses of this protein and a number of investigators have chemically modified the sulfhydrylgroup of CysiO2 to prevent dimer formation (Zuniga & Nall, 1983; Ramdas et al., 1986; Hickeyet al., 1991). However, this approach perturbs the structure of yeast iso-i cytochrome c inits C-terminal region and thereby complicates the analysis of changes induced by mutationsintroduced at other locations. To avoid this problem we have chosen to modify CysiO2 toeither an alanirie or threonine in our analyses of mutants of yeast iso-i cytochrome c. Thesetwo substitutions occur naturally at position 102 in other eukaryotic cytochrome c sequences(Hampsey et at., 1988; Moore & Pettigrew, 1990). Previous studies have shown that the C1O2Aand C1O2T substitutions have a minimal impact on the structural and functional properties ofyeast iso-i cytochrome c (Cutler et at., 1987; Hickey et cii., 1991; Berghuis & Brayer, 1992).It should be emphasized that even though the best efforts go into the design of mutantproteins, often the structural alterations that occur at the site of mutation are more extensivethan anticipated. Furthermore it is not uncommon for amino acid substitutions to induceChapter 2. Experimental Methods 30shifts far removed from the mutation site in ways that are difficult to predict given our currentunderstanding of the factors involved in protein folding and stability. This appears to beparticularly true for cytochrome c where previous studies of mutant proteins have uncoveredunexpected polypeptide chain refolding (Louie & Brayer, 1989), positional shifts in side chainsand internally bound water molecules (Louie et al., 1988b; Berghuis et al., 1994a), and changesin polypeptide chain mobility (Berghuis et al., 1994a,b). In the absence of structural studies,these unforeseen changes would have led to incorrect interpretation of functional data sincethe assumption would have been made that, other than at the mutation site, the structure ofthese mutant proteins had remained the same. For these reasons special diligence was appliedin the studies conducted in this thesis to characterize any unanticipated structural alterationsinduced by mutagenesis so as to achieve the best interpretation of the functional perturbationsobserved.2.3 The Crystalline StateThe utility of using X-ray diffraction techniques in the analysis of the structures of proteinsrequires that these proteins be in the crystalline state. A protein crystal consists of a wellordered three-dimensional lattice of a large number of individual protein molecules packed in asymmetrical manner. Such crystals are held together by intermolecular interactions, otherwiseknown as lattice contacts, formed between neighboring molecules. The fundamental repeatingunit within a protein crystal is the unit cell which contains one or more identical asymmetric units related to each other through defined symmetry operations. The goal of the X-raydiffraction experiment is the determination of the atomic structure of the protein molecule(s)that are within the asymmetric unit.Before structural analyses by X-ray diffraction can proceed, it is necessary to grow suitablecrystals of the protein to be investigated. In this section a general overview of the methods usedto grow diffraction quality crystals of mutants of yeast iso-i cytochrome c is given. In this worktwo major methods of crystallization were used, the hanging drop vapour diffusion and freeChapter 2. Experimental Methods 31interface diffusion approaches, both of which were supplemented by various seeding techniques.For a more detailed discussion of these and other crystallization methods the reader is referredto excellent reviews in McPherson (1982, 1990) and Carter, Jr. (1990).2.3.1 The hanging drop vapour diffusion methodIn this method, an inverted drop containing the protein solution is placed over a well containingthe precipitant solution or “mother liquor”, as illustrated in Figure 2.11. This protein drop contains a low concentration of precipitant whereas the mother liquor contains a high concentrationof precipitant, resulting in the formation of a concentration gradient within the crystaffizationwell. Due to this gradient, the protein drop and the mother liquor equilibrate over time throughthe diffusion of water vapour, causing a gradual increase in the precipitant concentration in theprotein drop as the system reaches equilibrium. As the precipitant concentration in the dropincreases, this solution becomes saturated and the protein comes out of solution, forming eithera precipitate or the desired result, crystals. In order to encourage crystal formation, the proteindrop can be seeded with either macro-crystals (Thaller et al., 1985) or micro-crystals (Leunget al., 1989) which can assist the formation of nucleation sites for crystal growth.A typical hanging drop crystaffization experiment was carried out using a twenty-four wellLinbro plate (Flow Laboratories, McLean VA) to allow one to scan different variables to optimize the conditions for crystaffization. In general these plates permitted two variables to bescreened at one time. Variables studied included type of precipitant, pH, temperature, concentration of protein and precipitant in the protein drop and in the well solution, as well as theeffect of reducing agents and other additives in the protein droplet. In crystallization trials,1 ml of precipitant solution was placed in each well of the test plate and hanging protein dropscontaining 5—20 1d of protein solution (40—100 mg/ml) were dispensed onto coverslips. Coyerslips were pre-treated with Sigmacote (Sigma Chemical Company, St. Louis), a siliconizingagent which prevented the protein drop from spreading out on the glass surface. Silicone greasewas applied to the rim of each well to form an airtight seal between the well and the coverslip.Chapter 2. Experimental Methods 32cover slipprotein solution /we1l_/Is/IIs/_s_____ _/5..’ ‘s___...precipitant ‘‘‘ - -Figure 2.11: Schematic diagram of the hanging drop vapour diffusion method used in thecrystallization of mutant yeast iso-i cytochromes c.In experiments utilizing crystal seeding, seeds were introduced into the protein drop just beforeplacing the coverslip over its well on the Linbro plate. After assembly, each crystaffization platewas further protected for storage by being wrapped in a thin plastic film.2.3.2 The free interface diffusion methodA second method used to grow protein crystals was the free interface diffusion technique(Salemme, 1972). The setup for this method is illustrated in Figure 2.12 and involves placing a layer of protein solution on top of a layer of precipitant solution within a small diametercapillary tube and allowing these solutions to slowly mix by diffusion through the interfaceformed between these two layers. The main difference between this method and the hangingdrop method is that the protein solution and mother liquor are in direct contact with one another and gradually mix across the contact interface. A consequence of this direct mnbdng is thatprotein initially encounters a supersaturating concentration of precipitant upon introductioninto the capillary tube and this favors the formation of crystal nucleation sites at the interfacebetween solutions. The largest crystals are often found to grow at this interface.A typical free interface diffusion experiment was carried out using a capillary tube withChapter 2. Experimental Methods 33capillary tubeprotein solution-.precipitant solutionFigure 2.12: Schematic diagram of the free interface diffusion method used in the crystallizationof mutant yeast iso-i cytochromes c.an inner diameter of 1.5 mm. Between 30—50 jiL of precipitant solution was injected into thecapillary with a syringe followed by the introduction of a layer of 5—10 jiL of the protein solutiondirectly on top of the precipitant solution. Manual centrifugation was employed to force thesolutions to the bottom of the capillary tube and then the top of the tube was sealed using abunsen burner. In experiments that utilized supplementary crystal seeding, these seeds wereintroduced into the protein solution layer before sealing the capillary tube. Capifiary tubeswere stored upright in styrofoam trays.2.3.3 Conditions for the crystallization of mutant proteinsYeast iso-i cytochrome c mutant protein crystals grown using the hanging drop vapour diffusionmethod used high concentrations of ammonium sulphate ((NH4)2S0as the precipitant agentinducing crystaffization. In a process referred to as “salting out”, water molecules preferentiallysolvate salt ions rather than the protein molecules as the concentration of the salt ions isChapter 2. Experimental Methods 34increased. Under these conditions, the protein molecules either precipitate out of solutionor, if the conditions are favorable, form crystals where lost protein-solvent interactions arereplaced with ordered protein-protein interactions. In a typical hanging drop experiment thestarting concentration of ammonium sulphate was r’.7O% saturation in the droplet and ‘“9O%saturation in the reservoir. Also essential for crystaffization of cytochrome c was the additionof a reducing agent such as dithiothreitol (DTT) or sodium dithionite (Na2SO4)to keep thisprotein in the reduced state. A summary of conditions under which crystals were obtained fordiffraction studies is shown in Table 2.3. In all cases, crystallizations were carried out at roomtemperature (.-.‘20°C), at which the best crystals were found to grow.The use of the free interface diffusion method for crystallizing yeast iso-i cytochrome cmutants employed conditions similar to those used in the hanging drop method, with ammoniumsulphate being the crystallizing agent. However no precipitant was initially present in the layercontaining the protein solution since the ammonium sulphate from the precipitant solution wasable to diffuse directly into the protein solution. In the case of the F82L cytochrome c mutantprotein, crystals could only be obtained for diffraction analysis by the free interface diffusionmethod. The conditions used are listed in Table 2.3 and involved a starting precipitant solutionof 75% saturated ammonium sulphate.Extensive scanning of crystaffization conditions was required to achieve the growth of diffraction quality crystals of yeast iso-i cytochrome c mutants. As shown in Table 2.3, the concentration of ammonium sulphate used to induce crystallization varied considerably (75%—94%)and the optimum pH varied from 5.3 to 6.5. In addition to the wide variance of potential conditions, crystallization trials were hampered by the narrow range of conditions under which eachindividual mutant protein would form satisfactory crystals. As a rule, variation of ammoniumsulphate concentration by as little as 2% or the pH of the protein droplet by 0.1 units would besufficient to result in precipitation of the protein rather than the formation of usable crystals.This is in contrast to other systems such as that of mutants of bacteriophage T4 lysozymewhich has a higher tolerance for variations in crystallization conditions (Brennan et al., 1988).Chapter 2. Experimental Methods 35Table 2.3: Growth conditions and maximum size of crystals formed by mutant yeast iso-icytochromes cIso-i cytochrome c Crystaffization conditions Crystal sizemutant (mm)R38A/N521 (C1O2A) 88% (NH4)2S0;70 mMNa2SO4; 0.4x0.3x0.050.1 M sodium phosphate, pH 6.5R38A/N521/F82S (C1O2A) 88% (NH4)2S0;70 mMNa2SO4; 0.5x0.4x0.20.1 M sodium phosphate, pH 6.5R38A/F82S (C1O2A) 88% (NH4)2S0;70 mMNa2SO4; 0.15x0.05x0.020.1 M sodium phosphate, pH 6.5N521/F82S (C1O2A) 88% (NH4)2S0;70 mMNa2SO4; 0.5x0.13x0.10.1 M sodium phosphate, pH 6.5F821 (C1O2T) 85% (NH4)2S0;150 mM NaC1; 0.17x0.17x0.0570 mMNa2SO4;0.1 M sodium phosphate, pH 6.4F82L (C1O2T) -, - 75% (NH)2SO; 70 mMNa2SO4; 0.32x0.15x0.130.1 M sodium phosphate, pH 5.6F82M (C1O2T) 90% (N114)2S0;140 mMNa2SO4; 0.15x0.08x0.050.1 M sodium phosphate, pH 6.2F82Y (C1O2T) 90% (NH4)2S0;20 mM DTT; 0.5x0.4x0.10.1 M sodium phosphate, pH 6.0F82Y/L85A (C1O2T) 90% (NH4)2S0;30 mM DTT; 0.4x0.4x0.10.1 M sodium phosphate, pH 5.3L85A (C1O2T) 94% (NH)SO; 80 mMNa2SO4; 0.5x0.43x0.150.1 M sodium phosphate, pH 5.5L85C (C1O2T) 90% (NH4)2S0;70 mMNa2SO4; 0.3x0.28x0.i0.1 M sodium phosphate, pH 6.4L85F (C1O2T) 92% (NH4)S0;70 mM DTT; 0.28x0.28x0.10.1 M sodium phosphate, pH 6.3L85M (C1O2T) 92% (NH4)S0;70 mM DTT; 0.58x0.35x0.130.1 M sodium phosphate, pH 6.2L94S (C1O2T) 92% (NH4)S0;70 mMNa2SO4; 0.43x0.43x0.050.1 M sodium phosphate, pH 6.5Chapter 2. Experimental Methods 36Further complicating the growth of suitable crystals for diffraction experiments was therequirement for seeding with micro-crystals. In almost all cases, suitable crystals could not beobtained without the introduction of seeds by a hair-seeding technique (Leung et al., 1989). Inthis method, yeast iso-i cytochrome c crystals were crushed in a solution of mother liquor toproduce microscopic seed crystals. A human hair was passed through this seed solution andthen through the protein solution used in the crystal growth trial. Even though the microcrystaffine seeds introduced into the protein drop by this process were too small to be observedunder a microscope, this process proved critical in initiating nucleation sites for crystal growth.A further major advantage of this method is that the resultant crystals of mutant proteins wereisomorphous to crystals of the wild-type protein, greatly simplifying the determination of thestructures of these mutant proteins.Even with a knowledge of the general conditions required for the crystallization of iso-icytochrome c mutant proteins and the application of the hair-seeding method to assist thisprocess, the growth of diffraction quality crystals was neither straightforward nor guaranteed.This was reflected in the large variability in time required to grow diffraction quality crystalsof these proteins. Although in optimal cases crystals appeared two weeks after initiation of thecrystallization experiment and grew to full size after a further two weeks, the time requiredfor the appearance of crystals could be as much as one year. In the extreme case of the F82Lmutant protein, crystals of sufficient quality appeared only after nearly 2 years of growth.It was also necessary to grow crystals large enough to produce acceptable experimentaldiffraction intensities, particularly for those mutant proteins studied using the laboratory EnrafNonius CAD4 diffractometer which was equipped with only a sealed beam X-ray tube. However,many of the mutant proteins studied yielded only very tiny crystals which could not be used forX-ray diffraction analysis. Even after repeated attempts to increase crysti size by successiverounds of micro- or macro-seeding with these very tiny crystals, oniy small crystals could beobtained for structural analysis in some cases (Table 2.3). This size requirement was greatlyChapter 2. Experimental Methods 37relieved by the installation of a Rigaku R-AXIS area detector equipped with a rotating anode Xray generator, which made possible diffraction analyses of those crystals too small to be studiedwith the CAD4 diffractometer. In some cases, even small crystals could not be obtained afterextensive crystallization trials. For example, over many years of crystallization trials employingthe hanging drop vapour diffusion method, the free interface diffusion method and various othertechniques, crystals were never obtained for the F82A, F8211, F82R or F82W mutants of yeastiso-i cytochrome c.With the exception of the F82L mutant protein, which is discussed in greater detail inChapter 5, all of the iso-i cytochrome c mutant proteins studied formed crystals which wereisomorphous to those of wild-type iso-i cytochrome c. These crystals belong to the tetragonalspace group P4321and for the wild-type protein have unit cell parameters of a = b = 36.46 A,c 136.86 A. Cell dimensions obtained for mutant protein crystals are discussed in each of theindividual experimental sections in Chapters 3—6. Both wild-type and mutant protein crystalsdisplayed a deep red color characteristic of cytochrome c and were shaped like pifiows (Louieet at., i988a), with the crystallographic c axis being coincident with the thinnest dimension.These crystals have a solvent content of ‘3O% and the asymmetric unit contains a singlemolecule of cytochrome c.2.4 Theoretical Aspects of X-ray DiffractionThis section briefly describes the fundamentals of X-ray diffraction theory as it applies to thedetermination of the three-dimensional structures of proteins. Its purpose is to provide thereader who is unfamiliar with diffraction methods with sufficient background to follow thecrystallographic aspects of the research described in this thesis. A more advanced discussion ofthese techniques is available in a number of excellent texts (for examples see Blundell & Johnson(1976), Ladd & Palmer (1985), Stout & Jensen (1989), McRee (1993) and Drenth (1994)).Chapter 2. Experimental Methods 382.4.1 Diffraction of X-rays by protein crystalsThe determination of a protein structure by X-ray diffraction analysis requires an understandingof the way in which X-rays interact with matter. An object will scatter an incident X-ray ina defined manner and this scattering is dependent on the distribution of the object’s electrondensity as follows:F(S)= f p(r) exp (27ri r S) dv (2.1)In this equation the total scattered wave F in the direction of the reciprocal space vector Sis the sum of the individual waves scattered by the electron density p at real space positionalvector r integrated over the whole volume v of the scattering object. Thus the total set ofscattered waves represents the Fourier transform of the object’s electron density distribution.Two advantageous consequences arise when a protein crystal is the scattering object. Thefirst of these is the multiplication of the intensity of the scattered waves to a measurable levelbrought about by the presence of large numbers of protein molecules in the crystal lattice. Thesecond consequence is that the waves diffracted from a crystal are discrete due to the periodicnature of the crystal lattice. This periodicity allows for the simplification of Equation 2.1 by thereplacement of the scattering vector S with a discrete direction specified by the Miller indiceshkl of a reflection. Thus Equation 2.1 can be rewritten in the following way for diffraction froma crystal:= f p(x, y, z) exp 2i(hx + ky + lz) dv (2.2)This is the equation which relates a structure factor Fhkl, which is composed of both an amplitude and a phase, to the electron density p(x, y, z) present in the unit cell.2.4.2 Calculation of electron density from structure factorsIn order to reconstruct an image of the protein of interest, a map of the distribution of theelectron density in the unit cell of the crystal is required. The calculation of such a map isbased on Equation 2.2, which can be inverted mathematically to give:Chapter 2. Experimental Methods 39p(x,y,z) = exp —27ri(hx + ky+lz) (2.3)where V is the volume of the unit cell. Thus the electron density at any real space position(x, y, z) in the unit cell can be calculated from a summation of the structure factors, therebyallowing a three-dimensional map of the overall distribution of electron density within the unitcell to be constructed. In turn, examination of such a map allows the determination of thepositions of atoms within the unit cell.A major obstacle in this process arises from the fact that each structure factor Fhkl hasboth an amplitude IFhkII and a phase hkl To reflect this, Equation 2.3 can be rewritten as:p(x, y, z) = IFiI exp iahkl exp —2iri(hx + ky + lz) (2.4)The structure factor amplitudes in this equation can be obtained directly from the intensitiesof the reflections measured in the X-ray diffraction experiment. Unfortunately, the phases ofthe structure factors cannot be obtained directly. Thus the calculation of an initial electrondensity map for a crystallized protein requires that estimates be made of the phases of thestructure factors. This step represents a significant impediment to the determination of proteinstructures by X-ray diffraction.2.4.3 Mutant protein structure determinationIn the current work, two different methods have been employed to obtain initial starting modelsof mutant proteins. The method chosen for each mutant protein depended on whether itcrystallized isomorphously to wild-type yeast iso-i cytochrome c or in some other space group.The general outline of these approaches are presented in the two following sections.2.4.3.1 Isomorphous crystal formsA distinct advantage of the growth of mutant cytochrome c crystals having the same spacegroup and unit cell as those of the known structure of wild-type yeast iso-i cytochrome c isthat this greatly facilitates the process of structure solution. This can be seen if we express theChapter 2. Experimental Methods 40structure factor equation (Equation 2.2) in terms directly related to the atomic centers withinthe unit cell as follows:Fhk( = K fj(shkl) exp —B4kl exp 2ri(hxj + ky3 + 1z) (2.5)In this equation, K is a scale factor and 5hkl is equal to (sin 8)/.) for the reflection with Mifierindices hkl. Each atom, j, is defined by its atomic position (xj, yj, z), its atomic scatteringfactor fj, which is dependent only on the identity of the atom (ie: its atomic number), andits thermal parameter B,, which is a measure of the mean square displacement of the atomfrom its average position. This form of the structure factor equation can be used to calculatestructure factors directly from knowledge of the positional and thermal parameters of the atomswithin a unit cell. Therefore if a starting model for a mutant protein in the unit cell can beconstructed, this can be used to estimate the phases of the observed structure amplitudes thathave been experimentally measured. For a mutant protein which crystaJiizes isomorphously toits wild-type counterpart, the structure of the wild-type protein provides an excellent startingmodel. This was the approach taken for many of the yeast iso-i cytochrome c mutants studied.Another advantage of crystailizing both mutant and wild-type proteins isomorphously isthat an electron density map which shows the structural differences between these two proteinscan be calculated. The modified version of Equation 2.4 needed to calculate such an electrondensity map is as follows:>Z (I-1’mut,hklI — IFwt,hklI) exp Owt,hk1 exp —2wi(hx + ky + lz) (2.6)In this equation, Lp(x, y, z) represents the differences between the electron density distributionsin the mutant and wild-type protein crystals, I1’mut,hklI are the structure factor amplitudesobserved in the X-ray diffraction experiment for the mutant protein, IFwt,hklI are the structurefactor amplitudes observed for the wild-type protein, and a,hkl are the phases of the relevantwild-type structure. This difference electron density map shows positive electron density peakswhere there are atoms present in the mutant structure that are absent in the wild-type proteinChapter 2. Experimental Methods 41and vice versa for the negative peaks. Three examples of initial difference electron density mapsof this type for the F82Y, L85A and F82Y/L85A mutant proteins of yeast iso-i cytochrome care shown in Figure 2.13. Inspection of such maps was used to confirm the presence of aminoacid substitutions within mutation sites and allow preliminary modeling of the positioning ofthese residues. In this way, a more representative starting model for a mutant protein could beconstructed for use in structural refinement.2.4.3.2 Non-isomorphous crystal formsWhen a mutant protein crystallizes with a unit cell different than that of wild-type yeast iso-1 cytochrome c, a molecular replacement approach (Rossman, 1972) is necessary to assist instructure determination. In this thesis, the analysis of the F82L mutant protein falls into thiscategory. The details of this specific analysis are given in Chapter 5 while a general discussionof this approach is presented here. A key ingredient in the molecular replacement method isthe calculation of an experimental Patterson map according to the following equation:P(u, v, to) = jFiI exp —2iri(hu + kv + 1w) (2.7)This calculation requires only the availability of the experimentally observed structure factoramplitudes for the mutant protein. In contrast to an electron density map where peaks occur atthe positions of atoms in the unit cell, the Patterson map has peaks at positions correspondingto all interatomic vectors within the unit cell. Thus a peak at (u, v, to) in the Patterson mapindicates that in the unit cell there are two atoms related positionally such that u = —v = Yi — y2 and to = — z2. In practice, an experimental Patterson map is too complex to bedirectly interpreted unless there are only a few atoms in the unit cell.Molecular replacement methods can make use of the experimentally observed Patterson mapto correctly position a search model of the mutant protein in its unit cell. For the structuresolution of the F82L mutant protein, the structure of wild-type yeast iso-i cytochrome c makesa good starting model for this search. The search for this positioning is carried out in twostages. To begin, a rotation function is used to correctly orient the search model. This processChapter 2. Experimental Methods 42ABFigure 2.13: continued on next page.Chapter 2. Experimental Methods 43CFigure 2.13: continued. Fmutant — Fwixd_type difference electron density maps drawn in theimmediate vicinity of the mutated residues in (a) F82Y, (b) L85A and (c) F82Y/L85A yeastiso-i cytochromes c. These maps were calculated using Equation 2.6 (Section 2.4.3.i) withpositive electron density shown as solid contours and negative electron density shown as dashedcontours. Overlaid are the final refined mutant protein structures obtained. Water moleculepositions are shown as small crosses. Difference electron density for the addition of the Tyr82hydroxyl group in the F82Y and F82Y/L85A mutant proteins is readily apparent. Also presentare difference electron density peaks representing the loss of the Leu85 side chain and movementin the side chain of Argl3 in the L85A and F82Y/L85A mutant proteins.Chapter 2. Experimental Methods 44involves the calculation of a set of intramolecular interatomic vectors for the search model whichis then systematically rotated in steps about three axes and compared with the experimentalPatterson map calculated with the observed diffraction data. This comparison is restricted toshort vectors so that only intramolecular interatomic vectors in the observed vector set areincluded. Maximal superposition of the modeled and observed sets of interatomic vectors willbe realized when the orientation of the search model coincides with that of the molecule in thenew unit cell of the mutant protein crystal.When the correct rotational orientation of the search model has been achieved, a translationfunction is utilized to position this model at the correct location within the unit cell. Tostart, a symmetry element is chosen and the correctly oriented search model is systematicallytranslated within the unit cell. Depending on the chosen symmetry element, this systematictranslation can be restricted to specific Harker sections. At each position during this process,the molecule related to the search model by the chosen symmetry element is generated alongwith the associated set of intermolecular vectors. These vectors are then compared againstthe experimentally derived Patterson map to determine the best correlation. In this way thepositioning of the search model with respect to this and other symmetry related molecules canbe determined to ultimately define the correct translational positioning of the search molecule.With the correct rotational and translational positioning of the search model, phase estimatescan be made for the experimentally observed structure factor amplitudes and an initial electrondensity map computed. This then allows for subsequent structural refinement of the mutantprotein structure.Another potentially useful component of the molecular replacement approach is the selfrotation search. This can be applied when multiple copies of a protein, which are related bynon-crystallographic symmetry, are present within the asymmetric unit of the unit cell. Thisprocedure is similar to the rotation function search described previously with the exceptionthat the Patterson map derived from the experimental structure factor amplitudes is also usedas the search Patterson map.Chapter 2. Experimental Methods 452.5 Practical Aspects of Diffraction Data CollectionIn this section, a brief description of the various steps involved in mounting crystals for Xray diffraction analysis, the collection of diffraction data and the processing of this data, isprovided. Specific details pertaining to data acquisition for individual mutants of yeast iso-icytochrome c are given in Chapters 3—6. It should be pointed out that midway through thestudies documented in this thesis, the laboratory changed from collecting data on an EnrafNonius CAD4-F11 diffractometer fitted with a sealed beam X-ray tube, to a newly acquiredRigaku R-AXIS II imaging plate area detector mounted on a RU300 rotating anode system. Theadditional intensity of the rotating anode X-ray beam and increased sensitivity of the imagingplate detector not only greatly improved the quality of diffraction data that could be collected,but also allowed for structure determinations where only very small crystals of mutant proteinscould be grown. The approach taken by the diffractonieter and area detector data collectionmethods are substantially different and general aspects of the methodology used for each aredescribed below.2.5.1 Mounting crystals for data collectionProtein crystals require special handling since they are soft and easily deformed. A furtherconsideration is the need to keep these crystals moist due to their high solvent content. Toprotect crystals during data collection these were mounted in thin walled glass capillary tubeshaving diameters varying from 0.7—1.5 mm, depending on crystal size. Before crystal mounting,a solution identical to the original crysta]iization well solution, except for a 5—10% higherammonium sulphate concentration, was prepared for use in the transfer of the crystal fromits growth well to the capillary tube. After mounting, a limited amount of buffer was leftnear the crystal, with any excess being wicked away using either thin strips of filter paper ormicrocapillary tubes. Also included in the glass capillary before it was sealed with beeswaxwere columns of mounting solution placed above and below the crystal to ensure that it did notdry out during data collection. One end of the capillary was then glued into a mounting pin atChapter 2. Experimental Methods 46the appropriate height for data collection and placed on a goniometer head.2.5.2 Diffractometer data collection and processingX-ray diffraction data sets for crystals of the F82Y, F82Y/L85A, L85A, L85F, L85M andL94S mutants of yeast iso-i cytochrome c were collected using an Enraf Nonius CAD4-F11diffractorneter. The sizes of each of the crystals used in data collection are summarized inTable 2.3. The diffractometer used was equipped with a helium filled beam tunnel and thedistance between the crystal and detector was set to 36.8 cm. A nickel-ifitered copper-targetX-ray tube with focal spot 0.75 mm x 0.15 mm was used to generate the incident radiationand was operated at 26 mA and 40 kV. The ambient temperature during data collection wasmaintained at 15°C.The strategy used in data collection was chosen to maximize the amount of data that couldbe collected from any one crystal, employing fairly narrow scan widths coupled with relativelyslow scan speeds. Typically, continuous ! scans of 0.6° at a speed of 0.55°/minute were used.Data collection was carried out in shells based on resolution, with higher resolution shells beingcollected earlier to offset the greater decay in the intensities of these reflections. Followingthe acquisition of a complete diffraction data set, when time permitted and the decay in theintensities of high resolution reflections had not progressed beyond 30%, repeat measurementsof the high resolution data shells were made. The total time required to collect each completedata set varied between three weeks and one month.The measurement of each reflection was divided into three parts. The middle 2/3 of eachscan was assigned as representing the intensity of the reflection, whereas the first and final1/6 of each scan was used to assess the background radiation and thereby provide a means ofcorrecting the reflection intensity for this factor. An improvement on this background correctioncould be achieved by employing background averaging (Murphy et al., 1992). In this method,an estimate of the background radiation for a reflection is not only based on the backgroundmeasurements of that reflection but also on the backgrounds of reflections which are nearbyChapter 2. Experimental Methods 47in reciprocal space. From experience, it was found that using the backgrounds from 8—12neighboring reflections for this purpose gave the best results.For each data collection, an absorption correction was applied using the empirical method ofNorth et al. (1968). This was necessitated primarily by the non-spherical shape of the crystalsused. In this method, a reflection aligned along the 4 axis of data collection is used. Themeasured intensity of this independent reflection (monitored in 5° increments) varies withrotation of the 4 axis due to absorption and this variation can be used to determine the amountof correction required for a general reflection.Also performed was a correction for intensity decay occurring due to crystal degradationduring X-ray exposure. This correction was based on the periodic measurement of standardreflections over the course of the full exposure time. Generally monitor reflections were groupedaccording to resolution so that the decay correction could be applied on this basis as well sincehigher resolution reflections are in general more sensitive to crystal decay than their lowerresolution counterparts.Following correction of measured intensities for background, absorption and decay effects,two additional corrections are required. The Lorentz correction is necessary because the reflection intensity is dependent on the geometry of the data collection method. The polarizationcorrection adjusts for the variation of reflection intensity with the amount of refracted X-raybeam polarization. Once these two corrections are made, the data are converted into structurefactors by taking the square roots of the reflection intensities.2.5.3 Area detector data collection and processingWith the installation of a Rigaku R-AXIS II imaging plate area detector system and RU300rotating anode, all subsequent data collections were collected using these instruments. Thisincludes data from crystals of all mutant proteins not specifically listed at the beginning ofSection 2.5.2. Unlike the CAD4 diffractometer where each reflection is measured individually,on the Rigaku area detector many reflections are simultaneously measured in a manner similarChapter 2. Experimental Methods 48to the photographic rotation method (Arndt & Wonacott, 1977). The physical basis of measurement used by this area detector consists of two radiation sensitive phosphor plates, each ofwhich allows up to r’.’iO,O00 diffraction intensities to be measured in a single exposure underoptimal conditions. The end result is that on the area detector a complete data set can becollected much more rapidly and usually within 2—3 days, whereas diffractometer data collections required on the order of 3—4 weeks. A further advantage of the area detector system isits sensitivity which allows for the measurement of data to higher resolution and from crystalsthat would be too small for use with a diffractometer. An important factor here is also thegreatly increased intensity of the X-ray beam generated by the rotating anode source.One disadvantage of the area detector is that due to geometric reasons it is not alwayspossible to collect 100% of the available data. Nonetheless, by using multiple scans at alternativesettings, almost all the available data is usually accessible. To assure that data collection is ascomplete as possible, a software program is used to model data collection based on the proposedcollection parameters. These parameters can then be adjusted as necessary to optimize datacollection. For data collections of mutant yeast iso-i cytochrome c crystals, it was found thatby doing two separate series of scans, one of which was collected with an adapter that offsetthe spindle axis of the capifiary holding the crystal by 450 from the axis of the area detector,that >95% of the theoretically available diffraction data could be collected.For data collected on the area detector, the rotating anode generator was operated at 90100 mA and 50—60 kV. The sizes of the crystals used in these data collections are summarized inTable 2.3 for each mutant protein. Data were collected using the oscillation method in which thecrystal was moved through a angle of 1.00 and exposed to the X-ray beam for 20—30 minutes foreach frame collected. Since the method of data collection is similar to the screenless oscillation orrotation method, processing of the collected data resembles the procedure described by Rossmanet cii. (1979), as implemented by Higashi and coworkers (Higashi, 1990; Sato et al., 1992). In thismethod, corrections for background, Lorentz and polarization effects are initially applied to theindividual diffraction intensities measured on each frame of data. The background correctionChapter 2. Experimental Methods 49applied to a particular diffraction measurement is based on the level of background radiationdetected in the vicinity of that reflection. Following these corrections, the frames collected arescaled to account for crystal decay and absorption effects in a resolution dependent manner.Only fully recorded diffraction intensities are included in the computation of scaling factors.After scaling, multiple measurements of intensities are merged and then reduced to structurefactor amplitudes.2.5.4 Determination of absolute scaleStructure factors obtained after data processing are on a relative scale and before the refinementof a model structure can proceed these must be multiplied by a scale factor so that they can becompared to calculated structure factors. The absolute scale factor is initially estimated andthen later improved upon during subsequent structural refinement.Estimation of the absolute scale factor can be accomplished by the method developed byWilson (1942), requiring that only a knowledge of the chemical composition of the unit cellcontents be available. This method assumes the atomic contents of the unit cell are randomlydistributed and therefore the theoretical mean value of diffraction intensities as a function ofresolution is given by:— 2 Isin0N2‘abs = >Zf3 exp _2B__-) (2.8)where ‘abs is the theoretical average intensity. The unit cell contents are represented by theatomic scattering factors f2 of all the atoms, j, in the unit cell and the thermal motion of theseatoms is represented by the thermal factor B.If the average observed intensity Irel is related to ‘abe by scale factor C, then Equation 2.8can be rewritten as:I 7rei \ (sin6”2lflf2} =lnC—2B—-,--)2.9This relationship can then be plotted to give a straight line with a slope of —2B and an interceptof in C. As can be seen in the example given in Figure 2.14, at medium and high resolutionChapter 2. Experimental Methods 5010-3-40.00 0.01 0.08(sin (9)! A.)2Figure 2.14: An example of a Wilson plot based on the R38A/N521/F82S yeast iso-i cytochrome c diffraction data set and used to determine the absolute scale factor. For this analysisreciprocal space was divided into shells according to resolution with each containing at least250 reflections. To determine the absolute scale factor, a least-squares fit line (solid) to the datapoints was calculated using only those data points between 5.0—1.8 A resolution (indicated bythe vertical dashed lines). This analysis suggested that the absolute scale factor for this dataset was 0.76.0.02 0.03 0.04 0.05 0.06 0.07Chapter 2. Experimental Methods 51a straight line is found as predicted, but at low resolution the data points fit less well. Inpractice, this poor fit at low resolution arises from the fact that the contents of the unit cellare not randomly distributed at this resolution. Thus in the determination of absolute scalefactor it is common practice to apply resolution cutoffs, after which a line is least-squares fitto the data and used to obtain this value. The scale factor thus obtained can be used to scalethe observed diffraction data set and the overall estimate of thermal factor serves as a usefulindicator of the isotropic thermal B for the structure as a whole.2.6 Refinement of Structural ModelsOnce an initial model of a mutant protein has been derived by the methods outlined in Section 2.4.3, further optimization is necessary to achieve the best fit between the observed structure factors and those that can be calculated from the structural model. This process involvessuccessive cycles of computational refinement, reference to difference electron density maps andmanual readjustments. One way in which the progress of structural refinement is typicallymonitored is by periodic calculation of the crystallographic R-factor as follows:R — hk1 I IF0,hklI — IFc,hklI I (2 10)— Zhk1 IF0,hklIwhere I are the observed structure factor amplitudes measured from the mutant proteincrystal and IFc,hklI are those calculated from the corresponding model structure by Equation 2.5.Thus the crystallographic R-factor is a direct measure of the agreement between the experimental data and the model structure, with a lower R-factor implying a better fit between these.For well refined protein structures, the value for the R-factor can vary between 0.10 and 0.25.2.6.1 Stereochemically restrained refinementThe primary method employed in structural refinement in this thesis is the restrained parameterleast-squares technique (Hendrickson, 1985). This refinement method operates by minimizingthe differences between the structure factor amplitudes observed in the X-ray diffraction experiment and those calculated from the model structure (by Equation 2.5), while ensuring thatChapter 2. Experimental Methods 52the stereochemistry of the protein structure remains within acceptable limits through the useof geometric restraints. This is accomplished by minimizing the following equation:2 hkl ( o,hklI — IFc,hkII)2 + geometric(2.11)hkl UF( )where 0F determines the relative weight of the structure factor terms in the equation andçbgeomtrjC are the terms defining the geometric restraints.The geometric terms in Equation 2.11 serve to restrain the relative atomic positions of theprotein structure so that the stereochemical parameters associated with these positions do notdeviate significantly from expected values. For example, the geometric term used to restrainthe distances between individual atoms connected through covalent bonds is:distances 14distances = 2‘.(dideal,j dmodel,j)2 (2.12)DJ)where 0D is the expected standard deviation for the distribution of the distances, djdeal,j isthe ideal distance between an atom pair j, and dmodel,j is the value for this distance calculatedfrom the model structure. Other geometric terms serve to restrain planar groups, chiral centers,non-bonded contacts, torsion angles and thermal parameters.2.6.2 Simulated annealing refinementA primary constraint of stereochemically restrained refinement is its relatively small radius ofconvergence. As discussed in detail in Chapter 5 for the F82L and F82M mutant proteins,this can lead to unsatisfactory refinements when larger structural corrections are required. Onetechnique that is useful under these circumstances is simulated annealing refinement (Briingeret aL, 1987) which is capable of sampling possible structural conformations beyond the localconformational energy minimum. This is carried out by linking the minimization of the differences between the observed and calculated structure factor amplitudes to a molecular dynamicssimulation. Such a simulation involves the solution of Newton’s equations of motion for eachindividual atom over a fixed span of time, with these atomic motions being dependent on theChapter 2. Experimental Methods 53interaction of this atom with every other atom in the simulation. These interactions are represented in the form of a potential energy function which consists of the sum of the energieswhich define the forces that act on each atom. In the simulated annealing method, the potentialenergy function is modified by the inclusion of a term representing the diffraction data:Etotat = Wxray (IFo,hk1 — IFc,hkll)2+ Egeometrjc (2.13)hklwhere Etotal is the total energy in the system, Wxray determines the relative weight of the structure factor terms in the equation and Egeometric are the potential energy terms calculated fromthe geometry of the molecule. Although this equation is similar to Equation 2.11, its application is different. Whereas in conventional least-squares refinement Equation 2.11 is minimized,in simulated annealing refinement the total energy in Equation 2.13 is set to an arbitrarilyhigh level and large time-dependent structural fluctuations of the molecule are simulated tosearch conformational space for the best fit between the observed and calculated structurefactor amplitudes.An example of a term in the potential energy function related to structural geometry duringsimulated annealing refinement is the following:bonds7 iiJ’bonds = jUmode1,j — Uaverage,j)3where k3 is a constant describing the variability about the average length daverage,j of bondj. As the bond distance calculated from the model structure deviates from the expected average bond distance, the potential energy of the bond increases to an unfavorable level. Otherpotential energy terms representing bond angles, dihedral angles, van der Waals’ contacts andelectrostatic interactions are also calculated.2.6.3 General considerations in refinement of mutant proteins2.6.3.1 X-ray diffraction data setsA number of factors can affect a particular X-ray diffraction data set. For example the completeness of data collected from crystals of yeast iso-i cytochrome c mutant proteins was dependentChapter 2. Experimental Methods 54on the size of these that could be grown (Table 2.3). Thus bigger crystals resulted in a greaternumber of observable reflections being collected than for small crystals where a larger portionof the reflections were too weak to be recorded. For example, only very small crystals couldbe grown for the F82M and F821 mutant proteins. Using the best of these, 2143 and 2460reflections (6.0—2.3 A resolution range), respectively, out of the ‘.‘4300 theoretically availablecould be measured using an Enraf-Nonius CAD4-F11 diffractometer equipped with a sealedbeam X-ray tube.Beyond crystal size, another significant factor in the numbers of reflections that can bemeasured from a given crystal is the detection instrumentation used. As noted earlier, midwaythrough the studies conducted in this thesis, the laboratory acquired a Rigaku R-AXIS II areadetector and RU300 rotating anode generator. The increase in X-ray beam intensity available,coupled with the increased sensitivity of the area detector, meant that more complete diffraction data sets could be collected even when crystal size was a severely limiting parameter. Forexample, for the F82M and F821 mutant protein crystals discussed above, collection of cliffraction data using the area detector resulted in 2913 and 3727 observed reflections, respectively.These significantly improved data sets greatly facifitated the structural refinements of thesemutant proteins (Chapter 5). However, it should be noted that despite this improvement, thelimitations imposed by only being able to grow relatively small crystals could not be entirelyovercome and a number of reflections from the F821 and F82M mutant protein crystals remainedtoo weak to be observed. For those crystals of other mutant proteins for which only the CAD4diffractometer was available during the collection of X-ray diffraction data, the completenessof the data sets measured is poorer, corresponding to the less intense X-ray source used andtherefore the overall weaker intensity of the data available. A plot of data completeness versusresolution for each mutant protein diffraction data set is provided in the experimental sectionsof Chapters 3—6.After collecting the best possible diffraction data set, another concern is how the weakest,and therefore most inaccurately measured, reflections should be handled. These reflections areChapter 2. Experimental Methods 55the most likely to contribute to noise in electron density maps, thereby obscuring the interpretation of such maps during the process of structural refinement. To remedy this situation, itis common to impose a threshold value or a cutoff to filter reflections used in refinements. Asurvey of the literature reveals that there is a wide range of o cutoffs in use, with the mostcommon values ranging from IFI (O—3)a(F) (McRee, 1993; Watkin, 1994). Close scrutiny ofthe effects of using different cutoffs of this magnitude to exclude weak data show that these donot adversely affect the structure resulting from refinement (Stenkamp & Jensen, 1975; Smalas& Hordvik, 1993).In the current structural studies a consistent a cutoff of 2a(F) was employed, except forthe L85A mutant protein data set where a 2.5a(F) cutoff was applied (problems with crystalslippage were encountered during data collection for the L85A mutant protein — see Chapter 3).Experience showed that this cutoff level met the dual goals of including the greatest amount ofdata possible and being able to facifitate structure refinement with the clearest electron densitymaps. For data collected on the CAD4-F11 diffractometer, these cutoffs did affect a fair numberof reflections depending on the size of the crystal used. In contrast, data sets collected on theRigaku area detector had very few reflections faffing below this threshold value, with the worstcase involving the F82M mutant protein data set where only 51 reflections fell into this category.To assess the utility of the 2a(F) cutoff imposed, a test was made with the L85M mutantprotein data set which had been collected with the CAD4-F11 diffractonieter. Using the finalrefined structure of this protein, an additional eight cycles of restrained parameter least-squaresrefinement (Section 2.6.1) were carried out using all observed data between 6.0 and 1.9 A resolution (6396 reflections, 83% completeness). This resulted in an average cumulative positionalshift of 0.07 A for all atoms over the 8 cycles of refinement, a value similar to the cumulative shiftof 0.06 A observed over the last 6 cycles for the original data set with a 2a(F) cutoff. Therefore,at the end of structural refinement, shifts with the inclusion of all data were of the same orderas those with the 2a(F) cutoff and a comparison of the two refined structures showed therewere no significant structural differences. One factor that was affected was the crystallographicChapter 2. Experimental Methods 56R-factor which was 0.208, compared to 0.178 for the 5210 reflections with IFI 2(F) in thesame resolution range (Table 4.15; Chapter 4). Thus as expected, inclusion of all data has littleimpact on structural shifts while leading to a considerably higher R-factor and noisier electrondensity maps.Two other factors affecting diffraction data were contributions resulting from the disorderedsolvent continuum and alternative crystal lattice packing arrangements. The first of these is adominant contributor to low resolution reflections; to account for this, diffraction data below6.0 A resolution were excluded from refinement (Blundell & Johnson, 1976; Schoenborn, 1988).An example of the latter factor was found for the F82L mutant protein crystals which werefound to be of a different space group than those grown for other mutant proteins. As discussedin Chapter 5, these crystals diffracted relatively poorly even when crystal size was taken intoaccount. This resulted in a correspondingly higher rate of fall off in the number of observedreflections at high resolution as illustrated in Figure 5.30. For this protein, even with the useof the Rigaku area detector system, the resolution of the best structure that could be obtainedwas 2.5 A.2.6.3.2 Difference electron density mapsDuring structural refinements, it was necessary to carry out manual fittings on poorly behavedamino acids based on reference to different types of difference electron density maps. In addition,such maps covering the whole of the polypeptide chain were calculated periodically (from 4—7times) during refinements to provide a means of checking the course of refinement and for anyother adjustments to the structure that might be necessary. This approach also provided theprimary means for locating solvent molecules bound to the protein surface or within the proteinitself.All of the difference electron density maps utilized were based on the calculation shownin Equation 2.6, but for which different structure factor and phase coefficients were used. Inone case this was a simple difference map with coefficients of IFo,hklI — IFc,hkll and exp 2tc,hk1.Chapter 2. Experimental Methods 57Here, JF0,hkzI represents the experimentally measured structure factor amplitudes and IFc,hklIand c,hkl are the structure factor amplitudes and phases computed from the refinement model.A related difference electron density map calculated with coefficients of21F0,hklj — IFc,hklI andexp icr,hk1 also proved very useful in assessing the fit of the refinement model to its electrondensity. A third type of map that particularly facilitated interpretations in those regions whereit proved difficult to fit the polypeptide chain was the omit difference electron density map. Herethe atoms of the structural feature being analyzed (this might be just a few atoms or those ofseveral amino acids) are excluded and structure factor amplitudes I Fomjt,jjk and phases cromjt,hklare calculated based on the resultant structural model. In this way a difference electron densitymap can be calculated by using the coefficients IFO,hklI — IFomit,hklI and exp omit,hk1 in aformulation of Equation 2.6. In this map the excluded feature should be present as a largepositive peak and thereby facilitate fitting of this part of the structure. The use of these threetypes of difference electron density maps was a powerful tool in achieving the best fit of mutantprotein structural models to their electron density maps.2.6.3.3 Course of a typical refinementThe course of the refinement of the N521/F82S mutant of yeast iso-i cytochrome c is ifiustratedin Figure 2.15 and provides a typical example of that followed for other mutant proteins. Notethat the precise details for this and other mutant protein structural refinements are given inChapters 3—6. After initial placement of the starting model for the N521/F82S structure inthe unit cell as discussed in Section 2.4.3.1, a series of refinement cycles was conducted usinglower resolution data until the crystallographic R-factor was sufficiently lowered. Then furtherrefinement was carried out with higher resolution data, interspersed with rounds of manualinterventions based on difference electron density maps. The water molecules associated withthe protein structure were also examined during each manual intervention and were added ordeleted to the refinement model as necessary. Possible water molecules were also identifiedby a peak search of a F1, — F difference electron density map or with the assistance of anChapter 2. Experimental Methods 580 10 20 30 40 50 60 70 80Refinement CycleFigure 2.15: The course of the structural refinement of N521/F82S yeast iso-i cytochrome c bythe restrained parameter least-squares method. This plot shows the crystallographic R-factor(Equation 2.10) as a function of refinement cycle number. Modeling of the side chains of 11e52and Ser82 into F0 — F difference electron density maps occurred after refinement cycle 24.Manual interventions during which the fit of the entire polypeptide chain was examined and awater search conducted using 2F0 — F, F0 — F and omit difference electron density maps areI I I I4Refinement with 6.0-2.2 A data,11e52 and Ser82 modelled as alaninesI iiiii....I0.40.30.2 -C0with 6.0-1.8 A dataModeling of fle52 andSer82 side chainsof data scale• Manual intervention,Water searchindicated.Chapter 2. Experimental Methods 59automated procedure utilizing alternating cycles of peak searching and refinement cycles (Tonget at., 1994). The criteria for the inclusion of a water molecule into the protein structure werethat the potential water position had to be within 3.5 A of a hydrogen bond donor or acceptoratom and that it refine to an isotropic thermal factor under 50 A2. The refinement processwas continued until convergence was reached for both positional shifts and the crystallographicR-factor. A paramount consideration during refinements was that acceptable stereochemistrybe maintained for the mutant protein structure. A compilation of stereochemical values andthe restraint weighting used during refinements is given in Table 2.4. Values pertaining toindividual mutant proteins can be found in Chapters 3—6.2.6.3.4 Calculation of atomic coordinate errorsIn this thesis, error estimates for the atomic coordinates of iso-i cytochrome c mutant structureswere determined by two methods. One method was that of Luzzati (1952) where overall coordinate errors are estimated by comparison of the dependence of the crystallographic R-factor onresolution with theoretical estimates of this dependence assuming that these errors are the onlysource of differences between observed and calculated structure factors. Luzzati plots for eachof the mutant protein structures determined are presented as part of the experimental sectionsof Chapters 3—6. Coordinate errors estimated by this method ranged between 0.16 and 0.26 Afor the fourteen mutant iso-i cytochromes c studied. In a second method (Cruickshank, 1949,1954), the errors in the coordinates of a given atomic position are dependent on the fit betweenobserved and calculated structure factors, the atom type and the thermal factor of the atom.Estimates for the overall r.m.s. coordinate error using this method ranged between 0.12 and0.21 A for all mutant protein structures studied. In general there was a close correspondencebetween the error estimates obtained by the Luzzati and Cruickshank methods. Nonetheless itmust be remembered that these are both empirical methods relying on assumptions that arenot completely valid and care must be taken when using such error estimates as an absolutemeasure of coordinate error.Chapter 2. Experimental Methods 60Table 2.4: Compilation of stereochemistry observed in refined yeast iso-i cytochrome c mutantstructuresStereochemical parameter r.m.s. deviation from Restraint weightingideal values valueDistances (A)Bond (1-2) 0.019 0.020Angle (1-3) 0.037—0.051 0.030Planar (1-4) 0.045—0.057 0.045Planes (A) 0.012—0.015 0.018Chiral volumes (A3) 0.132—0.202 0.120Non-bonded contacts (A)tSingle torsion 0.211—0.229 0.250Multiple torsion 0.179—0.220 0.250Possible hydrogen bonds 0.185—0.239 0.250Torsion angles (°)Planar (0° or 180°) 1.9—2.5 2.5Staggered (±60° ,180°) 19.0—28.4 20.0Orthonormal (+90°) 17.4—27.1 15.0f The r.m.s. deviations from ideality for this class of restraint incorporates a reduction of 0.2 Afrom the radius of each atom involved in a contact.Chapter 3The Roles of Residues 82 and 85 at the Interactive Face of Cytochrome c3.1 IntroductionThe objective of the studies conducted in this chapter is to examine the roles of Phe82 andLeu85 as participants in forming a part of the interactive surface between cytochrome c andcomplexed electron transfer partners. The side chain of Phe82 forms an integral part of thehydrophobic heme pocket in cytochrome c and packs directly against the heme prosthetic groupin a coplanar fashion. The side chain of Leu85, in turn, packs against the distal edge of thearomatic ring of Phe82, as illustrated in Figure 1.9. The side chains of both of these residues areexposed to the external solvent environment in that surface region of cytochrome c which, alongwith the exposed edge of the heme moiety, is located at the putative site of complex formationand electron transfer between cytochrome c and its redox partner proteins. Evidence for theinvolvement of Phe82 and Leu85 in the formation of electron transfer complexes has come frommolecular modeling (Salemme, 1976; Poulos & Kraut, 1980; Mauk et aL, 1986; Lum et al.,1987), computational studies (Wendoloski et aL, 1987; Eltis et al., 1991; Northrup et aL, 1993),NMR (Pielak et aL, 1988; Burch et aL, 1990), X-ray crystallography (Pelletier & Kraut, 1992)and comparative studies of electron transfer kinetics (Nocek et al., 1991). Following chaptersdeal in greater depth with other aspects of the contributions of Phe82 (Chapters 5 and 6) andLeu85 (Chapter 4) to the structural and functional properties of cytochrome c.Phe82 is a phylogeneticaily invariant residue (Hampsey et al., 1988; Moore & Pettigrew,1990) and it has been shown that substitution of this residue by glycine, serine or tyrosine resultsin decreased steady-state rates of electron transfer between cytochrome c and cytochrome cperoxidase (Pielak et at., 1985) and cytochrome c oxidase (Michel et a!., 1989). Furthermore,61Chapter 3. Roles of Residues 82 and 85 in Cytochrome c 62these substitutions and others at Phe82 also have a detrimental effect on yeast growth rates(Hilgen & Pielak, 1991; Inglis et al., 1991) and modify the kinetics of electron transfer mediatedby cytochrome c (Everest et at., 1991; Hazzard et at., 1992). These results serve to emphasizethe important role Phe82 plays in the formation of complexes between cytochrome c and redoxpartner proteins.With respect to the electrochemical reduction potential of cytochrome c, Phe82 also appearsto play an important role that is related to the close packing of the side chain of this residue nextto the heme group (Rafferty et at., 1990). Previous structural studies of the F82S and F82Gmutants of yeast iso-i cytochrome c revealed that the lowered reduction potentials observedfor these two mutant proteins arise from changes in the hydrophobic heme environment (Louieet at., 1988b; Louie & Brayer, 1989). Specifically, the mutation of Phe82 to serine results in theintroduction of a solvent channel into the heme pocket and the substitution of glycine at thislocation causes a comformational change in the local polypeptide backbone that places polargroups next to the heme moiety. Both substitutions substantially disrupt the hydrophobicnature of the heme pocket and thereby provide an explanation for the changes observed in thefunctional behavior of these mutant proteins.For the case of the substitution of tyrosine for Phe82, it is surprising that there is littleimpact on heme reduction potential (Rafferty et at., 1990) even though this replacement hasa significant impact on electron transfer complex formation (Pielak et at., 1985) and overallprotein stability (Greene et at., 1993). Molecular graphics modeling of the F82Y amino acidreplacement, as described in Section 2.2, indicates that a structural conflict is likely to arisebetween the hydroxyl group of Tyr82 and the side chain of the adjacent Leu85. This conifictwould be expected to necessitate a shift in either of the side chains of TyrS2 or Leu85, orperhaps in both. An earlier NMR study has tentatively assigned the residue shifted as beingLeu85 (Pielak et at., 1988; Greene et at., 1993). Although this result provides an adequateexplanation for the maintenance of heme reduction potential, it does not explain the significantimpact this mutation has on the formation of electron transfer protein complexes.Chapter 3. Roles of Residues 82 and 85 in Cytochrome C 63In this chapter, to address the issue of precisely what structural changes occur as the resultof replacement of Phe82 by tyrosine, the three-dimensional structure of this mutant proteinhas been elucidated using X-ray diffraction techniques. Furthermore the role of Leu85 in thesestructural shifts was assessed by examining the structure of the F82Y/L85A mutant protein inwhich the steric conflict that would normally occur between these residues should be relieved.To assure that the L85A mutation in itself does not lead to unexpected structural consequences,the structure of the mutant protein having this single amino acid substitution was also determined. By examining the structures of all three of these mutant proteins and that of wild-typeyeast iso-i cytochrome c, it was hoped that a comprehensive structural understanding of therelationship between Phe82 and Leu85 could be achieved in light of the available functionalstudies. Note that the studies in this chapter form an important part of the following chapters where additional mutations at both Phe82 and Leu85 are examined. The results of thestructural analyses in the present chapter have been published in Lo et a!. (1995a).3.2 Experimental ProceduresCrystals of the F82Y, L85A and F82Y/L85A mutants of yeast iso-i cytochrome c were grownunder reducing conditions as discussed in Section 2.3. For these proteins the hanging dropvapour diffusion method was augmented with either seeding from macro-crystals or hair-seedingfrom micro-crystals. As summarized in Table 3.5, crystals of all three mutant proteins grewisomorphously to those of wild-type yeast iso-i cytochrome c.Following the methods outlined in Section 2.5.2, each mutant cytochrome c data set wascollected from a single crystal on an Enraf Nonius CAD4-Fii diffractometer having a crystalto counter distance of 36.8 cm and equipped with a helium filled beam tunnel. A nickel-ifiteredcopper-target X-ray tube with a focal spot of 0.75 mm x 0.15 mm and operated at 26 mA and40 kV was used to generate the incident radiation. The ambient temperature was maintainedat 15°C during data collection.During data collection for the L85A mutant protein, an examination of the reflections usedChapter 3. Roles of Residues 82 and 85 in Cytochrome c 64Table 3.5: Data collection parameters for F82Y, L85A and F82Y/L85A yeast iso-i cytochromes cIso-i cytochrome c mutantParameter F82Y L85A F82Y/L85ASpace group P4321 P4321 F4321Cell dimensions (A)a = b 36.43 36.60 36.52c 136.67 138.30 136.74Number of reflections collected 7073 12191 10278Number of unique reflections 6173 6909 5408Resolution (A) 1.97 1.9 2.0to monitor decay and slippage indicated that the crystal was gradually shifting within the X-raycapillary tube, probably due to the presence of excess mounting buffer around the crystal. Tocorrect this, the X-ray capillary tube was unsealed and the excess buffer was removed with amicrocapillary pipette, with special care being taken to avoid disturbing the protein crystal.The column of buffer above the crystal was then replaced with fresh mounting buffer and the Xray capillary was resealed with wax. Following this process, the matrix defining the orientationof the L85A mutant protein crystal relative to the diffractometer was redetermined to confirmthat no significant shift in the position of the crystal within the X-ray capillary had occurred.No further problem with slippage of this crystal was encountered.The intensity data sets for all three mutant proteins were corrected for backgrounds, absorption (North et al., 1968), crystal decay, and Lorentz and polarization effects as described inSection 2.5.2. In order to statistically improve high resolution intensity measurements for thethree diffraction data sets. local background averaging was employed (Louie & Brayer, 1990;Chapter 3. Roles of Residues 82 and 85 in Cytochrome C 65Murphy et al., 1992). Prior to structural refinement, the diffraction data sets for the F82Y,L85A and F82Y/L85A mutant proteins were placed on an absolute scale by the method ofWilson (1942) (Section 2.5.4).Starting models for the three mutant protein structure refinements were constructed basedon the wild-type reduced iso-i cytochrome c structure determined to 1.2 A resolution (Louie &Brayer, 1990). The initial Fmutant — Fwjld_type difference electron density maps for each mutantstructure are shown in Chapter 2 (Figure 2.13) and have clear positive electron density peaksfor the addition of the Tyr82 hydroxyl groups in the F82Y and F82Y/L85A mutant proteins,along with negative electron density peaks representing the truncation of the Leu85 side chainin the L85A and F82Y/L85A mutant proteins. In addition, these maps have electron densitypeaks representing conformational shifts in the side chains of Argi3 and Leu94 in the L85Amutant protein and in the side chain of Argl3 in the F82Y/L85A mutant protein. Examinationof these maps allowed the modeling of the new positions of the side chains of Argi3, Tyr82and Leu94 into the starting models for the structural refinement of these mutant proteins. Allordered waters with refined isotropic thermal factors under 50 A2 from the high resolution wild-type structure were used in the starting models except for those which were in close proximity(< 4 A) to the mutation sites. The sulphate anion bound to the amino terminal end of theos-helix comprised of residues 3 through 12 in the wild-type structure was also included in thestarting models of all three mutant proteins.Structural refinements were carried out using a restrained parameter least-squares approach(Hendrickson, 1985) following the general protocols outlined in Section 2.6. This process utilized structure factors having a resolution greater than 6.0 A and a F/u(F) ratio greater than2.0 (2.5 in the case of the L85A data set) as discussed in Section 2.6.3.1, and employed therestraint weights listed in Table 2.4. The solvent water molecules included in structural refinement were modeled as neutral oxygen atoms with full occupancies. Overall, between 60—90individual cycles of least-squares computational refinement were carried out for each structuredetermination. During structure refinements, an average of 5—8 sequential examinations of theChapter 3. Roles of Residues 82 and 85 in Cytochrome c 66entire polypeptide chain using omit, F0 — F and 2F0 — F difference electron density maps wereconducted at regular intervals to check the progress of refinement and make manual correctionswhere necessary. These manual corrections included the identification of additional solventwaters by searching for strong difference electron density peaks within 3.5 A of hydrogen bonddonor or acceptor atoms, and the deletion of weakly resolved water molecules. Additional examinations of specific areas of each protein were required using F0 — F and omit differenceelectron density maps due to poorly defined electron density resulting from the high thermalmotion of these regions. The most significant of these problem areas were the first 5 amino-terminal residues of each protein and the placement of various surface side chains includingthat of Argl3 in the F82Y mutant protein. Whereas the side chain of Argl3 is substantiallydisordered in the F82Y and wild-type proteins, in the case of the L85A and F82Y/L85A mutantstructures well ordered electron density is present for Argl3 as demonstrated in Figure 2.13.This feature of Argl3 is probably due to the replacement of Leu85 by alanine as discussed inthe results section of this chapter. The final refined structures of all three mutant proteins exhibit good stereochemistry and this is summarized in Table 3.6 along with the final refinementparameters for these structures.Atomic coordinate errors for mutant protein structures have been estimated using two methods (see Section 2.6.3.4). Inspection of the Luzzati (1952) plot drawn in Figure 3.16 indicatesthat these errors are in the range of 0.18—0.22 A. Overall atomic coordinate errors can also beestimated by evaluating the individual atomic errors (Cruickshank, 1949, 1954). On the basisof this method, the estimated overall r.m.s. coordinate error is 0.16 A for the F82Y structure,0.18 A for the L85A structure, and 0.20 A for the F82Y/L85A structure.3.3 ResultsTo allow for a comprehensive analysis of the effects of mutations in the F82Y, L85A andF82Y/L85A mutant proteins, all the available coordinate sets, along with that of wild-type yeastiso-i cytochrome c, were superimposed using a least-squares procedure based on all of-carbonChapter 3. Roles of Residues 82 and 85 in Cytochrome c 67Table 3.6: Refinement results and stereochemistry for the F82Y, L85A and F82Y/L85A yeastiso-i cytochrome c mutant structuresF82Y L85A F82Y/L85A1. Refinement resultsResolution range (A) 6.0—1.97 6.0—1.9 6.0—2.0Number of observed reflections 4546 4886 3761Completeness in resolution range (%) 66.4 62.3 57.2Number of protein atoms 895 891 892Number of solvent atoms 70 79 72Average thermal factors (A2)Protein atoms 14.8 15.4 16.6Solvent atoms 23.4 27.2 27.4R-factor 0.186 0.196 0.1852. Stereochemistry of final modelsr.m.s. deviation fromideal valuesDistances (A)Bond (1-2) 0.019 0.019 0.019Angle (1-3) 0.038 0.039 0.039Planar (1-4) 0.045 0.046 0.045Planes (A) 0.013 0.012 0.013Chiral volumes (A3) 0.163 0.138 0.159Non-bonded contacts (A)fSingle torsion 0.212 0.216 0.226Multiple torsion 0.192 0.220 0.207Possible hydrogen bonds 0.215 0.239 0.234Torsion angles (°)Planar (0° or 180°) 2.0 1.9 2.0Staggered (±60°,180°) 24.7 23.1 23.6Orthonormal (±90°) 21.7 19.5 17.4t The r.m.s. deviations from ideality for this class of restraint incorporates a reduction of 0.2 Afrom the radius of each atom involved in a contact.Chapter 3. Roles of Residues 82 and 85 in Cytochrome C 68Resolution (A)5.0 3.3 2.5 2.0 1.7-0.0 ()0.35 -O.30-0.24 A025- --_. 0.20A0.16 A0.10— i I I I I I I I I I I I I I I I I I I I0.05 0.10 0.15 0.20 0.25 0.30(sin 0)!Figure 3.16: A plot of the crystallographic R-factor at the end of refinement as a function of resolution for the F82Y (0), L85A (Q) and F82Y/L85A (vi) mutants of yeast iso-i cytochrome c.Tue theoretical dependence of R-factor on resolution assuming various levels of r.m.s. errorin the atomic positions of the model (Luzzati, 1952) is shown as broken lines. This analysissuggested an overall r.m.s. coordinate error for the mutant structuresof between 0.18 and0.22 A. The top portion of this figure (axes at top and right) shows the fraction of reflectionsobserved and used in refinement as a function of resolution.Chapter 3. Roles of Residues 82 and 85 in Cytochrome c 69Table 3.7: Overall average positional deviations (A) between wild-type yeast iso-i cytochrome cand the F82Y, L85A and F82Y/L85A mutant proteinsIso-i cytochrome c mutantAtom groups F82Y L85A F82Y/L85AAll common protein atoms 0.31 0.37 0.35All main chain atoms 0.18 0.21 0.22All common side chain atoms 0.47 0.57 0.52All heme atoms 0.13 0.15 0.15atoms. This comparison is shown in Figure 3.17 and reveals that the overall polypeptide chainfold of yeast iso-i cytochrome c is retained upon introduction of these mutations at residues82 and 85. As indicated in Table 3.7, average deviations for main chain atoms between thewild-type and three mutant proteins are on the order of 0.18—0.22 A.Plots of the average positional deviations of residues along the polypeptide chain of themutant proteins are shown in Figure 3.18. The N-terminal residues, Thr(-5) and Glu(-4), aresubstantially disordered and therefore the large positional deviations seen at these positionsare likely the result of positional mobility rather than a reflection of the induced mutations.Tables 3.8 and 3.9 show that the overall geometry and solvent exposure of the heme groupof each mutant protein is comparable to that of wild-type cytochrome c. For each individualmutant protein, specific structural alterations are observed at or near the sites of mutation andthese are discussed in detail in the following sections.Chapter 3. Roles of Residues 82 and 85 in Cytochrorne c 70ABFigure 3.17: Stereo diagrams of the a-carbon backbones of the wild-type, F82Y, L85A andF82Y/L85A iso-i cytochrome c structures in (a) the standard view looking at the heme edge-onand (b) an alternate view looking directly down into the mutation sites at residues 82 and 85.Also drawn are the side chains of the mutated residues, Phe82 and Leu85, and the heme moietiesof all four proteins, along with the ligands to the heme iron atom (Hisi8 and Met8O) andcysteines 14 and 17, which form covalent thioether bonds to the heme porphyrin ring. Everyfifth amino acid residue is indicated by its one-letter amino acid designation and sequencenumber.Chapter 3. Roles of Residues 82 and 85 in Cytochrome c 714.0• III I I I II IIIIIjII• I liii,,,,,, I3.5,_. 3.0i...i-1 10 20 30 40 50 60 70 80 90 100Residue Number13 4.0 I I I II I3.53.02.52.01.51.00.50.0 I I-1 10 20 30 40 50 60 70 80 90 100Residue NumberFigure 3.18: continued on next page.Chapter 3. Roles of Residues 82 and 85 in Cytochrome c 72C>ti)Figure 3.18: continued. Plots of average positional deviations from the wild-type iso-i cytochrome c structure for the (a) F82Y, (b) L85A and (c) F82Y/L85A mutant proteins. Thicklines indicate average deviations of main chain atoms while thin lines indicate average deviations of the equivalent side chain atoms. In each diagram, the filled circle at residue position104 represents the average positional deviation of the heme group and the horizontal dashedline represents the average positional deviation for all main chain atoms.-1 10 20 30 40 50 60 70 80 90 100Residue NumberChapter 3. Roles of Residues 82 and 85 in Cytochrome C 73Table 3.8: Heme geometry of F82Y, L85A, F82Y/L85A and wild-type yeast iso-i cytochromes cWild-type F82Y L85A F82Y/L85A1. Angular deviations (°) between the pyrrole nitrogen plane normaland the four individual pyrrole ring plane normals and the hemecoordinate bonds.A 9.4 9.0 9.2 13.5B 11.1 11.2 8.8 11.6C 8.8 8.5 9.2 10.1D 8.1 9.5 10.2 10.2Fe - Hisl8 NE2 2.2 2.8 3.3 8.3Fe - Met8O SD 4.9 2.1 4.3 5.62. Angular deviations (0) between the porphyrin ring plane normaland the four pyrrole ring plane normals, the pyrrole nitrogenplane normal and the heme coordinate bonds.A 6.7 5.9 5.7 7.9B 11.9 10.7 10.1 12.4C 9.8 10.8 9.8 10.5D 6.0 7.0 6.7 4.7NNNN 2.6 3.1 3.4 5.6Fe - Hisl8 NE2 3.2 4.7 0.3 2.9Fe - Met8O SD 7.5 4.9 7.9 11.13. Bond distances (A) between the heme iron atom and itssix ligands.HisiS NE2 1.98 1.95 1.92 2.14Met8O SD 2.36 2.28 2.38 2.33Heme NA 1.97 1.98 1.99 2.01Heme NB 2.00 2.02 1.98 2.00Heme NC 1.99 2.02 1.98 2.02Heme ND 2.01 2.05 2.11 2.01The pyrrole nitrogen plane is defined by the four pyrrole nitrogens of the heme group. The fourpyrrole ring planes are each defined by the five atoms of the ring and the first carbon atomattached to each of the four carbons of the ring. The porphyrin ring is defined by the five atomsin each of the four pyrrole rings, the four bridging methine carbon atoms, the first carbon atomof each of the eight side chains of the heme and the central iron atom of the heme. The hemeatom nomenclature used in this table follows the conventions of the Protein Data Bank (seeFigure 1.2).Chapter 3. Roles of Residues 82 and 85 in Cytochrome c 74Table 3.9: Heme solvent accessibility in F82Y, L85A, F82Y/L85A and wild-type yeast iso-icytochromes cYeast iso-i cytochrome c structureWild-type F82Y L85A F82Y/L85A1. Solvent accessible lieme atomsand surface area exposed (A2)CHD 2.9 2.9 2.8 3.1CMC 9.2 11.2 12.3 9.2CAC 3.4 4.1 3.3 4.4CBC 20.1 20.6 18.4 18.3CMD 10.8 11.2 10.3 10.12. Total heme exposure (A2) 46.4 50.0 47.1 45.13. Total heme surface (A2) 513.1 515.0 514.7 512.24. % heme surface area exposed 9.0 9.7 9.2 8.8Solvent exposure was determined by the method of Connolly (1983) with a probe sphere havinga 1.4 A radius.Chapter 3. Roles of Residues 82 and 85 in Cytochrome c 753.3.1 Structure of F82Y cytochrome cIn wild-type yeast iso-i cytochrome c, the side chain of Leu85 lies directly adjacent to the distaledge of the aromatic ring of the side chain of Phe82. Thus, mutation of Phe82 to a tyrosineplaces the additional hydroxyl group into direct spatial conifict with the Leu85 side chain. Asthe current structural analysis shows, this potential conflict is resolved by the side chain ofTyr82 undergoing a significant positional displacement (zd = 0.7 A ), involving a rotationtoward the surface of the protein (‘.xi 30; zX2 = 100; Figure 3.19). At this location, theside chain of Tyr82 exhibits significantly increased thermal mobility (zB 8 A2; Table 3.10)and the hydroxyl group of this residue hydrogen bonds to a new surface water molecule. Alsohaving increased thermal mobilities are the main chain atoms of residues 81 and 82. In contrast,the Leu85 side chain is shown to retain a position comparable to that in the wild-type structure,with an average deviation of 0.3 A for side chain atoms. Tyr82 displacement also decreases theangle between the normal of its planar group and that of the heme to 9° from the 23° observedin the wild-type protein. The difference in these values, 14°, is significantly larger than the 40and 50 observed in the LS5A and F82Y/L85A structures, respectively.Leu9 and Argl3 are two residues in the region of the mutation site which adopt alteredconformations. The average side chain deviations observed are 1.1 A and 2.5 A, respectively(Figure 3.19). The shift of the side chain of Tyr82 to a more solvent exposed position necessitatesthe displacement of the nearby side chain of Argl3. The new conformation of Argl3 as well asthe positional shift of Tyr82 toward the protein surface (Figure 3.19) account for the markedincrease in the solvent exposure of the aromatic ring of Tyr82 (Table 3.11). In contrast, thealtered position of the side chain of Argi3 partially masks the side chain of Leu85 from solventexposure (Table 3.11). A further spatial consequence of Argl3 movement is the displacementof the side chain of Leu9. The conformation observed for Leu9 (Xi = -104°; X2 = 128° ) inthe F82Y protein corresponds well to an alternative conformation observed for this side chainin the high resolution wild-type structure (Xi = -111°; X2 = 123°; Louie & Brayer, 1990). Inwild-type yeast iso-i cytochrome c, this alternative site appears to have an occupancy of -.30%.Chapter 3. Roles of Residues 82 and 85 in Cytochrome c 76AL9ByL9D2Y)ji3Figure 3.19: Stereo diagrams showing two views of the region about Tyr82 in F82Y iso-icytochrome c. In each diagram the wild-type protein structure has been superimposed and isshown with thick lines, while the mutant protein structure is depicted by thin lines. Alteredside chain conformations for Leu9, Argi3 and Tyr82 are clearly evident.L94Chapter 3. Roles of Residues 82 and 85 in Cytochrome c 77Table 3.10: Average thermal factors (A2) of residues 80 through 85 in F82Y, L85A, F82Y/L85Aand wild-type yeast iso-i cytochromes cYeast iso-i cytochrome c structureAtom groups Wild-type F82Y L85A F82Y/L85AMet8O main chain 9.9 9.5 10.7 7.4Met8O side chain 5.2 5.5 8.0 6.2A1a81 12.0 16.7 11.5 10.7Phe/Tyr82 main chain 15.2 20.4 13.6 12.1Phe/Tyr82 side chain 16.9 25.2 10.9 13.2Gly83 19.0 21.2 16.4 12.4G1y84 17.4 21.2 17.2 14.3Leu/Ala85 main chain 13.8 16.3 12.8 13.6Leu/A1a85 side chain 17.9 12.6 12.9 8.9All heme atoms 5.3 6.0 7.9 7.6The average thermal factor for all protein atoms of the wild-type iso-i cytochrome c structurewas 16.5 A2. The distributions of protein atomic thermal factors for the three mutants werenormalized to this value for this comparison.Chapter 3. Roles of Residues 82 and 85 in Cytochrome c 78Table 3.11: Solvent accessibility of residues 82 and 85 in F82Y, L85A, F82Y/L85A and wild-typeyeast iso-i cytochromes cYeast iso-i cytochrome c structureWild-type F82Y L85A F82Y/L85AtSide chain surface area exposure (A2)Phe/Tyr82 CD2 10.9 11.8 8.6 10.3Phe/Tyr82 CE2 6.8 14.2 11.4 12.9Tyr82 OH — 6.9 — 5.7Leu85 CD2 13.3 8.7 — —Total Phe/Tyr82 aromatic ringexposure (A2)(atoms CD2 and CE2) 17.7 26.0 20.0 23.2Solvent exposed areas were calculated by the method of Connolly (1983) with aprobe sphere of 1.4 A radius.t The internal water molecules numbered 24 and 248 were considered to be anintegral part of the protein for this calculation.Other prominent features in Figure 3. 18a involving side chain groups represent amino acids inthe disordered amino terminal region of the polypeptide chain or polar side chains extendingfrom the protein into solvent (Lys54, Lys86 and LyslOO).3.3.2 Structure of L85A cytochrome cMutation of leucine 85 to a much smaller alanine residue would be expected to create considerable free volume in this region. The current structural studies show that in response, theChapter 3. Roles of Residues 82 and 85 in Cytochronie c 79ABFigure 3.20: Stereo diagrams showing two views of the region about A1a85 in L85A iso-icytochrome c. In each diagram the wild-type protein structure has been superimposed and isshown with thick lines, while the mutant protein structure is depicted by thin lines. Alteredside chain conformations for Leu9, Argi3 and Leu94 are clearly evident.L94D1’13Chapter 3. Roles of Residues 82 and 85 in Cytoclirome c 80nearby side chains of Leu9, Argl3 and Leu94 have all adopted new conformations (Figures 3.18band 3.20). The side chain of Leu94 is found to shift (zd = 1.8 A ) into the additional spaceavailable near A1a85. The side chain of Argl3 also moves toward A1a85 (Ld = 2.4 A ) and atthis alternate location forms a new interaction with the side chain of Asp9O (Argl3 NH1-Asp9OOD1; d = 3.6 A ). This new interaction has a marked effect on the thermal factors of the atomsof the Argl3 side chain, which have an average B of 32.4 A2 in the wild-type structure and16.7 A2 in the L85A protein. The movements of both Leu94 and Argl3 appear to cause a shiftin the side chain of Leu9 away from the mutation site (M = 1.9 A ) in order to avoid spatialconflicts with these residues. The new conformation of Leu9 differs from both of those seen inthe wild-type protein (Louie & Brayer, 1990).3.3.3 Structure of F82Y/L85A cytochrome cAs can be seen in Figure 3.21, in the F82Y/L85A mutant protein the removal of the side chainof Leu85 allows the additional hydroxyl group of Tyr82 to be accommodated with minimalperturbation of the overall position of the phenyl ring of this residue (zd 0.4 A ). Thishydroxyl group is found to point directly into the region formerly occupied by the Leu85 sidechain in the wild-type protein. In addition, two new water molecules are bound in this region,one of which (Wat 248) forms a hydrogen bond to the hydroxyl group of Tyr82. The secondwater molecule (Wat224) hydrogen bonds to the guanidinium group of Argl3 as well as toWat248 (Figure 3.21).As observed for the L85A protein, the side chain of Argl3 in the F82Y/L85A double mutantmoves toward the region vacated by the Leu85 side chain (i..d = 2.6 A ) and forms a newinteraction with the side chain of Asp9O (Argl3 NH1-Asp9O OD1; d = 3.3 A ). The averagethermal factor for Argl3 side chain atoms is 19.3 A2 and, as in the L85A mutant protein,this is considerably lower than found in the wild-type protein. New conformations are alsoassumed by both Leu9 (zd = 1.5 A ) and Leu94 (Ld = 1.5 A ). As seen in Figure 3.21, theLeu9 side chain has shifted away from the mutation site to avoid spatial conflicts with theChapter 3. Roles of Residues 82 and 85 in Cytochrome c 81ABD)Figure 3.21: Stereo diagrams showing two views of the region about the mutated residues inF82Y/L85A iso-I cytochrome c. In each diagram the wild-type protein structure has beensuperimposed and is shown with thick lines, while the mutant protein structure is depicted bythin lines. Altered side chain conformations for Leu9, Argl3 and Leu94 are clearly evident.Two internally bound water molecules (Wat224 and Wat248) observed in the mutant proteinare depicted as asterisks.Chapter 3. Roles of Residues 82 and 85 in Cytochrome c 82repositioned side chain of Argi3. Unlike the Leu94 shift observed in the L85A mutant protein,in the F82Y/L85A double mutant, reorientation of this residue is more localized likely due tothe two newly bound internal water molecules in this region.3.4 Discussion3.4.1 Structural consequences of residue 82 and 85 mutationsA summary of the structural changes seen in the F82Y, L85A and F82Y/L85A mutants ofyeast iso-i cytochrome c is given in Table 3.12. From this overview it can be seen that mostof the structural changes present are localized to the region about the mutation sites. In theF82Y protein, increased spatial requirements for the larger tyrosyl side chain appear to be thedominant factor and lead to the rotation of the side chain of Tyr82 toward the protein surfacewhere its hydroxyl group hydrogen bonds to a surface water molecule. The increased thermalmobifity of the side chain of this residue is indicative of less than optimal packing of this sidechain with neighboring amino acids which may explain the observed decrease in the stability ofthis mutant protein (Greene et al., 1993). Interestingly, earlier modeling and NMR studies ofthe F82Y mutant protein suggested that the Tyr82 side chain occupied a position equivalentto that of the Phe82 side chain in wild-type iso-i cytochrome c and that the side chain ofLeu85 had an altered conformation (Pielak et al., 1988; Greene et al., 1993). These resultsare in disagreement with the present structural analysis. However, the clarity of the currentFmutant — Fwjld_type (Figure 2.i3a), fragment deleted and 2F0 — F difference electron densitymaps in the vicinity of residues 82 and 85 clearly indicate that Tyr82 as opposed to Leu85 isshifted in the F82Y mutant protein. It should be noted that crystaffine lattice contacts arenot likely to be a potential factor in the disagreement with earlier data since the region aboutLeu85 and Tyr82 is unconstrained by such interactions.The induced structural incompatibility between Tyr82 and Leu85 in the F82Y mutant protein is relieved in the F82Y/L85A mutant protein. The mutation of Leu85 to aianine opens upadditional space next to the Tyr82 side chain so that it can retain a conformation comparableTable3.12:SummaryofpositionaldifferencesobservedbetweentheF82Y,L85AandF82Y/L85Amutantyeastiso-icytochromescandthewild-typeproteinF82YL85AF82Y/L85A1.Tyr82sidechainshiftstowardssurface1.Leu94sidechainmovestowardsA1a851.Leu9andLeu94sidechainsadoptnewofproteinanditshydroxylgroupinter-conformationsactswithasurfacewatermolecule2.Leu9andArgl3sidechainsadoptnewconformations2.Argl3andAsp9Osidechainsforma2.Tyrosyl82ringrotates14°relativetonewinteractionhemeplane3.Argl3andAsp9Osidechainsformanewinteraction3.Twowatermoleculesareintroduced3.Leu9andArgi3sidechainsadoptnewintothemutationsiteconformations4.Tyr82retainsoriginalPhe82positionwithitshydroxylgroupinteractingwithanewlyintroducedinternalwatermoleculeC cI ct 0oChapter 3. Roles of Residues 82 and 85 in Cytochrome C 84to that of the normally resident Phe82. This additional space also allows the hydrogen bondingpotential of the hydroxyl group of Tyr82 to be satisfied by permitting the binding of two newinternal water molecules. In the L85A mutant protein, the hydrophobic cavity that otherwisewould be created is filled by the side chain of Leu94 and the nonpolar portion of the side chainof Argi3. No internal water molecules are found in this highly hydrophobic interior region andthe positioning of the aromatic group of Phe82 is not affected.A further point of interest arising from these studies is that the conformation of Argi3 isclearly dependent on the presence of both Phe82 and Leu85 in the structure of wild-type yeastiso-i cytochrome c. One role of the side chains of Phe82 and Leu85 may be to restrict thepotential interaction of Argi3 and Asp9O in order for the former residue to be fully availableat the interactive face of cytochrome c with redox partners (Pelletier & Kraut, 1992; Northrupet al., 1993; Guillemette et al., 1994; Huang et aL, 1994). Shifts in the side chain of Argi3 havealso been observed in other mutant proteins with replacements at Phe82 (Louie et al., 1988b;Louie & Brayer, 1989). Overall, the structural changes observed for the three mutant proteinsexamined in this chapter are directly related to satisfying the spatial requirements of new sidechains, the minimization of unoccupied internal space and the solvent exposure of hydrophobicgroups, as well as attaining the maximal hydrogen bonding between available polar groups.3.4.2 Impact of mutations on reduction potentialThe dielectric constant of the heme pocket is a primary determinant in setting the reductionpotential of cytochrome c (Kassner, 1973; Louie & Brayer, 1989; Moore & Pettigrew, 1990).Two essential factors affecting the heme pocket dielectric constant are the solvent exposure ofthe heme moiety and the polarity of amino acids in close proximity to this group. One importantstructural element in this regard is the side chain of Phe82. For example, the mutation of thisresidue to serine leads to the creation of a solvent channel into the heme pocket (Louie et aL,1988b), thereby dramatically increasing heme solvent exposure and decreasing the observedreduction potential of this mutant protein (iEm = -35 mV; Rafferty et al., 1990). Similarly,Chapter 3. Roles of Residues 82 and 85 in Cytochrome c 85the introduction of polar groups into the heme pocket, occurring as the result of polypeptiderearrangements in the F82G mutant protein (Louie & Brayer, 1989), has a pronounced effecton reduction potential (LEm = -43 mV). By examining a wide range of side chain replacementsfor Phe82, it is possible to determine that this residue’s contribution to the observed +290 mVreduction potential of wild-type yeast iso-i cytochrome c is ‘—+40 mV (Rafferty et al., 1990).Replacement of Phe82 by a tyrosine has a modest impact on heme reduction potential (tEm= -10 mV; Table 3.13). The increased polarity of the tyrosyl side chain might be expected tohave a larger effect, but this is apparently offset by a small positional shift in this group sothat the tyrosyl hydroxyl group is oriented out toward the protein surface where it can interactwith solvent molecules (Table 3.11). In this conformation, the Tyr82 hydroxyl group is quitedistant (d = 6.5 A ) from the porphyrin ring of the heme and is shielded from direct hemeplane contact by the CBB side chain atom of the heme group. In addition, the hydrophobicityof the heme pocket remains intact since the tyrosyl ring remains in position to shield the hemefrom any significant increase in solvent exposure (Table 3.9).In both the L85A and F82Y/L85A mutant cytochromes c, even smaller effects are seen onheme reduction potential (Table 3.13). For the L85A mutant protein, this is to be expectedsince the side chain of Phe82 retains its wild-type conformation and the internal cavity left byreplacement of Leu85 is filled by the hydrophobic portions of other side chains (Figure 3.20).Thus the hydrophobic integrity of the heme pocket is preserved in this mutant protein and thereduction potential is not significantly perturbed.Although the hydroxyl group of Tyr82 takes on a more internal positioning and two newwater molecules are bound in the FS2Y/L85A mutant (Figure 3.21, Table 3.9), this has onlya small effect on heme reduction potential (Table 3.13). This result is somewhat surprisinggiven the larger impacts of the glycine and serine substitutions at Phe82 (Louie et al., 1988b;Louie & Brayer, 1989). However this might be explained by the fact that the tyrosyl hydroxyland the two newly bound internal water molecules which contribute to increased polarity inthe mutation site are substantially further removed from the heme plane than are the polarChapter 3. Roles of Residues 82 and 85 in Cytochrome c 86Table 3.13: Reduction potentials for F82Y, L85A, F82Y/L85A and wild-type yeast iso-i cytochromes cCytochrome C Em (mV)Wild-type 290±2F82Y 280±2L85A 285±3F82Y/L85A 283±3Experimental conditions were: 25°C, pH 6.0 and i = 0.1 M. Values are listed relative to a standard hydrogen electrode reference. The first two values listed were taken from Rafferty et al.(1990) while the last two values have resulted from collaborative work performed in conjunctionwith this thesis and presented in Guillemette et al. (1994).groups introduced in the F82S and F82G mutant proteins. In these latter proteins, the newlyintroduced polar groups not only are located immediately adjacent to the central face of theheme plane, but are also near the central heme iron atom. In the F82Y/L85A protein, thetyrosyl hydroxyl group is projected away from the heme porphyrin ring (d = 6.5 A ) and thenew internal water molecules are located along the heme edge where heme substituent sidechains shield direct access to the porphyrin ring. The shortest heme porphyrin ring plane towater distance is 5.4 A. Thus it appears that the increased local polarity introduced into theF82Y/L85A mutant protein is too distant from the central heme porphyrin core to substantiallyaffect heme reduction potential.Chapter 3. Roles ofResidues 82 and 85 in Cytochrome c 873.4.3 Electron transfer in mutant proteinsThe steady-state rate of the electron transfer reaction of yeast iso-i cytochrome c with yeastcytochrome c peroxidase varies considerably upon mutation of Phe82, with the F82Y, F82S andF82G proteins having 30%, 70% and 20% of wild-type activity, respectively (Pielak et aL, 1985).Similar trends were observed for the activity of these mutants in the electron transfer reactionwith bovine cytochrome c oxidase (Michel et al., 1989). Such decreases in reaction rate may arisefrom perturbation of either the intrinsic rate of electron transfer within these protein-proteincomplexes or the interactions that are a part of the formation of such complexes. Recent studieshave shown that the multiphasic kinetics observed in the electron transfer reaction betweenreduced iso-i cytochrome c and a Zn-substituted cytochrome c peroxidase porphyrin ir cationradical does not require the presence of an aromatic residue at position 82 of cytochrome c(Everest et al., 1991). Studies of electron transfer between bovine cytochrome c oxidase andiso-i cytochromes c with replacements of Phe82 produced similar conclusions (Hazzard et al.,1992).The fact that the kinetics of intracomplex electron transfer involving Phe82 mutants ofiso-i cytochrome c do not correspond to the steady-state activity decreases observed for theseproteins suggests that complex formation with redox partners is the major factor influencingthese changes in steady-state activity. This is supported by the proposed interactive surface ofcytochrome c as elucidated by modeling of such complexes (Poulos & Kraut, 1980; bum et al.,1987; Northrup et al., 1988) and the recent determination of the structure of the complexformed between yeast iso-i cytochrome c and cytochrome c peroxidase (Pelletier & Kraut,1992). In this latter complex, Phe82 is completely sequestered within the region of protein-protein interactions, although the side chain of Argi3 and several water molecules preventPhe82 from making direct contact with the cytochrome c peroxidase molecule. In the case ofthe F82Y mutant protein, the side chain of Tyr82 would protrude off the surface of the proteininto the site of complexation and, in conjunction with corresponding movements of the Argi3side chain, would disrupt the formation of a productive electron transfer complex. This couldChapter 3. Roles of Residues 82 and 85 in Cytochrome C 88explain the considerably lower steady-state activity of this mutant protein. In contrast, themutation of Phe82 to serine does not present such a physical intrusion into the interactionregion of the complex and steady-state activity is affected to a lesser extent. The lower rate ofthe Ser82 protein could be attributed to the increase of solvent molecules in this region leadingto somewhat less specific complex surface interactions. Another example of a large change inthe contour of the complex contact surface can be found in the mutation of Phe82 to glycine inwhich the exposed face of cytochrome c is drastically altered (Louie & Brayer, 1989). As wouldbe expected from the present work, this greatly reduces the steady-state electron transfer rate.Thus there appears to be a close relationship between the integrity of the complexationsurface of cytochrome c and the overall electron transfer activity observed. The importance ofthis factor is further emphasized by the observed species specificity in attaining optimal proteinprotein electron transfer reactions involving cytochrome c (Ho et al., 1985; Nocek et al., 1991;Moench et al., 1992, 1993). Such effects on electron transfer most likely arise from differences inthe makeup of the interactive surface of cytochrome c between species, not unlike the structuralchanges that have been introduced by site-directed mutagenesis.Chapter 4Replacements in a Conserved Leucine Cluster in the Hydrophobic Heme Pocketof Cytochrome c4.1 IntroductionIn Chapter 3, the structural and functional effects arising from the introduction of the F82Yand L85A substitutions into cytochrome c were examined along with the complementaryF82Y/L85A double mutant protein having both of these amino acid replacements. The objective of the experiments presented in this chapter is to expand on studies of the role of LeuS5in cytochrome c as well as the role of the adjacent Leu94 residue. These two leucine residues,along with leucines 9, 68 and 98, form a cluster of conserved leucines in the hydrophobic hemepocket of cytochrome c. This leucine cluster is also adjacent to, and in part forms the boundaryof, an internal hydrophobic cavity that is positioned toward the back of the heme pocket nearone edge of this moiety. In addition to these features, previous studies have implicated Leu85as being a primary hydrophobic contact point in complexes formed between cytochrome c andits electron transfer partners. Evidence for this arises not only from the results of modelingstudies (Salemme, 1976; Northrup et at., 1993) but also from the results of studies utilizingNMR that indicate a shift in the side chain of residue 85 in the complex of cytochrome c withcytochrome b5 (Burch et al., 1990). It has also been pointed out that amino acid replacementsor conformational shifts at Leu85 may influence the kinetics of cytochrome c function as thesewould require different packing restraints with complexed electron transfer partners (Pielaket at., 1988; Nocek et at., 1991). Such replacements and conformational shifts of Leu85 havepreviously been examined in conjunction with structure-function studies of the adjacent Phe82residue (see Chapter 3; Pielak et at., 1988; Greene et at., 1993).89Chapter 4. Replacement of Conserved Leucines in Cytochrome c 90To evaluate Leu85 from both the structural and functional perspectives, a family of variantproteins has been constructed in which this residue is replaced with several alternative aminoacids. In almost all species for which the amino acid sequence of cytochrome c is known, residue85 is either a leucine or an isoleucine. In those few instances where this is not the case, thisresidue is phenylalanine or methionine (Hampsey et at., 1988; Moore & Pettigrew, 1990). Pastwork has shown the conformation of an isoleucine at residue 85 as the result of the structuredetermination of horse cytochrome c (Bushnell et at., 1990). In addition, the structure of theL85A variant of yeast iso-i cytochrome c was described in Chapter 3 as part of an assessmentof the role of the nearby invariant Phe82 residue.In the current chapter, the L85F and L85M variants of yeast iso-i cytochrome c, containing the two other residues that naturally occur at this position, are examined. Both of thesereplacement amino acids should alter the interactive face of cytochrome c in this region significantly and therefore act as probes of complexation interactions. These substitutions, one beingaromatic and the other a longer unbranched hydrophobic residue, are also likely to have alteredpacking interactions within the leucine cluster in which the side chaln of Leu85 normally resides.Also constructed was the L85C variant, with the cysteine side chain potentially providing twouseful functions. First, as a considerably smaller side chain one would expect the formationof a cavity within the leucine cluster but with a more moderate effect than that observed forthe L85A substitution (see Chapter 3). Second, the L85C mutation provides the opportunityto form a crosslink, via a disulfide bond, to a suitably altered and complexed electron transferpartner in future experiments.Leu94 is also of interest, not only because it is part of the leucine cluster in yeast iso-icytochrome c, but in addition its side chain packs directly against that of Leu85 and the central heme group (Louie & Brayer, 1990). Also of importance is the location of this residue atthe highly conserved interface formed by the nearly perpendicular packing of the N-terminal(residues 2 to 14) and C-terminal (residues 87 to 102) helices in the structure of cytochrome c.Chapter 4. Replacement of Conserved Leucines in Cytochrome c 91This helix-helix packing arrangement has been linked to early events in the folding of cytochrome c (Roder et al., 1988).An analysis of 100 cytochrome c amino acid sequences from different organisms (Hampseyet at., 1988; Fredericks & Pielak, 1993) shows that Leu94 is highly conserved, with a leucineoccurring at this position in 95 sequences, an isoleucine in 3 sequences and a valine in 2 sequences. Interestingly, mutagenic studies have shown substitution of serine at residue 94 orsimilar substitutions at the related leucine cluster residues Leu68 and Leu98 has a significantand deleterious effect on yeast iso-i cytochrome c function (Hampsey et at., 1986, 1988). Thusdespite nearly normal amounts of cytochrome c being observed for the L94S variant, only 30%of the normal yeast growth rate was observed at 22°C. At 30°C, substantially less folded protein was present, indicating that this substitution had an effect on protein stability and thatthe variant protein was thermally labile (Hampsey et at., 1988). This led to the proposal thatthe L94S substitution prevents proper folding of cytochrome c. Similar results were obtainedby Fredericks and Pielak (1993) who examined a wider range of substitutions at Leu94. Theseauthors propose that the helix-helix interface at Leu94 is couformationally flexible and shiftsin helical positions adjust to substitutions made in residues in this region. To assess the degreeto which this helix interface can adjust and at the same time study the role of Leu94 in thefunction of cytochrome c and as part of the hydrophobic leucine cluster in the heme pocket,yeast iso-i cytochrome c with the substitution of serine at this position has been constructedand examined in this chapter.At this point, little attention has been focused on the contributions of hydrophobic interactions on the stability of cytochrome c complexes with electron transfer partners or the rolesthat a heme pocket leucine cluster, an adjacent internal cavity and a nearby helix-helix interfaceplay in the function of this protein. Previous studies have shown that two residues important tothese features are leucines 85 and 94. To gain further insight into their functional and structuralroles in cytochrome c, the work in this chapter examines a total of four variant proteins withsubstitutions at these residues. This work has been published in Lo et at. (i995c).Chapter 4. Replacement of Conserved Leucines in Cytochrome c 92Table 4.14: Data collection parameters for the L85C, L85F, L85M and L94S mutant yeast iso-icytochromes cIso-i cytochrome c mutantParameter L85C L85F L85M L94SSpace group P4321 P4321 P4321 P4321Cell dimensions (A)a = b 36.51 36.55 36.46 36.50c 137.11 138.24 137.55 136.63Data collection method RAXIS CAD4 CAD4 CAD4Number of reflections collected 35500 12913 12574 10787Number of unique reflections 7920 6435 6741 6197Resolution (A) 1.81 1.9 1.9 1.94.2 Experimental ProceduresThe L85C, L85F, L85M and L94S yeast iso-i cytochrome c mutant proteins were initiallycrystallized by the hanging drop vapour diffusion method under reducing conditions as discussedin Section 2.3. Diffraction quality crystals were obtained after further seeding with micro-crystals (Leung et al., 1989). All crystals obtained were of the space group P4321and werefound to be isomorphous with those of wild-type yeast iso-i cytochrome c. Unit cell dimensionsobtained for mutant protein crystals are listed in Table 4.14.Each of the X-ray diffraction data sets for the L85F, L85M and L94S mutant proteins wascollected from a single crystal on an Enraf Nonius CAD4-Fii diffractometer as described inSection 2.5.2. The incident radiation was generated by a nickel-filtered copper-target X-raytube operating at 26 mA and 40 kV, with the crystal to counter distance set to 36.8 cm andthe ambient temperature maintained at 15°C during data collection. Corrections to intensityChapter 4. Replacement of Conserved Leucines in Cytochrome c 93data sets were applied to account for background radiation (Murphy et at., 1992), absorption(North et at., 1968), crystal decay (Louie et at., 1988a), and Lorentz and polarization effects asdescribed in Section 2.5.2. Diffraction data for the L85C mutant protein was collected from asingle crystal on a Rigaku R-AXIS II imaging plate area detector as described in Section 2.5.3.The incident radiation was generated by a RU-300 rotating anode generator operating at 100 mAand 60 kV. Crystals were oscillated through a 4 angle of 1.00 for each frame. Processing ofthis data set was carried out using the R-AXIS II data processing software (Higashi, 1990; Satoet at., 1992) as described in Section 2.5.3. All four mutant protein data sets were put on anabsolute scale by inspection of Wilson (1942) plots as discussed in Section 2.5.4.The crystals used for the collection of the diffraction data sets for the L85F and L94S mutantsof yeast iso-i cytochrome c were particularly small, as indicated in Table 2.3. The small sizeof these crystals was reflected in the relatively weak intensity data measured for these mutantprotein data sets and the correspondingly higher number of unobserved reflections. Althoughthe size of the crystal used in the collection of the L8SC mutant protein diffraction data setwas comparable to that of the L85F mutant protein (Table 2.3), this data collection benefitedgreatly from the increased intensity of the incident X-ray beam produced by the rotating anodegenerator of the R-AXIS area detector and the improved sensitivity of the imaging plate areadetector, as discussed in Sections 2.5.3 and 2.6.3.1. Thus a correspondingly more complete dataset to higher resolution was obtained for the L85C mutant protein (Table 4.15; Figure 4.22).Starting models for structural refinement of mutant proteins were constructed from thecoordinates of the wild-type iso-i cytochrome c protein (Louie & Brayer, 1990). This wasassisted by examining Fmutant Fwild_type difference electron density maps (Equation 2.6; Section 2.4.3.1). The replacement of Leu85 by cysteine and methionine, and the replacement ofLeu94 by serine could be readily built into the starting models based on these difference electrondensity maps. For the L85F structure, the — FjId_type difference electron density mapwas less clear and this residue was initially modeled as an alanine residue. Also included inthe starting models used for structural refinement of the mutant proteins were water moleculesChapter 4. Replacement of Conserved Leucines in Cytochrome c 94from the wild-type iso-i cytochrome c structure having isotropic thermal factors below 50 A2and a sulphate anion bound to the amino terminal end of the N-terminal helix of this structure.The four cytochrome c mutant structures were refined using a restrained parameter least-squares approach (Hendrickson, 1985) following the methods described in Sections 2.6.3.1 and2.6 and using the restraint weights listed in Table 2.4. The solvent water molecules includedin structural refinement were modeled as neutral oxygen atoms with full occupancies. Duringrefinement, as many as 80 individual cycles of least-squares refinement were carried out foreach mutant protein structure. The side chain of Phe85 in the L85F mutant protein structure,which as discussed above could not initially be satisfactorily modeled, was clearly visualized ina 2F0 — F difference electron density map obtained after 11 cycles of least-squares refinementand the appropriate missing side chain atoms were added at this point. During the refinementof each mutant protein structure, at least 3, and in some cases up to 6 (L8SF structure),complete examinations of the entire polypeptide chain and all associated solvent moleculeswere made with the aid of omit, F0 — F and 2F0 — F difference electron density maps to allowfor manual corrections to the refinement models. In addition to these overall examinationsof the polypeptide chain structure, less comprehensive manual interventions were carried outduring the course of the refinements, consisting for the most part of the adjustment of theconformations of individual side chains and the addition and deletion of water molecules. Awater molecule was included in the structural model if it was found to be within 3.5 A of ahydrogen bond donor or acceptor atom and refined to a thermal factor of less than 50 A2.The final refinement parameters and structural stereochemistry for all four mutant proteinstructures are summarized in Table 4.15.Atomic coordinate errors for the four structures studied were estimated by two methodsas described in Section 2.6.3.4. Figure 4.22 shows the resultant Luzzati (1952) plot indicatingcoordinate errors ranging between 0.16 A and 0.22 A for these structures. The Criiickshank(1949) method produces overall atomic coordinate errors of 0.12 A, 0.18 A, 0.15 A and 0.18 Afor the L85C, L85F, L85M and L94S structures, respectively.Chapter 4. Replacement of Conserved Leucines in Cytochrome c 95Table 4.15: Refinement results and stereochemistry for the L85C, L85F, L85M and L94S yeastiso-i cytochrome c mutant structuresL85C L85F L85M L94S1. Refinement resultsResolution range (A) 6.0—1.81 6.0—1.9 6.0—1.9 6.0—1.9Number of observed reflections 7614 4466 5210 4223Completeness in resolution range (%) 85.8 57.2 67.3 55.3Number of protein atoms 892 897 894 892Number of solvent atoms 76 63 69 83Average thermal factors (A2)Protein atoms 21.6 15.9 15.3 14.6Solvent atoms 33.3 24.4 27.6 24.7R-factor 0.200 0.190 0.178 0.1892. Stereochemistry of final modelsr.m.s. deviation from ideal valuesDistances (A)Bond (1-2) 0.019 0.019 0.019 0.019Angle (1-3) 0.037 0.040 0.039 0.039Planar (1-4) 0.049 0.047 0.048 0.048Planes (A) 0.014 0.013 0.015 0.014Chiral volumes (A3) 0.146 0.160 0.156 0.166Non-bonded contacts (A)fSingle torsion 0.211 0.214 0.212 0.216Multiple torsion 0.188 0.202 0.186 0.194Possible hydrogen bonds 0.194 0.225 0.198 0.236Torsion angles (°)Planar (0° or 180°) 2.3 1.9 2.5 2.2Staggered (+60°,180°) 20.3 23.2 19.2 21.9Orthonormal (±90°) 20.4 22.0 19.7 19.2t The r.m.s. deviations from ideality for this class of restraint incorporates a reduction of 0.2 Afrom the radius of each atom involved in a contact.Chapter 4. Replacement of Conserved Leucines in Cytochrome c 96Resolution (A)i0i3252O 1.7E 4-0.0 Q0.35 -0.30 - 0.24 A0.10— i • i_I_I_I_I —; I I I I I • • I0.05 0.10 0.15 0.20 0.25 0.30(sin 0)!Figure 4.22: A plot of the crystallographic R-factor at the end of refinement as a functionof resolution for the L85C (Q), L85F (v), L85M (s), and L94S (D’) mutants of yeast iso-icytochrome c. The theoretical dependence of R-factor on resolution assuming various levelsof r.m.s. error in the atomic positions of the model (Luzzati, i952) is shown as broken lines.This analysis suggested an overall r.m.s. coordinate error for the mutant structures of between0.16 and 0.22 A. The top portion of this figure (axes at top and right) shows the fraction ofreflections observed and used in refinement as a function of resolution.Chapter 4. Replacement of Conserved Leucines in Cytochrome c 97Table 4.16: Overall average positional deviations (A) between wild-type yeast iso-i cytochrome c and the L85C, L85F, L8SM and L94S mutant proteinsIso-i cytochrome c mutantAtom groups L85C L85F L85M L94SAll common protein atoms 0.19 0.39 0.27 0.30All main chain atoms 0.13 0.25 0.15 0.19All common side chain atoms 0.27 0.57 0.41 0.43All heme atoms 0.10 0.17 0.14 0.134.3 Results4.3.1 Structural comparison of mutant and wild-type cytochromes cTo facilitate structural comparisons, each of the four mutant protein structures was superimposed onto the backbone of the high resolution wild-type structure (Louie & Brayer, 1990) bya least-squares fit of of-carbon atoms. As shown in Table 4.16 and Figure 4.23, the global fold ofthe polypeptide chain of all four mutant proteins is similar to that found in wild-type yeast iso-icytochrome c. Plots of the average positional deviations of residues along the polypeptide chainof the mutant proteins are shown in Figure 4.24. The N-terminal residues, Thr(-5) throughPhe(-3), are largely disordered in the wild-type protein (Louie & Brayer, 1990) and displaylarge positional deviations in the mutant proteins. Other large positional deviations involvethe side chains of Lys(-2), Lys4, Lysil, Lys54, Asn63, G1u66, Lys87, Lys89, Lys99, LysiOOand GluiO3 which are all located on the protein surface and have large thermal factor values.The differences observed for these residues and those at the N-terminus are likely the result ofpositional disorder rather than a consequence of the mutations introduced in this study. AlsoChapter 4. Replacement of Conserved Leucines in Cytochrome c 98ABFigure 4.23: Stereo diagrams of the a-carbon backbones of the wild-type, L85C, L85F, L85Mand L94S iso-i cytochrome c structures in (a) the standard view looking at the heme edge-onand (b) an alternate view looking directly into the mutation sites at residues 85 and 94. Alsodrawn are the side chains of the mutated residues, Leu85 and Leu94, and the heme moietiesof all five proteins, along with the ligands to the heme iron atom (Hisi8 and Met8O) andcysteines 14 and 17, which form covalent tliioether bonds to the heme porphyrin ring. Everyfifth amino acid residue is indicated by its one-letter amino acid designation and sequencenumber. Residue numbering is based on an alignment of the primary sequence of yeast iso-Icytochrome c (Table 1.1) where the amino-terminal residue of this protein is designated asresidue (-5).Chapter 4. Replacement of Conserved Leucines in Cytochrome c 994.0 .,iiI .11.111.11 I ••II•tIII I I I3.5_— 3.02.52.0.•-1 10 20 30 40 50 60 70 80 90 100Residue Number4.0 ..1 I I I I I I LLLL35.3.0,.“‘. :-1 10 20 30 40 50 60 70 80 90 100Residue NumberFigure 4.24: continued on next page.Chapter 4. Replacement of Conserved Leucines in Cytochrome c 100CD0>>Figure 4.24: continued. Plots of average positional deviations from the wild-type yeast iso-icytochrome c structure for the (a) L85C, (b) L85F, (c) L85M and (d) L94S mutant proteins.Thick lines indicate average deviations of main chain atoms while the thin lines indicate averagedeviations of the equivalent side chain atoms. In each diagram the filled circle at residue position104 represents the average positional deviation of the heme group. The plotted horizontaldashed line represents the average positional deviation for all main chain atoms.-1 10 20 30 40 50 60 70 80 90 100Residue Number-1 10 20 30 40 50 60 70 80 90 100Residue NumberChapter 4. Replacement of Conserved Leacines in Cytochrorne c 101examined between the four mutant proteins studied and wild-type yeast iso-i cytochrome cwere overall heme group geometry (Table 4.17) and solvent exposure (Table 4.18). Additionalconformational differences are observed about substituted residues in each mutant cytochrome cand these are discussed below in subsequent sections.4.3.2 Structure of L85C cytochrome cExamination of the amino acid side chains directly in the mutation site reveals that the xitorsion angle of Cys85 (-94°) is similar to that of the normally resident Leu85 in the wild-typestructure (-95°). The SG atom of Cys85 is positioned close to the aromatic ring of Phe82,being only 5.0 A from the ring centroid and 3.7 A from the CZ atom. Thus, this sulfur atomforms a favorable nonpolar interaction with the edge of the Phe82 aromatic ring of the typedescribed by Reid et al. (1985). The sulfur atom of Cys85 also forms a hydrogen bond (d =2.8 A ) to a new water molecule, Wat227, which occupies a position similar to that of the CD2atom of Leu85 in the wild-type protein. This latter group constitutes the branch of the leucineside chain projected toward the surface of the protein (Figure 4.25). Although the smaller sizeof a cysteine side chain at residue 85 might be expected to result in the formation of a directinteraction between the side chains of Argl3 and Asp9O, as observed in the F82Y/L85A andL85A mutant proteins (Chapter 3), this is prevented in the current structure by the interveningWat227 and the positioning of the sulfhydryl moiety of Cys85 (Figure 4.25), which togetherprovide comparable steric bulk to that of the normally resident leucine side chain in the wild-type protein.Although there is no corresponding replacement for the buried CD1 carbon atom of Leu85in the L85C mutant protein and therefore no spatial constraint preventing reorientation ofthe side chain of Leu94, this latter residue does not exhibit the conformational rearrangementseen in the L85A mutant protein (Chapter 3). This may arise since the sulfur atom of Cys85effectively excludes bulk solvent from the hydrophobic interior of the protein unlike the solventchannel formed when an alanine residue is substituted at this location. This would suggest thatChapter 4. Replacement of Conserved Leacines in Cytochrome c 102Table 4.17: Heme geometry of L85C, L85F, L85M, L94S and wild-type yeast iso-i cytochromes cWild-type LS5C L85F L85M L94S1. Angular deviations (°) between the pyrrole nitrogen plane normaland the four individual pyrrole ring plane normals and the hemecoordinate bonds.A 9.4 11.4 10.4 11.5 14.1B 11.1 10.5 10.0 10.7 11.1C 8.8 9.0 13.3 9.1 11.6D 8.1 8.3 8.1 9.4 12.5Fe - His 18 NE2 2.2 3.9 3.5 4.2 4.6Fe - Met8O SD 4.9 3.9 0.9 3.5 1.92. Angular deviations (°) between the porphyrin ring plane normaland the four pyrrole ring plane normals, the pyrrole nitrogenplane normal and the heme coordinate bonds.A 6.7 8.3 6.4 8.7 7.9B 11.9 11.8 12.1 12.2 12.5C 9.8 9.4 11.5 9.2 11.4D 6.0 5.8 4.0 7.3 6.3NNNN 2.6 3.3 4.5 3.3 6.2Fe - Hisl8 NE2 3.2 2.4 1.8 6.7 2.9Fe - Ivlet8O SD 7.5 7.1 5.1 6.7 5.73. Bond distances (A) between the heme iron atom and itssix ligands.Hisi8 NE2 1.98 1.98 2.00 1.97 1.97Met8O SD 2.36 2.28 2.43 2.28 2.34Heme NA 1.97 2.01 2.01 1.98 1.99Heme NB 2.00 2.00 2.01 2.03 2.03Heme NC 1.99 2.01 1.98 2.01 2.01Heme ND 2.01 2.06 2.05 2.04 2.04The pyrrole nitrogen plane is defined by the four pyrrole nitrogens of the heme group. The fourpyrrole ring planes are each defined by the five atoms of the ring and the first carbon atomattached to each of the four carbons of the ring. The porphyrin ring is defined by the five atomsin each of the four pyrrole rings, the four bridging methine carbon atoms, the first carbon atomof each of the eight side chains of the heme and the central iron atom of the heme. The hemeatom nomenclature used in this table follows the conventions of the Protein Data Bank (seeFigure 1.2).Chapter 4. Replacement of Conserved Leucines in Cytochrome c 103Table 4.18: Heme solvent accessibility in L85C, L85F, L85M, L94S and wild-type yeast iso-icytochromes cYeast iso-i cytochrome c structureWild-type L85C L85F L85M L94S1. Solvent accessible heme atoms andsurface area exposed (A2)CHD 2.9 3.1 3.6 3.3 0.0CMC 9.2 9.8 9.7 10.4 9.7CAC 3.4 3.5 2.3 3.2 4.3CBC 20.1 19.5 20.2 19.4 20.9CMD 10.8 11.1 6.8 10.5 7.72. Total heme exposure (A2) . 46.4 47.0 42.6 46.8 42.63. Total heme surface (A2) 513.1 515.8 511.7 516.2 514.54. % heme surface area exposed 9.0 9.1 8.3 9.1 8.3Solvent exposure was determined by the method of Connolly (1983) with a probe sphere havinga 1.4 A radius.Chapter 4. Replacement of Conserved Leucines in Cytochrome c 104AD90 D9B13Figure 4.25: Stereo diagrams showing two views of the region about Cys85 in the L85C mutantprotein. In each diagram the structure of the wild-type protein (thick lines) has been superimposed on the mutant protein structure (thin lines). A water molecule (Wat227) which forms ahydrogen bond (d = 2.8 A; dashed line) to the sulfur atom of Cys85 is depicted by an asterisk.94L9L94L9L94Chapter 4. Replacement of Conserved Leucines in Cytochrome c 105Leu94 side chain reorientation in the L85A protein is not a result of simple spatial compensationcaused by the decrease in side chain size at residue 85, but arises from a need to minimize theexposure of its hydrophobic side chain to bulk solvent.Other conformational changes in the vicinity of the mutation site involve G1y83 and Leu9(Figure 4.24a). For G1y83, the average main chain deviation observed is 0.31 A, which ismore than two times the overall average deviation of 0.13 A between the L85C and wild-typeproteins (Table 4.16). It would seem that G1y83, being part of a flexible double glycine sequencewhich can readily adopt alternate conformations, is responding to the need to attain an optimalpacking arrangement in the vicinity of the mutation site. Such compensatory behavior by G1y83has been previously observed in the F82S (Louie et al., 1988b) and F82G (Louie & Brayer, 1989)mutant proteins. In the wild-type iso-i cytochrome c structure, the side chain of Leu9 is locatedat the protein surface and exhibits two distinct conformations (Louie & Brayer, 1990). Theside chain of Leu9 in the L85C structure adopts neither of these, but has a conformation (Xi-80°; X2 = 153° ) which resembles that of Leu9 in L85A cytochrome c (Xi = -70°; X2 =177°; Chapter 3). In this alternate conformation, the side chain of Leu9 is farther away fromthe newly bound Wat227 than would be the case if this residue were to adopt either of theconformations observed in the wild-type protein. This suggests that reorientation of the sidechain of Leu9 arises in order to decrease the interaction of this aliphatic side chain with Wat227.4.3.3 Structure of L85F cytochrome cIntroduction of a phenylalanine residue at position 85 allows the formation of a direct aromaticring to aromatic ring interaction with Phe82. This occurs between the edge of the phenylgroup of Phe82 and the face of the Phe85 aromatic ring (Figure 4.26), with the CZ atom ofPhe82 being 4.0 A from the centroid of the Phe85 ring. The centroids of these aromatic ringsare 5.4 A apart and the normals of their rings form an angle of 57°; values consistent withenergetically favorable aromatic ring to aromatic ring interactions (Burley & Petsko, 1985;Singh & Thornton, 1985). Notably, the larger size of the phenylalanine replacement of Leu85Chapter 4. Replacement of Conserved Leucines in Cytochrome c 106ABD9 713Figure 4.26: Stereo diagrams showing two views of the region about Phe85 in the L85F mutant protein. In each diagram the structure of the wild-type protein (thick lines) has beensuperimposed on the mutant protein structure (thin lines).L9L9L94L9L94Chapter 4. Replacement of Conserved Leacines in Cytochrome c 107causes the a-carbon of Phe85 to be shifted toward the protein surface (Lid = 0.4 A ), requiringcompensating adjustments in the conformations of nearby polypeptide chain, including themain chain atoms of residues 81 through 86 (Figures 4.24b and 4.26). Surprisingly, these shiftsdo not affect the orientation of the Phe82 side chain (LIXi = 50; X2= 40 ) Retention ofside chain torsional angles in combination with the displacement of main chain segments is notan uncommon response to the mutation-induced repacking of protein interiors (Baldwin et al.,1993).Of the residue side chains which pack directly against the side chain of residue 85, Leu9 andArgl3 undergo the largest shifts upon mutation of Leu85 to phenylalanine (Figure 4.26). Leu9assumes a novel side chain conformation not seen in either the wild-type protein or in othermutants with replacements of Leu85. Other nearby residues such as Leu68, Asp9O and Leu94occupy positions similar to those seen in the wild-type protein (Figure 4.26).4.3.4 Structure of L85M cytochrome cThe xi. torsion angle of Met85 exhibits a conformation (-154°) which differs from all otherreplacements at this position. If Met85 were to adopt a xi. conformation similar to that ofLeu85 in the wild-type protein (Xi= 950 ), the SD atom of this residue would form anunfavorable interaction with the ir-electron cloud of the aromatic ring of Phe82 (Reid et al.,1985). Thus the Met85 orientation observed apparently arises from the need to position thedelta sulfur atom of this residue away from the aromatic ring of Phe82. This is in contrastto the SG atom of Cys85 in the L8SC mutant protein which makes an energetically favorableinteraction with the edge of the side chain of Phe82. The added side chain length of Met85 alsoimpinges upon the region occupied by Leu9 in the wild-type protein. This potential conifictcauses Leu9 to adopt the conformation observed in Figure 4.27 (Xi = -80°; X2 = 164° ) whichis similar to that seen in the L85C and L85A mutant proteins.Chapter 4. Replacement of Conserved Leucines in Cytochrome c 108Figure 4.27: Stereo diagrams showing two views of the region about Met85 in the L85M mutant protein. In each diagram the structure of the wild-type protein (thick lines) has beensuperimposed on the mutant protein structure (thin lines).R 13AB94D9V 7L9L9413L9Chapter 4. Replacement of Conserved Leucines in Cytochrome c 1094.3.5 Structure of L94S cytochrome cIn wild-type yeast iso-i cytochrome c the side chain of Leu94 is buried in the hydrophobic hemepocket at the crossover point between the N (residues 2-14) and C (residues 87-102) terminalhelices (Figure 4.28). To accommodate the hydrophilic nature of a serine side chain at thisposition, a novel conformation is adopted. The Xi torsion angle of Ser94 rotates by -135°with respect to that of Leu94, placing the OG atom in a position to form bifurcated hydrogenbonds to the main chain carbonyl atoms of Asp9O and Arg9i (Figure 4.28). Despite thesenew interactions, both of these carbonyl groups retain their hydrogen bonds to the main chainamide groups of Leu94 and fle95, respectively. Conformations of serine residues in a-heliceswhich allow hydrogen bonding between the side chain hydroxyl group and main chain carbonyloxygen atoms of the previous helical turn are common when these residues are buried in thehydrophobic interiors of proteins (Gray Matthews, 1984).In response to the movement of Ser94 away from the interior of the protein, the side chainof Leu98 shifts toward the region vacated by residue 94 (Figure 4.28), with the CD1 atom ofthis residue moving by 1.3 A. The Leu9 side chain orientation seen in this mutant protein (Xi-129°; X2 = 137° ) corresponds to an alternative conformation of this residue seen in thewild-type protein (Xi = -111°; X2 123° ). It appears that this conformation of Leu9 is favoredto avoid direct contact with the hydroxyl group of Ser94.4.4 Discussion4.4.1 Structural consequences of residue 85 and 94 mutationsAs summarized in Table 4.19, significant structural changes observed in mutants of yeast iso-icytochrome c with replacements of the conserved leucines at positions 85 and 94 are localizedto the vicinity of the mutation sites. These appear to arise primarily as a consequence of thespatial requirements of the new side chains and the need to optimize the new intramolecularinteractions formed. For example both CysS5 and Phe85 are oriented to form energeticallyChapter 4. Replacement of Conserved Leucines in Cytochrome c 110ABFigure 4.28: Stereo diagrams showing two views of the region about Ser94 in the L94S mutantprotein. The structure of the wild-type protein has been superimposed and is shown withthick lines while the mutant protein structure is depicted by thin lines. Altered side chainconformations for Leu9 and Leu98 are clearly evident. Hydrogen bonds formed between thehydroxyl group of Ser94 and the main chain carbonyl oxygen atoms of Asp9O and Arg9l arerepresented by dashed lines. Intra-helical hydrogen bonds formed between these carbonyl groupsand the amide nitrogen atoms of Leu94 and 11e95 are also represented by dashed lines.L9 L9-ITable4.19:SummaryofpositionaldifferencesobservedbetweentheL85C,L85F,L85MandL94Smutantyeastiso-icytochromescandthewild-typeprotein0 I.IL85CL85FL85ML94S1.Cys85sidechaininteracts1.Phe82andPhe85side1.Met85adoptsaconforma-1.Ser94sidechainhydrogenwithedgeofPhe82aro-chainsformanaromatictiontopreventcontactofbondstoAsp9OandArg9lmaticringringtoaromaticringinter-SDatomwithPhe82r-mainchaincarbonylsactionelectroncloud2.Anewwatermoleculeis2.Leu9shiftsawayfromSer94boundinthemutationsite2.Mainchainspanning2.Leu9sidechainadoptsahydroxylgroupresidues81through86un-newconformation3.Leu9sidechainshiftsawaydergoesapositionalshift3.Leu98sidechainadoptsafromnewlyboundwaternewconformation3.Leu9andArgi3sidechains4.Gly83adoptsanewconfor-adoptnewconformationsmationChapter 4. Replacement of Conserved Leucines in Cytochrome c 112favorable interactions with the edge of the aromatic ring of Phe82. The conformation of Cys85is such that the SG atom of this residue interacts with the edge of the aromatic ring of Phe82(Reid et at., 1985) while for Phe85, an edge-face interaction with Phe82 is introduced (Burley &Petsko, 1985; Singh & Thornton, 1985). A similar perpendicular edge-face interaction betweentwo phenylalanine rings was recently noted in a mutant of T4 lysozyme and resulted in thestabilization of that protein (Anderson et at., 1993). Although the precise geometries of thesetwo interactions are somewhat different, it is likely that this interaction in L85F cytochrome cwould have a similar stabilizing effect. In contrast, in the L85M mutant protein, Met85 adoptsa more distant conformation which prevents a potentially unfavorable interaction of its SDatom and the 7r-electron cloud of Phe82 (Reid et at., 1985). Finally, the side chain of Ser94 inthe L94S mutant protein adopts a conformation which allows interaction with two main chaincarbonyl groups of the C-terminal a-helix (residues 90 and 91), thereby satisfying the hydrogenbonding potential of the Ser94 hydroxyl group.4.4.2 Effects on reduction potentialAn essential factor controffing cytochrome c reduction potential is the dielectric constant ofthe heme pocket (Kassner, 1973; Louie & Brayer, 1989; Moore & Pettigrew, 1990). This isdependent in large part on both the solvent exposure of the heme group and the polarityof those residues in the immediate heme environment. For example, the reduction potentialof yeast iso-i cytochrome c can be significantly decreased by an increase in either the solventexposure of the heme moiety (Louie et at., i988b) or the polarity of protein substituents packingdirectly against the heme plane (Louie & Brayer, 1989). However, this effect rapidly drops offwith distance so that the introduction of polar groups such as hydroxyl moieties or internalwater molecules at distances greater than A from the heme porphyrin plane do not result ina significant decrease in reduction potential (Chapter 3).The reduction potential of the L85C mutant protein is very similar to that of wild-typeyeast iso-i cytochrome c (Table 4.20), despite the introduction of a polar side chain and aChapter 4. Replacement of Conserved Leucines in Cytochrome c 113Table 4.20: Reduction potentials for L85C, L85F, L85M, L94S and wild-type yeast iso-i cytochromes cCytochrome c Em (mV)Wild-type 290±2L85A 285+3L85C 288+3L85F 285±2L85M 284+2L94S 280±2Experimental conditions were: 25°C, pH 6.0 and it = 0.1 M. Values are listed relative to a standard hydrogen electrode reference. The first value listed was taken from Rafferty et at. (1990)while the values for the mutant proteins have resulted from collaborative work performed inconjunction with this thesis and published in Lo et at. (1995c) and Guillemette et at. (1994).newly bound water molecule (Figure 4.25). Examination of this structure shows that the sulfuratom of Cys85 is pointed toward the protein surface and is distant from the heme porphyrinring (d 7.1 A ), while Wat227 is located at an even greater distance (d = 9.1 A ). Thus thesenew features of the L85C mutant protein do not change the dielectric constant in the directvicinity of the heme group. A further factor which preserves the hydrophobic pocket about theheme is the manner in which the orientation of the Cys85 side chain excludes entrance of bulksolvent into this region.Replacement of Leu85 by either phenylalanine or methionine also does not result in reductionpotentials which are significantly different from that of the wild-type protein (Table 4.20). ThisChapter 4. Replacement of Conserved Leucines in Cytochrome c 114is to be expected given that the substitutions are hydrophobic and the solvent exposure of theheme groups in these two proteins do not differ greatly from the wild-type protein (Table 4.18).These results indicate that conformational changes occurring to accommodate these amino acidreplacements do not significantly impact the heme environment.Replacement of Leu94 by a serine does result in a modest decrease in the reduction potentiaiof cytochrome C (LE = -10 mV; Table 4.20). This effect is unlikely to arise from changes inside chain polarity since the gamma oxygen atom of Ser94 is distant from the heme porphyrinplane (d = 7.4 A ). Temperature dependence measurements (data not shown) reveal that thedecreased reduction potential of L94S cytochrome c is more the result of changes in the entropyof reduction, LS° (-12.6 e.u. in L94S; -9.1 e.u. in wild-type), rather than changes in theenthalpic contribution, iH° (-14.9 kcal/mol in L94S; -14.0 kcal/mol in wild-type). Two factorsproposed to affect the entropy of reduction are the reordering of solvent around the polypeptidechain (Taniguchi et at., 1980) and alteration of side chain packing in the hydrophobic core(Murphy et at., 1992). It is possible that either or both of these factors may play a role in thedecreased reduction potential of the L94S mutant protein given the structural changes observed.4.4.3 Functional alterationsEarlier studies have implicated residue 85 in the formation of bimolecular electron transfercomplexes involving cytochrome c (Pielak et at., 1988; Burch et at., 1990; Nocek et at., 1991).However, a recent examination of model complexes formed between cytochrome c and cytochrome b5 shows that Leu85 is not likely to reside at the interface between these two proteins(Northrup et at., 1993; Guillemette et at., 1994). Furthermore, measurements of the rate of electron transfer in this complex have revealed no significant changes when Leu85 of cytochrome cis replaced by either cysteine, phenylalanine or methionine (Guillemette et at., 1994). Ourresults support the conclusion that Leu85 is not a major factor in bimolecular complexationwith cytochrome b5, given the absence of functional effects in light of the significant proteinsurface changes observed for the L85C, L85F and L85M mutant proteins.Chapter 4. Replacement of Conserved Leucines in Cytochrome c 115In the case of Leu94 in cytochrome c, mutation to serine has been found to significantlydecrease the growth of yeast (Hampsey et aL, 1986; Fredericks & Pielak, 1993). These effectshave been ascribed to both defective formation of the hydrophobic heme pocket (Hampseyet al., 1986) and alteration of the nature of the interaction between the N (residues 2-14)and C-terminal (residues 87-102) c-helices of cytochrome c (Fredericks & Pielak, 1993). Ourresults show that the L94S mutation does have a direct effect on the hydrophobic heme pocket,precipitating structural changes in the adjacent conserved leucine cluster and the hydrophobicinternal cavity. However, the N and C-terminal helices which intersect in a nearly perpendicularmanner (Figure 4.23) and appear to be an essential component in the folding of cytochrome c(Roder et aL, 1988) are unaffected (Figures 4.24 and 4.28) by the L94S mutation in terms ofstructural integrity and orientation. These results show that the yeast growth rate alterationscaused by this mutation are most likely the result of disruption of the heme pocket region.A study of a mutagenic library has also suggested that the large number of substitutionsobserved at Leu94 and at the adjacent Tyr97 are likely accommodated by shifts in the positionof the interacting N and C-terminal helices (Fredericks & Pielak, 1993). Our results show thatat least in the case of the L94S mutant protein, no such structural change is observed. It is likelythat the presence of a cluster of leucine side chains in this region of the protein, including Leu9,Leu68, Leu85, Leu94 and Leu98, in conjunction with a nearby internal cavity, provide sufficientstructural plasticity such that any of the mutant side chains introduced can be accommodatedwithout shifts in the relative orientations of these two helices.4.4.4 Hydrophobic internal cavity fluctuationsAs illustrated in Figure 4.29, a large internal cavity is found in yeast iso-i cytochrome c,bounded by the side chains of Leu32, 11e35, Met64 and Leu98, and one edge of the hemegroup (Louie & Brayer, 1990). There are no potential hydrogen bonding partners on the cavitysurface and no water molecules are observed to be present. This cavity can be eliminated bythe replacement of Leu98 by a methionine residue (Murphy et aL, 1992) or by a positional shiftChapter 4. Replacement of Conserved Leucines in Cytochrome c 116-9L98Figure 4.29: A stereo diagram showing the location of the hydrophobic internal cavity (dotsurface) found in wild-type iso-i cytochrome c. The heme group and nearby side chains thatdefine the outer limits of this cavity in the L85F (thick lines) and L94S (thin lines) mutantproteins have been superimposed on the structure of the wild-type protein (medium lines). Thedifferent conformations observed for the Leu98 side chain are largely responsible for determiningthe size of this internal cavity. In the L85F mutant protein, the Leu98 side chain moves intothe internal cavity, effectively eliminating this feature. In the L94S mutant protein, the Leu98side chain moves toward the side chain of Ser94, causing an increase in cavity size. The volumeand surface area of this cavity in a wide range of mutant proteins is tabulated in Table 4.21.of the side chain of Leu98 (Murphy et al., 1993). A tabulation of the volume and surface areaof this hydrophobic cavity for a series of mutant proteins having replacements of Leu85 andLeu94, as well as Phe82, is presented in Table 4.21.As evident in Table 4.21 there is a trend for the size of the internal cavity to be affected byvolumetric changes brought about by mutagenesis, even if the site of mutation is not immediately adjacent to the cavity wall. For example, mutations in which one residue is replaced bya larger one (F82Y, L85F) tend to eliminate this cavity. Conversely, substitution of a smallerresidue for the native amino acid (F82G, F82S, L85A, L94S) tends to cause an increase in cavityChapter 4. Replacement of Conserved Leucines in Cytochrome c 117Table 4.21: Surface area and volume of the hydrophobic internal cavity in mutant and wild-typeyeast iso-i cytochromes cProtein Surface area (A2) Volume (A3)Wild-type 40.4 35F82G 54.1 42F82S 50.0 40F82Yt — —F82Y/L85A 25.8 15L85A 47.8 40L85Ct — —L85Ft — —L85M 44.4 31L94S 72.1 56Surface areas and cavity volumes were calculated by the methods of Connolly (1983, 1985) withprobe spheres of radius 1.4 A and 1.2 A, respectively. The values for surface areas and volumesare estimated to have errors of ±5.0 A2 and ±5 A3, respectively. The structure of the wild-typeprotein was from Louie & Brayer (1990); the F82G mutant from Louie & Brayer (1989); theP825 mutant from Louie et al. (1988b); and the F82Y, F82Y/L85A and L85A mutants fromChapter 3.f no cavity could be detected with a 1.4 A radius probe sphere.Chapter 4. Replacement of Conserved Leucines in Cytochrorne c 118size. Furthermore, mutation to a similar sized (L85M) residue seems to help maintain the sizeof the cavity. Although the behavior of the F82Y/L85A and L85C mutants would seem tocontradict these trends, it must be realized that newly bound water molecules are introducedeither into the protein interior (F82Y/L85A; Chapter 3) or onto the protein surface (L85C;Figure 4.25) in these mutants, causing a net increase in steric bulk.The position of the side chain of Leu98 seems in large part to determine cavity size, witheven small shifts of this residue having significant effects. In some cases, this dependence canbe readily explained by changes in immediately adjacent residues such as in the L94S mutantprotein (see Figure 4.29). However, in other mutant proteins, the cause for a positional shiftin Leu98 and a subsequent change in cavity size is not as obvious. It is likely that small,cumulative changes in the positions of both main chain and side chain atoms result from theintroduced mutations, and that there are subtle structural strains arising from these changes.It is possible that the combination of a hydrophobic internal cavity and the presence of thenearby flexible leucine cluster composed of not only Leu98, but also Leu9, Leu68, Leu85 andLeu94, can act as a structural buffer in these cases.Previous studies of residue substitutions resulting in a decrease in side chain size can providesome insight into the current structural results. For example, such mutations in a hydrophobicprotein core can result in either the creation of internal cavities (Eriksson et at., 1992) or thecontraction of the protein around the affected region (Katz & Kossiakoff, 1990; McRee et at.,1990). From our results, we observe that for the internal L94S cytochrome c mutant, the nearbyinternal cavity has been enlarged by the largest amount (Table 4.21) without contraction of theprotein, indicating that the immediate region has considerable structural rigidity. In contrast,the introduction of size decreasing mutations at protein surfaces can result in the reorganizationof the local solvent structure around newly available hydrogen bonding protein moieties (Katz& Kossiakoff, 1990; Nair et at., 1991; Nair & Christianson, 1993) or a combination of positionalshifts in nearby side chains with the introduction of bound solvent molecules (Bone et at., 1989;Varadarajan & Richards, 1992). While the former situation applies in the case of the L85CChapter 4. Replacement of Conserved Leucines in Cytochrome c 119cytochrome c mutant, a unique scenario was observed in the L85A mutant protein (Chapter 3).Without compensating structural changes, this latter mutation would result in the creation ofa hydrophobic invagination exposed to bulk solvent. The presence of the flexible leucine clusterin this region of the protein allows compensatory adjustments which effectively exclude solventfrom the hydrophobic protein interior. Although the effect of these adjustments increasesthe size of the internal cavity of iso-i cytochrome c, the energetic cost involved appears tobe adequately offset by the gain from preserving the hydrophobic environment of the proteininterior. These examples show that the response to structurally disruptive mutations will varydepending on the rigidity of the protein backbone and the flexibility of nearby side chains.Overall, our results suggest that the conserved leucine cluster found in cytochrome c alongwith an internal hydrophobic cavity allows this protein to retain a measure of conformationalflexibifity. Apart from an abifity to adapt to mutations, it is also conceivable that these twofeatures have a role in cytochrome c function from the standpoint of providing the fiexibifityto assist in the structural switch of this protein between oxidation states (Berghuis & Brayer,1992). Many of those residues affected in an oxidation state dependent manner are locatedadjacent to the leucine cluster region, as well as the nearby internal hydrophobic cavity.Chapter 5Replacement of the Invariant Phenylalanine 82 in Cytochrome c by AliphaticResidues5.1 IntroductionThe previous two chapters have dealt with the potential role of Phe82 and Leu85 in forming contact face interactions, as well as looking at the contribution of Leu85 to an internalhydrophobic leucine cluster and a related internal hydrophobic cavity. As part of this latterstudy Leu94 was also investigated since it contributes to both the leucine cluster and cavitystructures in addition to being located at the juncture between the N and C-terminal helicesof the cytochrome c fold. In the present chapter a closer study is made of the role of Phe82, inparticular as this relates to its replacement by large aliphatic amino acids.As discussed in Chapter 3, Phe82 is a phylogenetically invariant residue in cytochrome c(Hampsey et at., 1988; Moore & Pettigrew, 1990) that plays a role in electron transfer complexformation (Pielak et at., 1985; Michel et at., 1989), electron transfer kinetics (Everest et at.,1991; Hazzard et al., 1992) and heme reduction potential (Rafferty et aL, 1990). Previousstructural studies have looked at mutant proteins wherein Phe82 was replaced by glycine (Louie& Brayer, 1989), serine (Louie et at., 1988b) or tyrosine (Chapter 3), but as yet these studieshave not been extended to replacements by large aliphatic side chains.Replacement of Phe82 by leucine, methionine or isoleucine leads to mutant proteins thatexhibit unique functional properties and it is therefore of considerable interest to be able tocorrelate structural studies with these data. For example, these mutant proteins all behave differently in their kinetics of electron transfer with cytochrome b5 (Wiffie et at., 1993; Guillemetteet ci., 1994) and cytochrome c peroxidase (Everest et at., 1991; Nocek et at., 1991). Especially120Chapter 5. Aliphatic Replacements of Phe82 in Cytochrome c 121notable is the F821 mutant protein, in which 50 x tighter binding with cytochrome c oxidaseis observed (Michel et at., 1989) along with strikingly different kinetics of electron transfer(Hazzard et at., 1992).Another aspect of cytochrome c function addressed in the present studies involves thereduction potential of cytochrome c. All three replacement aliphatic side chains are smallerand of different shape than that of the normally resident phenylalanine, and the substitutionof each could be expected to result in the formation of a cavity or even potentially a solventchannel into the heme pocket. This would be expected to produce a concomitant decrease inheme reduction potential. While this result is observed in the F821 mutant protein, albeit toa smaller than expected degree, this is not the case for the F82L and F82M mutant proteins(Rafferty et at., 1990). Thus the question arises as to how the side chains of leucine andmethionine, when substituted for Phe82, act to preserve the hydrophobic nature of the hemepocket. The goal of the present chapter is to address this question and others related to theunique features demonstrated by the F821, F82L and F82M mutants of cytochrome c. Thiswork has been submitted for publication as part of Lo & Brayer (1995).5.2 Experimental Procedures5.2.1 Crystallization and data collectionUsing the conditions detailed in Table 2.3, the F821 and F82M mutant proteins were crystallizedby the hanging drop vapour diffusion method, while the F82L mutant protein was crystallizedby the free interface diffusion technique. Both of these crystallization methods are describedin Section 2.3. All crystaffizations of these proteins were aided by seeding with micro-crystals(Leung et at., 1989). Crystals of the F821 and F82M mutant proteins grew isomorphously tothose of the wild-type protein and belong to the space group P4321,with the cell dimensionslisted in Table 5.22.In contrast, the morphology of the crystals formed by the F82L mutant protein was uniqueand these crystals were not isomorphous to those found for wild-type yeast iso-i cytochrome c.Chapter 5. Aliphatic Replacements of Phe82 in Cytochrome c 122Table 5.22: Data collection parameters for F821, F82L and F82M yeast iso-i cytochromes cIso-i cytochrome c mutantParameter F821 F82L F82MSpace group P4321 I222 P4321Cell dimensions (A)a 36.49 62.92 36.52b 36.49 69.15 36.52c 136.87 104.26 138.33Number of reflections collected 13825 13183 8514Number of unique reflections 6250 5101 3820Merging R-factort 0.069 0.094 0.134Resolution (A) 1.9 2.5 2.1V•’ T. TT Tiil- p ç . — Z_thkz L.d1 hki hklerguig - ac or —1_.hkZ L.,;=iInstead of the normally observed rounded pifiow shape found for crystals of the wild-typeand related mutant proteins, crystals of the F82L mutant protein were rectangular prismswith sharp well defined edges and faces. It is also notable that F82L mutant protein crystalsrequired considerably more time to grow (nearly 2 years) than those of other mutant proteins.Precession photography and subsequent diffraction data collection indicated that these crystalswere of either space group 1222 or I22i2, with the cell dimensions listed in Table 5.22.Unfortunately, these two orthorhombic body-centered space groups are indistinguishable fromone another on the basis of systematic absences, since the absences expected for the 2i screwaxes of space group I222 form a subset of the general absences caused by body-centering inboth space groups. It is of interest to note that if one assumes there is one protein moleculeper asymmetric unit, a solvent content of ‘-72% can be calculated for these crystals (Matthews,Chapter 5. Aliphatic Replacements of Phe82 in Cytochrome c 1231968). A solvent content of ‘.‘44% is expected if the assumption is made that there are twomolecules per asymmetric unit. As both of these situations are theoretically possible, each wasconsidered during the structure solution of this protein, as described below.Diffraction data for the F821, F82L and F82M mutant proteins were collected on a RigakuR-AXIS II imaging plate area detector system with incident radiation provided by a RU-300rotating anode generator operating at 90—100 mA and 50—60 kV, as discussed in Section 2.5.3.Crystals were oscifiated through a 4 angle of 1.00 for each frame, with the X-ray exposuretime ranging between 20 and 30 minutes. X-ray intensity data were processed to structurefactors following the procedures described in Section 2.5.3 using the R-AXIS II data processingsoftware (Higashi, 1990; Sato et al., 1992). The method of Wilson (1942) was used to put eachmutant protein data set on an absolute scale as described in Section 2.5.4.As illustrated in Figure 5.30, an examination of the diffraction data sets obtained for theF821, F82L and F821V1 mutant proteins revealed that the crystals of all three proteins diffractedpoorly, having a rapid fall off in intensities with increased resolution. For the F821 and F82Mmutant proteins this was likely a consequence of the extremely small size of the crystals thatcould be grown (Table 2.3) despite numerous crystallization trials to improve on these. Aprimary factor in successful data collection in these cases was undoubtedly the sensitivity ofthe R-AXIS II area detector and the availability of a rotating anode X-ray source. As discussedin Section 2.6.3.1, earlier attempts to collect diffraction data sets from crystals of these twomutant proteins using a CAD4 diffractometer with a sealed beam X-ray tube were unsuccessfulsince the poor quality of the data sets obtained precluded their use in structural refinement.For the F82L mutant protein, although the best crystal that could be grown was alsorelatively small, it was of a size comparable to a number of other crystals of mutant iso-icytochromes c used in structural analyses (Table 2.3). However, this F82L mutant proteincrystal diffracted considerably poorer than might be expected at higher resolution. This resultmay be related to the very open packing of the alternative crystalline lattice of these crystalsas discussed in later sections. As shown in Figure 5.30, the completeness of the diffraction dataChapter 5. Aliphatic Replacements of Phe82 in Cytochrome c 124Resolution (A)5.03.32.52.0 1.7p0:5-0.0 U0.35 - - 0.28 A0.30 - 0.24 A025 -- ---: 0.20A::: 016A0.10— i i ii J I I I I I I I I I I I I I p I0.05 0.10 0.15 0.20 0.25 0.30(sin 9)! ?,Figure 5.30: A plot of the crystallographic R-factor at the end of refinement as a function ofresolution for the F821 (Q), F82L (v) and F82M (0) mutants of yeast iso-i cytochrome c. Thetheoretical dependence of R-factor on resolution assuming various levels of r.m.s. error in theatomic positions of the model (Luzzati, 1952) is shown as broken lines. This analysis suggestedan overall r.m.s. coordinate error for the mutant structures of between 0.20 and 0.26 A. Thetop portion of this figure (axes at top and right) shows the fraction of reflections observed andused in refinement as a function of resolution.Chapter 5. Aliphatic Replacements of Phe82 in Cytochrome c 125that could be collected for the F82L mutant protein crystal was low beyond 3.0 A resolutionand was oniy 27.6% for the 2.6—2.5 A resolution shell. Even though it was limited, the collectionof diffraction data for the F82L mutant protein was extended to 2.5 A resolution in the hopeof getting a reasonable definition of the protein structure despite the weakness of the data athigh resolution.5.2.2 Structure solution for the F82L mutant protein using molecular replacementmethodsCrystallization of F82L yeast iso-i cytochrome c in a unique space group necessitated the useof the molecular replacement approach in the solution of the structure of this protein. A briefsummary of the theoretical basis for this method is presented in Section 2.4.3.2. The set ofmolecular replacement programs used in these studies was the MERLOT package (Fitzgerald,i988) running on a Silicon Graphics 4D/340 workstation.A key factor in the success of a molecular replacement search is the quality of the initialsearch model. For mutant proteins with a small number of amino acid substitutions, molecularreplacement analyses are facilitated since the search model that can be constructed closelyresembles the target structure. The search model used in this case was based on the highresolution structure of wild-type yeast iso-i cytochrome c (Louie & Brayer, 1990) and consistedof residues -.1 through 103, the heme group and four internal water molecules. The four residuesat the N-terminus were not included since these are largely disordered in the wild-type structure.The four internal water molecules included in the search model, WatilO, Wati2i, Wati66 andWati68, are highly conserved and can be considered to be an integral part of the proteinstructure. The side chain at residue 82 was modeled as an ala.nine since the conformation of aleucine at this position was not known.The search model was initially centered at the origin of a P1 unit cell with a=b=c=90 Aand o=/3=7=90°, and structure factors were calculated to 3.0 A resolution. The thermal factorsof all atoms were assumed to be 15.5 A2, the same as the average of the thermal factors ofChapter 5. Aliphatic Replacements of Phe82 in Cytochrome c 126the wild-type iso-i cytochrome c atoms included in the search model. A fast rotation functionsearch (Crowther, 1972) was carried out with data between 10.0 and 3.5 A resolution. Thissearch was performed for 0.0° a < 180.0° in 2.5° steps, 0.0° < /3 < 90.00 in 1.00 stepsand 0.0° ‘y < 360.0° using 5.0° steps. The radius of the search about the Patterson originwas limited to 19.0 A and origin removal was applied. The highest peak observed was 6.2aabove background, with the next highest peak being of height 4.5o as illustrated in Figure 5.31.Fine scans to optimize the best rotation function solution obtained (Lattman & Love, 1972)employed 1.00 steps spanning +5.0° in each of a, j3 and 7.The translation function of Crowther and Blow (1967) was used to determine the translational positioning of the rotationally oriented model of the F82L mutant protein within thecrystallographic unit cell. Due to crystallographic symmetry, there are four symmetry relatedrotation function solutions in each of the two orthorhombic body-centered space groups. Thetranslation function search was applied to each different pair of rotationally related moleculesutilizing diffraction data between 10.0 and 3.5 A resolution. The height of the largest peak fromeach translation search ranged from 6.5u to 8.3u (Figure 5.32). This set of peaks was consistentwith the I222 space group but not with the 1222 space group. The correctly oriented andtranslated search model obtained from this molecular replacement search resulted in a startingcrystallographic R-factor of 0.398 for data between 6.0 and 3.0 A resolution.Given the large size of the unit cell of the F82L mutant protein crystals (72% solvent content if only one molecule is present in the asymmetric unit), the possibifity of a second proteinmolecule being present within the asymmetric unit was carefully assessed during and after thestudies discussed above. In particular, application of translation function searches to the nexthighest peaks resulting from the rotation function search did not produce any additional solutions. Furthermore, a self rotation search was performed to determine if non-crystallographicsymmetry elements were present. The results of this analysis were also negative. After thecompletion of refinement of the F82L mutant protein structure (discussed in Section 5.2.3), afurther inspection of 2F, — F and F0 — F difference electron density maps also failed to showChapter 5. Aliphatic Replacements of Phe82 in Cytochrome c 127360-y(o)18000 180 0 180Figure 5.31: Two sections of the rotation function map calculated for F82L iso-i cytochrome c.The correct solution (a = 1°, fi = 83°, -y = 82°), having a peak height 6.2o above background,is shown on the left in the map section at ,8 = 83°. For comparison, on the right is shown themap section having the second highest peak found (a = 70°, /3 = 37°, -y = 235°) contoured atthe same levels.II II IChapter 5. Aliphatic Replacements of Phe82 in Cytochrome c 128Figure 5.32: A section of a translation function map for F82L iso-i cytochrome c. The correcttranslation function solution is represented by the highest peak in the map (zIy = 0.75, /z =0.77). This section was calculated at x = 0 for two molecules related to each other by a 1800rotation about the a axis. An equivalent solution exists in the section at ZSx 0.50 representinga molecule translated by half a unit cell length along each of the three orthogonal axes.1.0LZ 0.500 0.5 1.0Chapter 5. Aliphatic Replacements of Phe82 in Cytochrome c 129any evidence for an additional protein molecule in the asymmetric unit. Additional support forthis result was the ability to refine the crystallographic R-factor for this structure to a valuecomparable to that of the other mutant proteins studied.5.2.3 Refinement of mutant protein structuresThe starting models for the refinement of the F821 and F82M mutant proteins consisted of thehigh resolution structure of wild-type yeast iso-i cytochrome c (Louie & Brayer, 1990) in whichresidue 82 was represented as an alanine. With the exception of those near residue 82, all solventmolecules having isotropic thermal factors less than 50 A2 in the wild-type protein structurewere included in these two starting models, as was a bound sulphate anion. For the F82Lmutant protein, the starting model employed was that oriented in the molecular replacementanalysis described in Section 5.2.2 where residue 82 was also represented as an alanine. Initialrefinements of the three mutant proteins were carried out by the restrained parameter least-squares method (Hendrickson, 1985) as described in Section 2.6 using the restraint weightslisted in Table 2.4.lii the case of the F821 mutant protein, the side chain of 11e82 was built into a 2F0 —difference electron density map calculated after 14 cycles of refinement. Interspersed at 5points during a subsequent 66 cycles of refinement were complete examinations of the entirepolypeptide chain and all solvent molecules using omit, F0 — F and 2F0 — F difference electrondensity maps as a guide, with manual adjustments being made as necessary. An iterative peaksearching and refinement procedure was used to identify the likely location of additional watermolecules (Tong et al., 1994). Water molecules which refined to isotropic thermal factors greaterthan o A2 were eliminated from the refinement model, as were those more than 3.5 A fromhydrogen bond donor or acceptor atoms.The structural refinements of the F82L and F82M mutant proteins proved to be considerablymore difficult. For the F82L mutant protein, clearly defined electron density for the side chainof Leu82 was observed in a F0 — F difference electron density map after 24 cycles of refinementChapter 5. Aliphatic Replacements of Phe82 in Cytochrome c 130and on this basis this side chain was fit. The four amino-terminal residues which had been leftout of the molecular replacement model were added after a further 8 cycles of refinement basedon 2F0 — F and F0 — F difference electron density maps. Despite the clarity of the Leu82 sidechain in difference electron density maps, problems arose during refinement in maintaining theconformation of this residue within normally accepted geometric restraints. It proved fruitlessto correct the conformation of this residue through manual fittings since further refinementsimply resulted in a return to the original unacceptable conformation. In addition, the progressof the structural refinement of the F82L mutant protein as a whole proved unsatisfactory asindicated by convergence to a standard crystallographic R-factor of 0.250 after 72 cycles. Forthe F821v1 mutant protein, the side chain of Met82 was placed on the basis of a F0 — F differenceelectron density map calculated after 32 cycles of least-squares refinement. Although refinementof the conformation of Met82 was within accepted norms, the overall progress of the refinementof the F82M mutant protein also proved unsatisfactory as indicated by convergence of thecrystallographic R-factor to an unacceptably high value of 0.237.To overcome the problems encountered in the refinement of the F82L and F82M mutantproteins, simulated annealing refinement using the program X-PLOR (Brünger, 1992) wasperformed. The objective of simulated annealing refinement is to conduct an enhanced searchof conformational space while optimizing the fit of the observed and calculated structure factoramplitudes. The major advantage of this technique is that it permits the structural model accessto a large number of alternative conformations which are beyond the local potential energyminimum and are not accessible in conventional least-squares refinement. A brief descriptionof this method is presented in Section 2.6.2.For the F82L and F82M iso-i cytochrome c structures, the slow-cooling protocol (Briinger,1990) was employed for simulated annealing refinement. The first step in this process involvesthe minimization of both structure factor and energy terms to optimize the conformation of therefinement model. In the second step, velocities are assigned to individual atoms based on thekinetic energy that would be present at an arbitrarily high temperature, in this case 2500 K. ThisChapter 5. Aliphatic Replacements of Phe82 in Cytochrome c 131kinetic energy drives the conformational search of the model, allowing large barriers of potentialenergy to be overcome in the search for a best fit. In the third step, the molecular dynamicssearch is allowed to continue, with the kinetic energy of the system being periodically decreasedin steps of 25 K until an energy representing 300 K is reached. Finally, the structural modelis extensively refined to minimize both the crystallographic R-factor and the potential energyfunction, in effect giving the optimal structure in the local energy minimum. The simulatedannealing refinement of the F82L and F82M mutant proteins resulted in significant overall shiftsin the positions of atoms in these two structures (d = 0.56 A for F82L; id = 0.60 A for F82M).The phases of the calculated structure factors for these structures also underwent significantchanges, averaging 32° and 41° for the F82L and F82M structures, respectively. Surprisingly,the overall crystallographic R-factors resulting from the simulated annealing process were foundto be only slightly lowered (0.236 for F82L; 0.225 for F82M) from starting values.Examination of electron density maps showed that despite the small lowering of R-factor observed, a much improved fit of polypeptide chain for both the F82L and F82M mutant proteinswas observed after the application of simulated annealing refinement. Of special note was theconformation around Leu82, where a shift in nearby polypeptide chain allowed for an improvedfit of this side chain that satisfied all normal stereochemical requirements. As had been hoped,these optimized structural fits allowed conventional restrained parameter least-squares refinement to proceed. Thus each of these two mutant protein structures were subjected to a further60—70 cycles of refinement, in addition to 4 overall examinations of the complete polypeptidechain to allow for manual adjustment using difference electron density maps, as previously outlined for the F821 mutant protein. These manual interventions were concerned with optimizingthe conformations of surface side chains, locating water molecules and adjusting the N-terminalregion of these proteins (residues (-2)—(-5) ). The final parameters and stereochemistry for therefinement of all three cytochrome c mutant protein structures studied in this chapter are givenin Table 5.23.The two methods described in Section 2.6.3.4 were used to estimate atomic coordinateChapter 5. Aliphatic Replacements of Phe82 in Cytochrome c 132Table 5.23: Refinement results and stereochemistry for the F821, F82L and F82M yeast iso-icytochrome c mutant structuresF821 F82L F82M1. Refinement resultsResolution range (A) 6.0—1.9 6.0—2.5 6.0—2.1Number of observed reflections 5991 4482 3520Completeness in resolution range (%) 78.3 59.5 61.3Number of protein atoms 891 891 891Number of solvent atoms 75 52 76Average thermal factors (A2)Protein atoms 19.6 20.9 19.7Solvent atoms 31.3 32.6 17.6R-factor 0.188 0.197 0.1922. Stereochemistry of final modelsr.m.s. deviation fromideal valuesDistances (A)Bond (1-2) 0.019 0.019 0.019Angle (1-3) 0.042 0.051 0.042Planar (1-4) 0.052 0.057 0.051Planes (A) 0.013 0.015 0.014Chiral volumes (A3) 0.143 0.202 0.164Non-bonded contacts (A)$Single torsion 0.215 0.229 0.224Multiple torsion 0.179 0.214 0.211Possible hydrogen bonds 0.187 0.231 0.227Torsion angles (°)Planar (0° or 180°) 2.3 2.5 2.1Staggered (±60°,180°) 22.5 28.4 27.6Orthonormal (±90°) 21.2 22.9 27.1t The r.m.s. deviations from ideality for this class of restraint incorporates a reduction of 0.2 Afrom the radius of each atom involved in a contact.Chapter 5. Aliphatic Replacements of Phe82 in Cytochrome c 133errors for the refined structures obtained. The Luzzati (1952) plot shown in Figure 5.30 givesestimates for overall coordinate errors for the F821, F82L and F82M mutant proteins of 0.20 A,0.26 A and 0.22 A, respectively. Evaluation of overall atomic coordinate errors by the methodof Cruickshank (1949) results in r.m.s. coordinate errors of 0.13 A for the F821 structure, 0.18 Afor the F82L structure, and 0.21 A for the F82M structure.5.2.4 Direct electrochemistryAs part of the work described in this thesis, the midpoint reduction potential of the F82Mmutant of yeast iso-i cytochrome c was obtained by cyclic voltammetry using the methods ofRafferty et al. (1990). However, a detailed discussion of the theory and methodology used inthe measurement of this value is beyond the scope of this work and the reader is referred toexcellent descriptions of this technique in Rafferty et al. (1990) and Burrows et al. (1991).The experimental apparatus used in this experiment was described by Rafferty et al. (1990)and the conditions employed are briefly described here. Prior to the measurement of theelectrochemical potential of the F82M mutant protein, a gold disk electrode was carefullypolished. This electrode was then cleaned by immersion in a solution of 100 mM NaC1O4having20 mM phosphate buffer and adjusted to pH 6, while cycling a current over a potential rangeof -1 to 1.2 V through the electrode. The gold surface of the electrode was then modified byimmersion in a saturated solution of 4,4’-dithiodipyridine. This treatment promotes the transferof electrons between the electrode and cytochrome c. A sample of F82M iso-i cytochrome c (at0.4 mM concentration) in a buffer with ionic strength = 0.1 M and at pH 6.0 (containing50 mM KC1 and enough sodium phosphate to provide the remaining ionic strength) was usedfor the determination of electrochemical potential. A voltammogram was obtained at 25°C byscanning a potential range of 44 to 544 mV at a rate of 20 mV s. The midpoint reductionpotential of the protein was calculated by averaging the peaks from the oxidizing and reducingcurrents.Chapter 5. Aliphatic Replacements of Phe82 in Cytochrome c 134Table 5.24: Overall average positional deviations (A) between wild-type yeast iso-i cytochrome c and the F821, F82L and F82M mutant proteins.Iso-i cytochrome c mutantAtom groups F821 F82L F82MAll common protein atoms 0.23 0.77 0.63All main chain atoms 0.15 0.54 0.43All common side chain atoms 0.33 1.06 0.87All heme atoms 0.10 0.32 0.43The three N-terminal residues (Thr(-5) - Phe(-3) ) were excluded fromthese calculations.5.3 Results5.3.1 Structural comparison of mutant and wild-type cytochromes cPrior to a detailed analysis of the F821, F82L and F82M mutant protein structures, all threewere superimposed onto the wild-type yeast iso-i cytoclirome c structure using a least-squaresfitting procedure that included all a-carbon atoms in the polypeptide chain with the exception ofthe three at the amino-terminal end. As discussed later, these three residues were found to havelarge positional deviations between these structures. Although the global fold of cytoclirome cis retained in the mutant proteins (Figure 5.33), it is apparent from Table 5.24 that the mainchain atoms of the F82L and F82M mutant proteins have undergone significant conformationalchanges. Plots of the average positional deviations of residues along the polypeptide chain of themutant proteins are shown in Figure 5.34. For the F82L and F82M mutant proteins, particularlylarge positional deviations are observed for the N-terminal residues, Thr(-5) through Lys(-2),Chapter 5. Aliphatic Replacements of Phe82 in Cytochrome c 135and a number of flexible surface side chains.5.3.2 Structure of F821 cytochrome cThe longer branch of the substituted 11e82 side chain is directed toward the interior of theprotein (Xi 66°; X2 = 136°) and resides near the bottom of the pocket normally occupiedby the aromatic ring of Phe82. The a-carbon atom of 11e82 is found in a more surface exposedlocation relative to that of Phe82 and this appears to result in shifts in the flexible Gly83-G1y84portion of the polypeptide chain (Ad 0.32 A; Figures 5.34a and 5.35). The shorter branchof the 11e82 side chain is projected toward the heme group and makes van der Waals’ contactwith the CMC methyl group of the heme (d = 3.2 A ). The conformation and geometry of theheme group in the F821 mutant protein is tabulated in Tables 5.24 and 5.25. As indicated inTable 5.26, there is a small increase in the heme solvent exposure in this mutant protein.5.3.3 Structure of F82L cytochrome cAs shown in Figures 5.34b and 5.36, substitution of leucine for Phe82 results in significantshifts in nearby polypeptide chain backbone (Ad = 0.70 A for residues 81—84). Other largemain chain shifts occur around G1y23 (Ad = 1.3 A for residues 22—23) and Lys54 (Ad =1.3 A for residues 53—54), as well as at the amino (Ad = 5.7 A for residues (-5)—(-4) ) andcarboxy (Ad = 1.0 A for residues 102—103) termini. At least some of the shifts in theselatter four regions are likely due to alternative crystaffine lattice packing interactions formedby the F82L mutant protein as compared to those found in crystals of wild-type yeast iso-icytochrome c. This is supported by a crystal packing analysis (Figure 5.37) which shows thatthree of these polypeptide chain segments do form different lattice interactions while the fourth,the C-terminus of the polypeptide chain, forms lattice interactions in wild-type protein crystalsbut not in those of the F82L mutant. This packing analysis also shows that large solventchannels are present between the protein molecules in crystals of the F82L mutant protein.It is not clear what has led to the new space group observed for crystals of the F82L mutantChapter 5. Aliphatic Replacements of Phe82 in Cytochrome c 136ABFigure 5.33: Stereo diagrams of the overlapped os-carbon backbones of the wild-type, F821,F82L and F82M iso-i cytochrome c structures in (a) the standard view looking at the hemeedge-on and (b) an alternate view looking directly down into the mutation site at residue 82.Also drawn for all four proteins are the side chain of residue 82, the heme group, the ligands tothe heme iron atom (Hisl8 and Met8O) and cysteines 14 and 17, which form covalent thioetherbonds to the heme porphyrin ring. Every fifth amino acid residue is indicated by its one-letteramino acid designation and sequence number. A complete listing of the primary sequence ofyeast iso-i cytochrome c can be found in Table 1.1.Chapter 5. Aliphatic Replacements of Phe82 in Cytochrome c 13740 •II•IIItI I I 1.1,1.1.1.1 I II I •tiiiiii I3.53.0-1 10 20 30 40 50 60 70 80 90 100Residue NumberB .1 I I I I I-1 10 20 30 40 50 60 70 80 90 100Residue NumberFigure 5.34: continued on next page.Chapter 5. Aliphatic Replacements of Phe82 in Cytochrome c 138CFigure 5.34: continued. Plots of average positional deviations from the wild-type iso-i cytochrome c structure for the (a) F821, (b) F82L and (c) F82M mutant proteins. Thick linesindicate average deviations of main chain atoms while thin lines indicate average deviations ofthe equivalent side chain atoms. In each diagram the filled circle at position 104 representsthe average positional deviation of the heme group. For each structure, the average positionaldeviation for all main chain atoms (except those of the three N-terminal residues) is displayedas a horizontal dashed line.-1 10 20 30 40 50 60 70 80 90 100Residue NumberChapter 5. Aliphatic Replacements of Phe82 in Cytochrome c 139ABFigure 5.35: Stereo diagrams showing two views of the region about 11e82 in the F821 mutantprotein. The structure of the wild-type protein (thick lines) has been superimposed on themutant protein structure (thin lines).aP7113P71Chapter .5. Aliphatic Replacements of Phe82 in Cytochrome c 140Table 5.25: Heme geometry of F821, F82L, F82M and wild-type yeast iso-i cytochromes cWild-type F821 F82L F82M1. Angular deviations (°) between the pyrrole nitrogen plane normaland the four individual pyrrole ring plane normals and the hemecoordinate bonds.A 9.4 11.3 12.9 9.4B 11.1 10.4 13.6 6.2C 8.8 10.3 11.7 12.3D 8.1 9.6 17.9 7.8Fe - Hisi8 NE2 2.2 5.5 4.9 6.8Fe - Met8O SD 4.9 1.7 12.2 10.02. Angular deviations (°) between the porphyrin ring plane normaland the four pyrrole ring plane normals, the pyrrole nitrogenplane normal and the heme coordinate bonds.A 6.7 7.2 8.2 7.1B 11.9 11.8 13.1 10.4C 9.8 10.1 15.9 8.7D 6.0 5.8 13.7 4.1NNNN 2.6 4.3 4.8 4.2Fe - Hisl8 NE2 3.2 1.7 4.7 8.4Fe - Met8O SD 7.5 5.5 16.9 8.23. Bond distances (A) between the heme iron atom and itssix ligands.Hisl8 NE2 1.98 2.07 2.10 1.92Met8O SD 2.36 2.21 2.03 2.20Heme NA 1.97 1.98 2.00 2.05Heme NB 2.00 2.02 2.02 1.99Heme NC 1.99 2.00 2.03 1.98Heme ND 2.01 2.04 2.01 2.02The pyrrole nitrogen plane is defined by the four pyrrole nitrogens of the heme group. The fourpyrrole ring planes are each defined by the five atoms of the ring and the first carbon atomattached to each of the four carbons of the ring. The porphyrin ring is defined by the five atomsin each of the four pyrrole rings, the four bridging methine carbon atoms, the first carbon atomof each of the eight side chains of the heme and the central iron atom of the heme. The hemeatom nomenclature used in this table follows the conventions of the Protein Data Bank (seeFigure 1.2).Chapter 5. Aliphatic Replacements of Phe82 in Cytochrome c 141Table 5.26: Heme solvent accessibility in F821, F82L, F82M and wild-type yeast iso-i cytochromes cYeast iso-i cytochrome c structureWild-type F821 F82L F82M1. Solvent accessible heme atoms andsurface area exposed (A2)CBB 0.0 7.4 0.0 0.0CHD 2.9 2.6 0.0 4.4CMC 9.2 7.8 10.3 11.5CAC 3.4 2.8 4.7 1.4CBC 20.1 20.0 18.1 25.4CMD 10.8 10.8 12.6 12.52. Total heme exposure (A2) 46.4 51.4 45.7 55.23. Total heme surface (A2) 513.1 515.8 514.3 513.54. % heme surface area exposed 9.0 10.0 8.9 10.7Solvent exposure was determined by the method of Connolly (1983) with a probe sphere havinga 1.4 A radius.Chapter 5. Aliphatic Replacements of Phe82 in Cytochrome c 142ABFigure 5.36: Stereo diagrams showing two views of the region about Leu82 in the F82L mutantprotein. The structure of the wild-type protein (thick lines) has been superimposed on themutant protein structure (thin lines).13aP71 P71P71Chapter 5. Aliphatic Replacements of Phe82 in Cytochrome c 143ABFigure 5.37: A packing diagram of the crystalline lattice containing the F82L iso-i cytochrome cstructure showing (a) the eight symmetry related molecules which pack within a single unit celland (b) one such molecule (thick lines) and the three neighboring cytochrome c molecules (#2,#3 #4) with which it forms direct protein-protein contacts. These latter three moleculesare related to the origin molecule by the following symmetry operations: #2:(1/2 — x, y, —z);#3:(x, —y, 1/2 — z); #4(—x, 1/2—y, z).Chapter 5. Aliphatic Replacements of Phe82 in Cytochrome c 144protein (Table 5.22). While there are significant conformational shifts in the vicinity of themutation site, in light of other studies of related mutant proteins (for example see those havingF82Y: Figure 3.19, Chapter 3; F82M: Figure 5.40; F82G: Louie & Brayer, 1989), these seeminsufficient to promote alternative lattice packing. Other major polypeptide changes are inflexible exterior ioop regions and as such, it is not surprising that these changes accompany anew lattice packing arrangement. It is notable that earlier crystaffization screens of wild-typeyeast iso-i cytochrome c did result in crystals of the same morphology as those of the F82Lmutant protein, although these were much too small for use in diffraction analyses. Apparentlysome feature of the F82L mutant protein enhances the ability of such crystals to attain a greatersize.A novel feature of the F82L structure is the unique orientation found for the N-terminalportion of its polypeptide chain (Figure 5.34b). This new conformation is accompanied by asignificant decrease in average main chain thermal factors for the three N-terminal residues,Thr(-5) to Phe(-3) (47.4 A2 in wild-type; 31.4 A2 in F82L; Figure 5.38), allowing for muchbetter definition of this region in electron density maps. This is likely due to the alternativelattice contacts formed by this region of the polypeptide chain in the F82L mutant proteincrystals, as noted above. As shown in Figure 5.38, other regions of polypeptide have significantchanges in main chain thermal factors, including increases about Leu9, Pro25, A1a43 and Lys79and a decrease around Asp6O. These changes in polypeptide chain mobility may be related toconformational changes observed in nearby segments of polypeptide chain (Figure 5.34b), orchanges in the lattice packing contacts of these regions.To demonstrate the good fit achieved, the final refined conformation of Leu82 overlaid on itscorresponding omit difference electron density map is shown in Figure 5.39. It was anticipatedthat substitution of a smaller leucine side chain at residue 82 would lead to increased hemesolvent exposure. Unexpectedly, this does not occur due to a shift in the main chain atomsof Ala.8i and Leu82 toward the interior of the heme pocket (Table 5.26; Figure 5.36). As canbe seen in Figure 5.36, the side chain of Leu82 is also found positioned much deeper in theChapter 5. Aliphatic Replacements of Phe82 in Cytochrome c 145/125-! 7- ....- S 56 26 26 26 50 60 76 55 56Residue NumberLegendI 15.Oandgreater10.Oto 15.05.Oto 10.0..5 S 56 26 36 26 50 66 76 26 IIResidue NumberFigure 5.38: Thermal factor difference matrices from the comparison of wild-type iso-i cytochrome c with the (a) F821, (b) F82L and (c) F82M mutant proteins. Each matrix point P,represents a pairing of amino acids x and y. The value of the pairing is calculated from theequation: = (B — By)mutant(Bx — By)wild_type where B is the average thermal factor(A2) of the main chain atoms of a given amino acid i. Positive matrix values are displayedaccording to the scale shown. Due to the inverse symmetry of the matrix across the diagonal,negative values are redundant and have been omitted. Amino acids producing vertical streakswithin the matrix have significantly higher main chain thermal factors in the mutant structurerelative to the wild-type structure while horizontal streaks indicate significantly lower mainchain thermal factors in the mutant protein.V.Chapter 5. Aliphatic Replacements of Phe82 in Cytochrome c 146LEU\Figure 5.39: Omit difference electron density map of the F82L iso-i cytochrome c structure inthe vicinity of Leu82 computed with the omission of the whole of this residue. Contours areshown at the 3cr level. Superimposed on this map is the final refined structure of the F82Lmutant protein.Chapter 5. Aliphatic Replacements of Phe82 in Cytochrome c 147hydrophobic pocket than expected. The main chain shift of residues 81 and 82 coupled withthe formation of a close contact between the Leu82 CD2 side chain methyl group and theCE methyl group of Met8O (d = 3.1 A; Figure 5.36), appear to play a role in perturbing theconformation of the side chain of Met8O. Primarily affected are the length and orientation ofthe heme iron ligand bond formed by Met8O (Table 5.25).5.3.4 Structure of F82M cytochrome cPlacement of a methionine at residue 82 is accompanied by positional shifts in nearby polypeptide chain, including residues 8 1—84 (Figure 5.40). A further adjustment in the side chain ofLeu85 is observed, wherein it shifts toward the Met82 side chain (Ld 1.1 A; Figure 5.40)and fills much of the free volume created by the substitution of a smaller methionine for thenormally resident phenylalanine. As discussed in Chapter 4 (Section 4.4.4), Leu85 also appearsto have a certain amount of flexibility to respond to nearby mutations by virtue of belonging to a cluster of conserved leucines, including those at positions 9, 68, 94 and 98, which isable to accommodate structural changes in adjacent protein groups by undergoing positionaldisplacements in concert with changes in the size of an adjacent internal hydrophobic cavity.A methionine at position 82 also results in a conformational shift in the side chain of theMet8O heme ligand. As can be seen in Figure 5.40, the CE methyl group of Met82 points into thehydrophobic heme pocket where it forms a close contact with the CE methyl group of Met8O (d= 3.3 A) and displaces this latter group by 0.9 A from its normal position. This displacementoccurs primarily via a rotation of -41° in. the X3 torsion angle of Met8O. While movementof Met8O CE avoids a potential steric conflict, it does result in an increase in the thermalparameters of the polypeptide backbone atoms of Met8O and the adjacent Lys79 (Figure 5.38).A contributing factor to these higher thermal factors may be the loss of a hydrogen bondbetween the side chain of Lys79 and the main chain carbonyl oxygen of Ser47 (d = 2.6 A inwild-type; 4.0 A in FS2M mutant). This hydrogen bond is formed across the solvent exposededge of the heme crevice and is believed to play a role in stabilizing the structure of cytochrome cChapter 5. Aliphatic Replacements of Phe82 in Cytochrome c 148ABFigure 5.40: Stereo diagrams showing two views of the region about Met82 in the F82M mutantprotein. The structure of the wild-type protein (thick lines) has been superimposed on themutant protein structure (thin lines).F82MaP71 P71P71Chapter 5. Aliphatic Replacements of Phe82 in Cytochrome c 149(Hampsey et at., 1986).One apparent consequence of the shift observed in the side chain of Met8O is a ‘5° rotation ofthe heme moiety as a whole about an axis normal to the heme plane and running approximatelythrough the heme NA atom. Heme rotation leads to a correspondingly large displacement ofheme atoms in the F82M mutant protein compared to their positions in wild-type yeast iso-icytochrome c (zd = 0.43 A; Table 5.24), but does not greatly perturb the internal geometry ofthe heme (Table 5.25). Two heme substituents particularly perturbed include the CBC methylgroup (d = 0.5 A ) which becomes more exposed to solvent (Table 5.26) and the CBB methylgroup (td = 1.1 A ) which comes into contact with the CD2 methyl group of the side chainof Leu94 (d = 3.2 A). This latter interaction may account for the positional shift observed forthe side chain of Leu98 (d = 1.4 A ) and the subsequent abolition of the normally presentinternal hydrophobic cavity in the back of the heme pocket. The potential roles of Leu94 andthis internal cavity are discussed in more detail in Chapter 4.Another prominent structural change observed in the F82M mutant protein, but which isfar removed from the mutation site at residue 82, involves the conformationaily flexible regionaround residues 37 and 38, which forms part of a type 11/3-turn on the surface of yeast iso-icytochrome c (Table 1.2; Louie & Brayer, 1990). Here G1y37 has adopted a new conformation(4 = 168°; & = -77°) which differs markedly from that found in the wild-type protein (qS =87°; b = -11°). Although conformational variability about G1y37 has previously been notedin different cytochronies c, this has been attributed to local amino acid substitutions (Murphyet at., 1992). The current study suggests that while this may be a contributing factor, regionalflexibility of the polypeptide chain is sufficient to result in different structural isoforms. In theF82M mutant protein it seems likely that heme group rotation is responsible for inducing thisrearrangement.In addition to conformational shifts in the main chain atoms of Arg38, the position of theside chain of this residue also undergoes a significant alteration in the F82M mutant protein(d = 1.1 A ). Normally, the guanidinium group of this side chain participates in hydrogenChapter 5. Aliphatic Replacements of Phe82 in Cytochrome c 150bonding interactions with two conserved water molecules, Wat 121 and Wat 168, which are inturn hydrogen bonded to the propionate A group of the heme (Figure 1.5; Louie & Brayer,1990). Heme rotation in the F82M mutant protein shifts the position of both the propionate Agroup (zd = 0.55 A) and Watl2l (tid = 0.73 A). The conformational shifts observed for Arg38compensate for heme movement and preserve the hydrogen bonding interactions observed inthe wild-type protein. However, as evident from Figure 5.38, these shifts do result in increasedthermal mobility for residues 39—41.5.4 Discussion5.4.1 Structural implicationsReplacement of Phe82 by the aliphatic residues isoleucine, leucine and methionine substantiallyalters the character of the side chain present at this position both in terms of bulk and overallshape. Surprisingly, substitution by isoleucine, which has the fewest degrees of side chainconformational freedom among these three residues, results in the smallest structural changes.The positional shifts observed (Figures 5.34a and 5.35) appear to be a consequence of the betabranching of the 11e82 side chain and the difficulty in packing this group into the shape ofthe available space. Particularly affected by this are G1y83 and G1y84 which provide someconformational flexibility in this region to compensate for this new group. The orientation ofthe side chain of 11e82 appears to be for the most part set by the close contact of the CG2methyl group of this residue with the plane of the heme group. Some increased heme solventexposure (Table 5.26) is observed but the conformation of 11e82 is such as to avoid the formationof a solvent channel into the hydrophobic heme pocket.As illustrated in Figure 5.34b, substantive conformational shifts are found in the F82Lmutant protein. These are evident both in the direct vicinity of the mutation site (Figure 5.36)as well as more globally about the outer surface of the protein. These latter changes are relatedto the new lattice packing arrangement adopted by this protein in the crystalline state, withthe regions about G1y23, Lys54 and the N and C-termini being the most affected. All of theseChapter 5. Aliphatic Replacements of Phe82 in Cytochrome c 151regions form prominent lattice contacts in one or both of the F82L mutant and wild-type yeastiso-i cytochromes c. Directly in the mutation site, significant shifts in the adjacent backboneof residues 81 to 84 are observed and these serve to partially fill open space present as the resultof the smaller leucine side chain. These adjustments coupled with the positioning of this sidechain further than expected into the hydrophobic heme pocket (Figure 5.36) mean that for thismutant protein no increase in heme solvent exposure is observed (Table 5.26). One result ofthe tight packing of Leu82 in the heme pocket is a close contact with the heme ligand residueMet8O, which perturbs the nature of the heme ligand bond formed by this residue (Table 5.25).A completely unexpected consequence of methionine substitution at Phe82 is the novelorientation observed for the heme group (Figure 5.40), which is found rotated by .-‘5° withinthe hydrophobic pocket in which it is bound. Heme rotation appears to be promoted by a closecontact between the side chains of Met82 and the heme ligand Met8O, and as a consequence theCD methyl group of the side chain of this latter residue is considerably displaced. The effectsof heme rotation are widespread and include conformational changes in a cluster of conservedleucines, rearrangement of a surface /3-turn formed by residues 35—38, positional shifts in theinternal Wati2i and the side chain of Arg38, and an increase in the thermal mobility of thepolypeptide segments involving residues 39—41 and 78—81 (Figure 5.38).In the F82M mutant protein at the mutation site itself, the nearby backbone of residues 81to 84, along with the side chain of Leu85, shift in toward the heme pocket and thereby fill in thefree volume that would otherwise be created by the smaller Met82 side chain. Nevertheless asmall increase in heme solvent exposure is observed for this protein (Table 5.26), which appearsto be related to the observed rotation of the heme group. The heme atoms most affected (CMC,CBC and CMD) occur on the solvent exposed edge of this group (Figures 1.2 and 5.40), a regionrotated up and further out into solvent. It is notable that in no other mutant form of yeastiso-i cytochrome c has such a substantive and concerted movement of the heme group beenobserved. Much of the ability of the heme group to undergo this rotation is likely the resultof the presence of an internal cavity and a nearby cluster of flexible leucine side chains at theChapter 5. Aliphatic Replacements of Phe82 in Cytochrome c 152back of the heme pocket (see Chapter 4 for further discussion in this regard).5.4.2 Functional effectsAs discussed in Chapter 3, Phe82 occupies a central position in what is believed to be thecontact face formed between cytochrome c and associated electron transfer partners. Changesin this interactive surface by mutation of P1ie82 to other amino acids lead to alterations in thekinetics of electron transfer within such complexes (Wiffie et al., 1993; Guillemette et al., 1994).Especially notable is the behavior of the F821 mutant protein, which binds 50 x tighter to cytochrome c oxidase than does the wild-type protein (Michel et al., 1989). This might be relatedto the unique shape of the aliphatic side chain of 11e82 which permits a tighter and more extensive fit between cytochrome c and cytochrome c oxidase when these proteins are complexed.Figure 5.35 suggests this might result from the smaller size of the side chain of isoleucine andperhaps the positioning of this residue as dictated by its /3-branched structure. Unfortunately,the structure of the complex formed between cytochrome c and cytochrome c oxidase has notbeen determined and therefore a more definitive understanding of the interactions occurringis not yet possible. It should be remembered that while 11e82 may lead to tighter binding inthis complex, the invariant conservation of a phenylalanine at this position suggests that otherfactors are of greater importance in the overall scheme of the electron transfer pathway of whichcytochrome c is only part.Of all the mutant proteins studied, the F821 mutant protein alone has such a marked effecton cytochrome c coinpiexation with cytochrome c oxidase, with other substitutions at Phe82leading to less drastic changes in the interaction of these two electron transfer proteins. Forexample, the positional shifts in the polypeptide backbone observed in the F82L (Figure 5.36)and F82G (Louie & Brayer, 1989) mutant proteins lead to only slightly increased affinitiesfor cytochrome c oxidase (Michel et al., 1989). Furthermore, while the replacement of Phe82by serine in the F82S mutant protein might be expected to allow for a closer approach ofcytochrome c and cytochrome c oxidase, the polar nature of the serine side chain and theChapter 5. Aliphatic Replacements of Phe82 in Cytochrome c 153introduction of a newly bound water molecule into a solvent channel in the heme pocket (Louieet al., 1988b) appear to have a disruptive effect on complex formation. In the case of the F82Mmutant protein, it is conceivable that a critical factor perturbing electron transfer proteincomplexation is the unexpected discovery that the heme group of this protein is rotated. Anatural consequence of this will be the formation of altered interactions with the exposedheme edge in any complexes formed with this protein. Collectively these results show that thecharacter and surface contour of cytochrome c in the region of residue 82 is a determinant inthe ability of this protein to form productive electron transfer complexes, but that elucidationof the structures of a number of these wifi be required to understand all the factors involved.The alterations in the association of cytochrome c with redox partners observed to resultfrom amino acid substitutions at residue 82 account for the perturbation of the kinetics of electron transfer within the protein complexes formed (Everest et al., 1991; Hazzard et at., 1992).However, it should be pointed out that the relationship between complex formation and electrontransfer kinetics is not straightforward since multiple docking geometries having different reactivities are possible for these complexes (Northrup et at., 1993; Stemp & Hoffman, 1993; Zhou& Hoffman, 1993; Mauk et at., 1994; Zhou & Hoffman, 1994). Further complicating matters isthe dependence of electron transfer kinetics on environmental factors such as temperature andsolvent (Nocek et at., 1991). Thus it is difficult to obtain a quantitative correlation between thestructural changes and electron transfer kinetics observed for the mutant proteins examined.5.4.3 Effects on reduction potentialThe reduction potential of cytochrome c is dependent on the dielectric constant of the hydrophobic heme pocket (Kassner, 1973; Louie & Brayer, 1989). However, as discussed in Chapters 3and 4, when substitutions are made in or near the heme pocket, the effect on reduction potentialcan vary considerably and is dependent on the positioning of newly introduced groups relativeto the heme moiety. For example, when water molecules are introduced into the heme pocketalong the edge of the heme where intervening groups such as heme methyl substituents shieldChapter 5. Aliphatic Replacements of Phe82 in Cytochrome c 154the heme plane, the effect on heme reduction potential is small, even if these polar groups areclose (5.4 A) to the heme porphyrin ring (F82Y/L85A mutant protein; Chapter 3). In contrast,replacements which increase the polarity of the heme environment through the introduction ofpolar groups against the planar face of the heme have large effects on reduction potential. Twoexamples of this are found in the F82S mutant protein, in which a decrease in reduction potential results from the introduction of a polar hydroxyl group and a solvent channel into the hemepocket (Louie et at., 1988b; see also Chapter 6), and the F82G mutant protein, in which thepacking of polar main chain groups against the face of the heme porphyrin ring also results ina decrease in reduction potential (Louie & Brayer, 1989). Thus the parameters which influenceheme reduction potential are complex and numerous, making predictions of final values basedon modeling difficult.The small decrease observed for the reduction potential of the F821 mutant cytochrome c(Em = -17 mV; Table 5.27) can be rationalized by its slightly increased heme solvent exposure(Table 5.26). This increase in solvent exposure is for the most part localized at the CBB methylgroup which is normally completely buried in the hydrophobic heme pocket (Table 5.26). Theobserved decrease in reduction potential is not as large as that for the F82S mutant protein(iEm = -45 mV; Rafferty, 1990) in which more of the CBB methyl group is exposed to solvent.In this latter case, the polar nature of the serine side chain and the binding of a new watermolecule into the mutation site also contribute to the lower reduction potential observed.In the case of the F82L mutant protein, very little change is observed in the midpointreduction potential relative to that of the wild-type protein (Table 5.27). This result correspondsto the fact that residues 81—84 shift deeper into the heme pocket, preventing the formation ofa solvent channel that would be expected from the simple replacement of Phe82 by leucine.Thus both the nonpolar heme environment (Table 5.26) and the midpoint reduction potentialare preserved in the F82L mutant protein. Note that perturbations in the conformation of theMet8O ligand in this mutant protein do not appear to have any impact on reduction potential.More difficult to explain are the results for the F82M mutant protein. Here an increase inChapter .5. Aliphatic Replacements of Phe82 in Cytochrome c 155Table 5.27: Reduction potentials for F821, F82L, F82M and wild-type yeast iso-i cytochromes cCytochrome c Em (mV)Wild-typef 290±2F821t 273±2F82Lt 286±2F82M 288±2Experimental conditions were: 25°C, pH 6.0 and = 0.1 M. Values are listedrelative to a standard hydrogen electrode reference.f from Rafferty et al. (1990).The protocol used to determine the reduction potential for the F82M mutantprotein is described in Section 5.2.4.heme solvent exposure is observed (Table 5.26) without a corresponding decrease in reductionpotential (Table 5.27). A possible explanation for this discrepancy may be the fact that theincreased heme solvent exposure originates with the rotation of the heme moiety within itsbinding pocket. Since the resultant increased heme solvent exposure occurs to one edge of theheme, it is possible its effect on heme reduction potential is muted due to shielding by hemesubstituent side chains as was previously observed in the F82Y/L85A (Chapter 3) and L94S(Chapter 4) mutant proteins.Understanding the elements involved in regulation of reduction potential in the F82M mutant protein is further complicated by other structural features. Some of these involve shiftsthat occur in the immediate vicinity of the heme, including altered interactions between thisgroup and its binding pocket, and a new conformation for the side chain of the heme ligandChapter 5. Aliphatic Replacements of Phe82 in Cytochrome c 156Met8O. Others involve more distant groups, which cannot be discounted as unimportant giventhat similar shifts in the packing of polypeptide groups distant from the heme have been shownto have a significant effect on heme reduction potential in other instances (Murphy et al., 1993).In conclusion, the results for the F82M mutant protein show that heme solvent exposure byitself cannot be used as a simple predictor of the reduction potential of cytochrome c sincemany other factors contribute in different ways to the regulation of this functional property.Chapter 6The Structural and Functional Effects of Multiple Mutations at Distal Sites inCytochrome c6.1 IntroductionWhile an extensive analysis of the individual roles of Phe82 and Leu85 in the structure andfunction of cytochrome c has been made in Chapters 3—5, it must be remembered that theseresidues do not act in isolation and that other amino acids, both nearby and distally remote,influence many of the same properties. For example, two other regions of cytochrome c thatmake significant contributions to heme reduction potential, electron transfer rates and proteinstability are those in the vicinity of Arg38 and Asn52. Both of these highly conserved aminoacids have been studied individually with respect to the structural and functional consequencesof making amino acid substitutions. Figure 1.9 illustrates the placement of Arg38, Asn52 andPhe82 within the structure of yeast iso-i cytochrome c and shows that each is spatially distant(> 9 A) from the others. The goal of the present chapter is to study simultaneous mutationsat all three of these residues to gain insight into the interplay between these distal sites withrespect to functional properties that they all influence in common.Arginine 38 is an invariant residue (Hampsey et aL, 1988; Moore & Pettigrew, 1990) whichis partially buried in the interior of cytochrome c. The side chain of this residue is involved ina hydrogen bonding network with propionate A of the heme group and two conserved internalwater molecules, as illustrated in Figure 1.5. This residue is involved in the regulation of theheme reduction potential of cytochrome c through both the electrostatic stabilization providedby the charge on its guanidinium group and the electron withdrawing effect of this side chainon herne propionate A (Cutler et al., 1989; Davies et al., 1993). Previous studies have shown157Chapter 6. Multiple Distal Mutations in Cytochrome c 158that various amino acid substitutions at Arg38 result in a decrease in the midpoint reductionpotential of cytochrome c (Cutler et at., 1989). In structural terms, the substitution of Arg38by alanine results in the removal of a positively charged side chain and creates an access pointfor bulk solvent (H. Tong, personal communication).Asparagine 52 is a completely internal residue which is invariant among eukaryotic cytochrome c sequences (Moore & Pettigrew, 1990). This residue participates in a hydrogen bondingnetwork that involves the conserved internal water molecule, Wat 166 (Figure 1.5). This waterappears to be an important component in the structural transition between oxidation statesin cytochrome c (Berghuis & Brayer, 1992). Much interest has focused on the substitution ofAsn52 by isoleucine since this has been found to increase protein stability dramatically (Hickeyet at., 1991) and act as a suppressor of mutations which would normally abolish functionalactivity in cytochrome c (Das et at., 1989; Berroteran & Hampsey, 1991). The N521 substitution also results in a significant decrease in heme reduction potential (Burrows et at., 1991;Guillemette et at., 1994), likely as the result of the displacement of Wat166 and the resultantadjustment of internal hydrogen bonding (Berghuis et at., 1994a).The focus of the present studies is the analysis of mutant proteins having all possible combinations of the substitutions R38A, N521 and F82S. In the case of Phe82, the serine substitutionwas chosen because it has the largest impact on functional properties without leading to changesin polypeptide chain folding, a feature that might complicate interpretation of the results. TheR38A and N521 mutations were chosen since individually each of these results in well characterized perturbations of the functional behavior of cytochrome c. In addition, the structuresof the mutant proteins having each of the targeted substitutions at Arg38, Asn52 and Phe82have been determined and therefore a baseline exists for documenting any additional effectsthat might occur in multiply mutated proteins. The basic question posed by these studies iswhether Arg38, Asn52 and Phe82 act individually or in unison with respect to various properties of cytochrome c, and if the latter is the case, to what degree does this occur. This workhas been published as part of Lo et al. (1995b).Chapter 6. Multiple Distal Mutations in Cytochrome c 159Table 6.28: Data collection parameters for yeast iso-i cytochromes c with multiple distal mutationsIso-i cytochrome c mutantR38A/N521/Parameter R38A/N521 R38A/F82S N.521/F82S F82SSpace group P4321 P4321 P4321 P4321Cell dimensions (A)a = b 36.42 36.75 36.10 36.48c 137.03 137.39 137.47 137.28Number of reflections collected 19351 20721 36008 71744Number of unique reflections 6667 6291 8048 8999Merging R-factort 0.067 0.088 0.088 0.064Resolution (A) 1.8 1.9 1.8 1.8TM T f + — Zhki >1lhkL41Ierging - acor —•/_hk L_=i ‘hkl6.2 Experimental ProceduresCrystals of reduced yeast iso-i cytochrome c proteins containing combinations of the R38A,N521 and F82S mutations were grown from solutions of 88% ammonium sulphate and 70 mMsodium dithionite buffered at pH 6.5 by 0.1 M sodium phosphate (Table 2.3). The hanging dropvapour diffusion method (Section 2.3) was employed with seeding from micro-crystals (Leunget al., 1989). The crystals grown for each of the four mutant proteins are of the space groupP4321with unit cell dimensions as indicated in Table 6.28 and are isomorphous with crystalsof wild-type yeast iso-i cytochrome c.For each of the four mutant cytochromes c, X-ray diffraction data were collected on a RigakuR-AXIS II imaging plate area detector from a single crystal as described in Section 2.5.3. TheChapter 6. Multiple Distal Mutations in Cytochrome c 160incident radiation was provided by a RU-300 rotating anode generator operating at 90—100 mAand 50—60 kV. For each frame, each crystal was oscillated through a angle of 1.0° and exposedto the X-ray beam for 20—30 minutes. The relatively low number of total measurements madefor the R38A/N521 and R38A/F82S crystals (Table 6.28) was a result of these diffracting poorlyto high resolution and having a much higher rate of intensity decay. As documented in Table 2.3,the best crystals that could be grown of these mutant proteins were less than ideal for analysis,with the R38A/N521 crystal having one very thin dimension and the R38A/F82S crystal beingextremely small overall. X-ray intensity data were processed to structure factors (summarizedin Table 6.28) using the R-AXIS II data processing software (Higashi, 1990; Sato et al., 1992)and the procedures described in Section 2.5.3. Each mutant protein data set was put on anabsolute scale using the Wilson (1942) plot method (Section 2.5.4).Starting models for the structural refinement of the four mutant proteins were based on thehigh resolution structure of wild-type iso-i cytochrome c (Loule & Brayer, 1990). New sidechains at the three mutation sites were initially modeled as alanine residues. Also includedin the starting models were all water molecules from the wild-type structure having isotropicthermal factors below 50 A2, with the exception of those in the vicinity of residues 38, 52 and82. Additionally, the sulphate anion bound to the amino terminal end of the N-terminal helixof cytochrome c was included in each starting model.The restrained parameter least-squares approach (Hendrickson, 1985) was employed for therefinement of each cytochrome c mutant structure as described in Section 2.6. The restraintweights listed in Table 2.4 were used for refinement and all water molecules were treated asfully occupied neutral oxygen atoms. Each of the four mutant proteins was initially subjectedto 24 cycles of least-squares refinement, after which the first round of manual adjustments weremade based on F0 — F and 2F0 — F difference electron density maps at the three mutationsites. In this way mutation of Arg38 to alanine was confirmed by the lack of electron densityfor a longer side chain. Also observed in this case were electron density peaks representing newwater molecules bound in the cavity created by this mutation. At this point the side chain ofChapter 6. Multiple Distal Mutations in Cytochrome c 16111e52 could be clearly visualized and fit. These difference maps also confirmed the concomitantelimination of the internal water molecule, Wat166, when 11e52 is present as opposed to thenormally resident Asn52. In those mutant proteins where Phe82 was replaced by serine, theconformation of the Ser82 hydroxyl group was readily modeled, as was an adjacent newly boundwater molecule.For each of the four mutant proteins, a further 50—60 cycles of least-squares refinement werecarried out. In addition, omit, Fc, — F and 2F0 — F difference electron density maps coveringthe entire course of the polypeptide chain were examined periodically during the course of thisleast-squares refinement, with each mutant protein being subjected to a thorough examinationof this type a total of four times. These checks of the progress of refinement resulted in anumber of manual corrections of the five amino-terminal residues of the polypeptide chain andin surface side chain positions. Water molecules considered for inclusion into refinement modelswere primarily found by the use of an iterative procedure involving alternating rounds of peaksearching and reciprocal space refinement (Tong et al., 1994). All water molecule positionsdetermined by this method were also confirmed manually by reference to F0 — F and 2F0 — Fdifference electron density maps. Final refinement parameters and stereochemistry for all fourmutant protein structures are tabulated in Table 6.29.Atomic coordinate errors for each of the four mutant protein structures have been estimatedusing the two methods described in Section 2.6.3.4. Inspection of a Luzzati (1952) plot (Figure 6.41) provides estimates ofr.m.s. coordinate errors ranging from 0.18 A for the R38A/N521structure to 0.22 A for the R38A/F82S structure. The N521/F82S and R38A/N521/F82Sstructures both have r.m.s. coordinate errors of 0.20 A by this method. Overall atomic coordinate errors can also be estimated by evaluating individual atomic errors (Cruickshank, 1949,1954). Based on this method, the estimated overall r.m.s. coordinate error is 0.13 A for theR38A/N521 structure, 0.14 A for the R38A/F82S structure, and 0.12 A for the N52I/F82S andR38A/N521/F82S structures.Chapter 6. Multiple Distal Mutations in Cytochrome c 162Table 6.29: Refinement results and stereochemistry for the structures of yeast iso-i cytochromes c with multiple distal mutationsR38A/ R38A/ N521/ R38A/N521/N521 F82S F82S F82S1. Refinement resultsResolution range (A) 6.0—1.8 6.0—1.9 6.0—1.8 6.0—1.8Number of observed reflections 6358 6018 7740 8665Completeness in resolution range (%) 70.8 77.1 87.4 95.3Number of protein atoms 886 881 887 881Number of solvent atoms 75 80 69 73Average thermal factors (A2)Protein atoms 22.5 21.0 21.2 23.9Solvent atoms 34.5 29.0 34.4 36.8R-factor 0.188 0.198 0.197 0.1992. Stereochemistry of final modelsr.m.s. deviation from ideal valuesDistances (A)Bond (1-2) 0.019 0.019 0.019 0.019Angle (1-3) 0.039 0.039 0.038 0.039Planar (1-4) 0.050 0.046 0.047 0.049Planes (A) 0.014 0.014 0.014 0.015Chiral volumes (A3) 0.143 0.148 0.132 0.155Non-bonded contacts (A)TSingle torsion 0.212 0.211 0.218 0.211Multiple torsion 0.198 0.194 0.186 0.193Possible hydrogen bonds 0.201 0.227 0.205 0.185Torsion angles (°)Planar (0° or 180°) 2.3 2.2 2.2 2.5Staggered (+60°,180°) 22.1 21.9 20.9 19.0Orthonormal (±90°) 21.0 20.6 21.3 25.1I The r.m.s. deviations from ideality for this class of restraint incorporates a reduction of 0.2 Afrom the radius of each atom involved in a contact.Chapter 6. Multiple Distal Mutations in Cytochrome c 163Resolution (A)___ __1.7E°5 I-0.0 U0.35 -0.30 - 0.24 AE0.10— • . .i I I II I I I I I • • • I I I I I I0.05 0.10 0.15 0.20 0.25 0.30(sin 0)! ?Figure 6.41: A plot of the crystallographic R-factor at the end of refinement as a function ofresolution for the R38A/N521 (Q), R38A/FS2S (7), N521/F82S (0) and R38A/N521/F82S(D) mutants of yeast iso-i cytochrome c. The theoretical dependence of R-factor on resolutionassuming various levels of r.m.s. error in the atomic positions of the model (Luzzati, 1952) isshown as broken lines. This analysis suggested an overall r.m.s. coordinate error for the mutantstructures of between 0.18 and 0.22 A. The top portion of this figure (axes at top and right)shows the fraction of reflections observed and used in refinement as a function of resolution.Chapter 6. Multiple Distal Mutations in Cytochrome c 164Table 6.30: Overall average positional deviations (A) between yeast iso-i cytochromes c withmultiple distal mutations and the wild-type proteinIso-i cytochrome c mutantR3SA/N521/Atom groups R38A/N521 R38A/F82S N521/F82S F82SAll common protein atoms 0.28 0.26 0.32 0.30All main chain atoms 0.19 0.17 0.22 0.20All common side chain atoms 0.38 0.36 0.44 0.42All heme atoms 0.18 0.19 0.21 0.216.3 Results6.3.1 Structural comparison of mutant and wild-type cytochromes cTo obtain an accurate assessment of the individual and cumulative structural effects of theR38A, N521 and F82S amino acid substitutions, the three-dimensional structures of the fourmutant proteins were not only compared with each other, but also with the high resolutionstructure of wild-type yeast iso-i cytochrome c (Louie & Brayer, 1990). Prior to these detailedcomparisons, the structure of each mutant protein was superimposed onto the polypeptidechain of the wild-type protein by a least-squares fit of all os-carbon atoms. These structuralanalyses showed that significant changes were observed at each of the R38A, N521 and F82Smutation sites, and these are discussed individually in subsequent sections of this chapter.The conformational shifts involved are for the most part localized to these mutation sites, andno major disruptions to the overall fold of the polypeptide chain were observed (Table 6.30;Figure 6.42).The distribution of average positional deviations over the course of the polypeptide chainChapter 6. Multiple Distal Mutations in Cytochrome c 165Figure 6.42: A stereo diagram of the of-carbon backbone of the wild-type iso-i cytochrome Cstructure and those of the four mutant proteins with multiple distal mutations. The hememoieties of all five proteins are shown, along with the side chains of the mutated residues,Arg38, Asn52 and Phe82, the ligands to the heme iron atom, Hisl8 and Met8O, and cysteines14 and 17, which form covalent thioether bonds to the heme porphyrin ring. Every fifth aminoacid residue is indicated by its one-letter amino acid designation and sequence number.Chapter 6. Multiple Distal Mutations in Cytochrome c 166is shown for each of the four mutant cytochromes c in Figure 6.43. Larger variations in thepositions of the N-terminal residues Thr(-5) through Phe(-3) arise as a consequence of thermaldisorder in this region of the protein (Louie & Brayer, 1990) rather than as a consequence ofthe introduced mutations. In addition, the hydrophilic side chains of Lys(-2), Lys4, Glu2i,G1u44, Lys54, Lys55, Asn63, G1u66, Lys86, Lys87, Lys89 and LyslOO are all on the surface ofthe protein and are projected out into the surrounding solvent medium. As such, these sidechains are substantially disordered and display large positional deviations between the variousstructures studied. Three hydrophobic residues on the protein surface also undergo apparentpositional shifts, but these changes arise from either the presence of multiple residue conformations (Leu9), the high mobility of the local polypeptide backbone (Va157), or a combinationof these factors (Leu58). These three residues display similar characteristics in wild-type yeastiso-i cytochrome c (Louie & Brayer, 1990).Examination of the heme geometry of the mutant proteins (Table 6.31) reveals that allfour proteins are comparable to the wild-type protein in this regard. Of note is the shorterthan expected ligand distance between the NE2 atom of Hisi8 and the heme iron atom in theN521/F82S protein. This apparently arises from movement of the side chain of Hisi8 towardthe heme group (Lid = 0.2 A ), and leads to a perturbation of the angle between this bond andthe pyrrole nitrogen plane (becomes instead of the average of -‘2.5° ; Table 6.31).6.3.2 R38A mutation siteThe replacement of Arg38 by an alanine represents a substantial decrease in side chain sizeas well as the elimination of a positive charge located partially in the interior of the protein inclose proximity to the heme propionate A group. The space vacated by the Arg38 side chainis filled by two water molecules (Wat A and B) which serve to maintain the structure of thehydrogen bonding network centered around heme propionate A (Table 6.32; Figure 6.44). In theR38A/N521/F82S mutant protein only Wat A is present. Spatially, these new water moleculesare located at positions comparable to the NE and Nh atoms of Arg38 and participate inChapter 6. Multiple Distal Mutations in Cytochrome C 16740• I I I I 1.1..,,,., I I I3.5_-. 3.02.52.0441 M1-1 10 20 30 40 50 60 70 80 90 100Residue Number13 4.o ,..I I I I I I I I I3.5,—. 3.02.5-1 10 20 30 40 50 60 70 80 90 100Residue NumberFigure 6.43: continued on next page.Chapter 6. Multiple Distal Mutations in Cytochrome c 1684.0 I I I I I I I I3.5,—. 3.02.52.0-1 10 20 30 40 50 60 70 80 90 100Residue Number4.0 1,1.1 I I I I I I I I-1 10 20 30 40 50 60 70 80 90 100Residue NumberFigure 6.43: continued. Plots of average positional deviations from wild-type iso-i cytochrome c along the course of the polypeptide chain for the (a) R38A/N521, (b) R38A/F82S, (c)N521/F82S and (d) R38A/N521/F82S mutant proteins. Thick lines indicate average deviationsof main chain atoms while thin lines indicate average deviations of the equivalent side chainatoms. In each diagram the filled circle at residue position 104 represents the average positionaldeviation of the heme group and the horizontal dashed line represents the average positionaldeviation for all main chain atoms (Table 6.30).Chapter 6. Multiple Distal Mutations in Cytochrome c 169Table 6.31: Heme geometry of yeast iso-i cytochromes c with multiple distal mutationsR38A/N521Wild-type R38A/N521 R38A/F82S N521/F82S F82S1. Angular deviations (°) between the pyrrole nitrogen plane normal and the fourindividual pyrrole ring plane normals and the heme coordinate bonds.A 9.4 12.7 16.0 11.3 9.5B 11.1 10.4 11.4 8.7 8.3C 8.8 10.6 12.3 11.0 8.2D 8.1 10.9 13.2 9.0 10.0Fe - Hisl8 NE2 2.2 2.1 4.2 7.2 1.5Fe - Met8O SD 4.9 4.4 1.4 3.6 3.22. Angular deviations (°) between the porphyrin ring plane normal and the fourpyrrole ring plane normals, the pyrrole nitrogen plane normal and the hemecoordinate bonds.A 6.7 9.1 11.9 6.6 5.4B 11.9 12.3 14.8 10.6 9.1C 9.8 10.9 11.4 10.3 9.2D 6.0 7.5 9.1 4.2 5.9NNNN 2.6 4.0 4.8 4.7 4.1Fe - Hisl8 NE2 3.2 3.0 4.5 2.8 2.5Fe - Met8O SD 7.5 7.4 5.6 7.3 7.23. Bond distances (A) between the heme iron atom and its six ligands.Hisi8 NE2 1.98 1.91 1.96 1.79 1.94Met8O SD 2.36 2.34 2.39 2.34 2.38Heme NA 1.97 1.98 1.97 1.99 2.00Heme NB 2.00 2.00 2.01 2.01 2.04Heme NC 1.99 2.00 2.03 1.98 2.03Heme ND 2.01 2.06 2.02 2.08 2.04The pyrrole nitrogen plane is defined by the four pyrrole nitrogens of the heme group. The fourpyrrole ring planes are each defined by the five atoms of the ring and the first carbon atomattached to each of the four carbons of the ring. The porphyrin ring is defined by the five atomsin each of the four pyrrole rings, the four bridging methine carbon atoms, the first carbon atomof each of the eight side chains of the heme and the central iron atom of the heme. The hemeatom nomenclature used in this table follows the conventions of the Protein Data Bank (seeFigure 1.2).Chapter 6. Multiple Distal Mutations in Cytochrome c 170Table 6.32: Heme propionate hydrogen bond interactions in yeast iso-i cytochromes c withR38A, N521 and F82S replacementsDistances (A)Interaction 1t38A/Wild- R38A/ R38A/ N521/ N521/type R38A N521 F82SI N521 F82S F82S F82SHeme Tyr48 OH 2.81 2.89 2.63 2.92 2.84 3.10 2.76 2.57O1A Watl2l 2.81 2.58 3.48 2.88 2.89 3.47 2.69 2.69Wat168 2.84 3.18 3.50 3.10 3.37 3.42 3.14 3.20Heme G1y41 N 3.21 3.02 3.07 2.90 3.26 3.06 3.01 3.0602A Asn52 ND2 3.33 3.30 — 2.70 — 3.37 — —Trp59 NE1 3.10 2.88 2.84 (3.71) 2.89 3.08 2.95 3.10Wati2i (4.01) (3.50) (3.91) (3.70) (3.91) (4.22) (3.74) (3.56)Heme Thr49 N 2.94 2.72 2.79 3.07 2.92 2.72 2.82 2.89OlDHenie Thr49 OG1 2.64 2.56 2.49 3.16 2.73 2.76 2.53 2.6102D Thr78 OG1 2.90 3.11 2.76 (3.67) 2.86 3.01 2.72 2.80Lys79 N 3.18 2.95 3.03 3.01 3.09 3.13 3.03 3.12Wati2l Arg38 NE 2.81 — 2.97 (3.72) — — 3.09 —Wat168 (3.55) (4.31) (4.17) (4.67) (4.28) (4.92) (3.70) (4.12)WatA — (3.92) — — 2.54 2.47 — 3.31Wat168 Arg38 Nh 2.56 — 2.93 3.33 — — 2.63 —WatB— 3.22 — — 2.89 2.74 — —WatA WatB — 2.91 — — 2.60 2.79 — —Values listed are the distances between hydrogen donor and acceptor atoms. Values given inparentheses are not considered to be hydrogen bonds but are listed for comparison.t from H. Tong, unpublished results.4rom Berghuis et al. (1994a).¶from Louie et al. (1988b).§WatA and WatB are the water molecules which replace the Arg38 side chain and are closestto Wati2l and Wa.t168, respectively.Chapter 6. Multiple Distal Mutations in Cytochrome c 171ABFigure 6.44: Stereo diagrams of the region around the R38A mutation site in yeast iso-i cytochrome c. The structure of wild-type yeast iso-i cytochrome c is drawn with thick lines andsuperimposed on the structures of (a) the N521/F82S mutant protein (thin lines) and (b) theR38A, R38A/N521, R38A/F825 and R38A/N521/F82S mutant proteins (all drawn with thinlines). Water molecules are shown as asterisks, with the two conserved water molecules, Wat 121and Wat 168, labeled.Chapter 6. Multiple Distal Mutations in Cytochrome c 172similar hydrogen bonding interactions with two conserved water molecules, Watl2l and Wat168(Table 6.32). In turn, the conserved Watl2l and Wat168 retain positions comparable to thosefound in the wild-type protein and form hydrogen bonds to the O1A atom of heme propionateA.6.3.3 N521 mutation siteSubstitution of Asn52 by isoleucine involves the exchange of a polar side chain for a nonpolarside chain of approximately equivalent size but of different shape. In the N521 single-site mutantprotein, the highly conserved internal water molecule, Wat166, is excluded from the protein.In addition, the side chain of Tyr67 moves away from the side chain of Met8O (i.d = 0.6 A)and toward that of Thr78, forming a new hydrogen bond interaction (Table 6.33; Figure 6.45;Berghuis et al., 1994a). These changes are observed in all of the combinatorial mutant proteinshaving the N521 mutation (Table 6.33; Figure 6.45), with the Tyr67 side chain undergoing apositional shift of 0.5—0.6 A.6.3.4 F82S mutation siteThe mutation of Phe82 to serine brings about the creation of a solvent channel directly intothe heme pocket, thereby disrupting the nonpolar environment of the heme and leading toa significant drop in the reduction potential of cytochrome c (Louie et al., 1988b). In thecombination mutants containing the F825 mutation, a comparable phenomenon is observed,with a single water molecule being observed in the newly created solvent channel (Figure 6.46)and a significant increase in the solvent exposure of the heme porphyrin ring (Table 6.34; Louleet al., 1988b). As apparent from Table 6.34, increased heme solvent exposure is caused primarilyby the F82S mutation and is not affected by either the R38A or N521 mutations.The side chain of Argl3 is in the vicinity of the F82S mutation site and shows considerablevariability in the conformations adopted in the different combinatorial mutants (Figure 6.46).Nonetheless, the Argl3 side chain occupies the same general surface region in each case, andChapter 6. Multiple Distal Mutations in Cytochrome c 173Table 6.33: Wat166 hydrogen bond interactions in yeast iso-i cytochromes c with R38A, N521and F82S replacementsDistances (A)Interaction R38A/Wild- R38A/ R38A/ N521/ N521/type R38Af N521 F82SI N521 F82S F82S F82SWat166 Asn52 ND2 3.14 (3.72) — (4.13) — 3.21 — —Tyr67 OH 2.62 2.60 — 2.36 — 2.46 — —Thr78 OG1 2.72 2.53 — 2.89 — 2.90 — —Tyr67 Thr78 OG1 (4.18) (4.21) 3.41 (3.73) 3.45 (4.10) 3.38 3.32OH Met8O SD 3.25 3.19 (3.63) 3.02 (3.61) 3.26 (3.43) (3.60)Values listed are the distances between hydrogen donor and acceptor atoms. Values given inparentheses are not considered to be hydrogen bonds but are listed for comparison.t from H. Tong, unpublished results.tfrom Berghuis et al. (1994a).¶from Louie et al. (1988b).Chapter 6. Multiple Distal Mutations in Cytochrome c 174AM80 M80T78 T78BM8 M80\_166 166T78 T78Figure 6.45: Stereo diagrams of the region around the N521 mutation site in yeast iso-i cyto—chrome c. The structure of wild-type yeast iso-i cytochrome c is drawn with thick lines andsuperimposed on the structures of (a) the R38A/F82S mutant protein (thin lines) and (b)the N521, R38A/N521, N521/F82S and R38A/N521/F82S mutant proteins (all drawn with thinlines). The internally bound Wat166 molecule, found in wild-type iso-i cytochrome c and theR38A/F82S mutant protein and located adjacent to Asn52, Tyr67 and Thr78, is representedby an asterisk.Chapter 6. Multiple Distal Mutations in Cytochrome c 175ABFigure 6.46: Stereo diagrams of the region around the F82S mutation site in yeast iso-i cytochrome c. The structure of wild-type yeast iso-i cytochrome c is drawn with thick lines andsuperimposed on the structures of (a) the R38A/N521 mutant protein (thin lines) and (b) theF825, R38A/F82S, N521/F825 and R38A/N521/F82S mutant proteins (all drawn with thinlines). Water molecules bound in the solvent channel formed by the replacement of phenylalanine by serine at position 82 are shown as asterisks.Chapter 6. Multiple Distal Mutations in Cytochrome c 176Table 6.34: Heme solvent accessibility in yeast iso-i cytochromes c with R38A, N521 and F82SreplacementsYeast iso-i cytochrome c structureR38A/Wild- R38A/ R38A/ N521/ N521/type R38Af N521 F82S’ N521t F82Sf F82S F82Sf1. Solvent accessibleheme atoms andsurface areaexposed (A2)CBB 0.0 0.0 0.0 15.4 0.0 12.7 7.8 12.2CHD 2.9 2.6 2.3 3.6 3.6 3.8 0.0 2.3CMC 9.2 10.2 13.2 13.1 9.1 10.4 11.6 12.9CAC 3.4 2.6 2.6 3.4 3.2 3.6 5.1 3.7CBC 20.1 19.9 20.1 21.2 19.2 18.8 18.4 19.4CMD 10.8 11.3 9.8 11.4 10.7 10.5 10.7 9.92. Total heme 46.4 46.6 48.0 68.1 45.8 59.8 53.6 60.4exposure (A2)3. Total heme 513.1 516.2 521.2 511.8 512.4 512.2 514.5 511.0surface (A2)4. % heme surface 9.0 9.0 9.2 13.3 8.9 11.7 10.4 11.8area exposedSolvent exposure was determined by the method of Connolly (1983) with a probe sphere havinga 1.4 A radius.tfor all mutant proteins having the R38A mutation, Wati2i and Wat168 were considered anintegral part of the protein structure. Results for the R38A single mutant protein were providedby H. Tong (personal communication).lfrom Berghuis et al. (1994a).¶from Louie et al. (i988b).Chapter 6. Multiple Distal Mutations in Cytochrome c 177the variability observed for this residue likely arises from the high thermal factors observed forthis side chain (32.4 A2 ill the wild-type protein; Louie & Brayer, 1990; average B is 38.5 A2in F82S mutants).A further structural change is an increase in thermal parameters for residue 82 and thetwo glycines at positions 83 and 84 (Figure 6.47; average B of these three residues is 17.1 A2in wild-type and 32.3 A2 averaged over F82S mutants). This increase in thermal parametersis particularly marked for the N521/F82S (B = +15.0 A2) and R38A/N521/F82S (i.B =+18.3 A2) mutants (Figure 6.47). Increased mobility likely arises from the loss of the tightpacking interactions formed by the aromatic ring of Phe82 which is sandwiched between thissegment of polypeptide chain and the heme group. Another factor is that the water moleculeintroduced into the newly formed solvent channel at residue 82 can form a hydrogen bondto the carbonyl oxygen of Leu68 (d = 3.2 A ) and thereby interfere with the hydrogen bondnormally found between this latter group and the main chain nitrogen atom of Leu85. Thiswould allow more freedom of motion for the flexible Gly83-G1y84 segment of the polypeptidechain. Comparison of the thermal parameters for the F82S single mutant and the wild-typeprotein is not conclusive, probably due to the relatively low resolution (2.8 A) of the structuraldetermination for this mutant protein (Louie et al., 1988b).6.3.5 The conserved internal water, Wat166In all eukaryotic cytochromes c examined thus far, a conserved water molecule is centrallylocated and hydrogen bonded to Asn52, Tyr67 and Thr78 (Bushnell et al., 1990). A shift in theposition of Wat166 toward the heme iron atom is observed in the structure of the oxidized formof yeast iso-i (zd = 1.7 A ) and other cytochromes c (Berghuis & Brayer, 1992). This oxidationstate dependent shift results in the loss of the hydrogen bond between Wat166 and the sidechain of Asn52. Similar structural changes are observed in both the reduced R38A (H. Tong,personal communication) and reduced F82S mutant proteins (Louie et al., 1988b), suggestingthat these structures shift toward the structure of the oxidized state. In the present study, all ofChapter 6. Multiple Distal Mutations in Cytochrome c 178IILegend15.0 and greater10.0 to 15.05.0 to 10.0Figure 6.47: Thermal factor difference matrices from the comparison of wild-type iso-i cytochrome c with the (a) R38A/N521, (b) R38A/F82S, (c) N521/F82S and (d) R38A/N521/F82Smutants. Each matrix point P represents a pairing of amino acids x and y. The value of thepairing is calculated from the equation: = (Bar — By)mutant(B — By)wjId_type where Bis the average thermal factor (A2) of the main chain atoms of a given amino acid i. Positivematrix vaiues are displayed according to the scale shown. Due to the inverse symmetry of thematrix across the diagonal, negative values are redundant and have been omitted. Amino acidsproducing vertical streaJs within the matrix have significantly higher main chain thermal factors in the mutant structure relative to the wild-type structure while horizontal streaks indicatesignificantly lower main chain thermal factors in the mutant protein.Residue Number Residue NumberChapter 6. Multiple Distal Mutations in Cytochrome c 179the protein structures determined were in the reduced state. Since the N521 mutation eliminatesWat 166 from the protein interior, the only protein in the present study which retains this watermolecule is the R38A/F82S double mutant. Surprisingly, Wat166 in the R38A/F82S proteindoes not move from the position it occupies in the reduced wild-type structure (Figure 6.45a)and it retains a hydrogen bond with Asn52 (Table 6.33).6.4 DiscussionThe site-directed mutagenesis technique provides the opportunity to specifically make multiplesite mutations within a single protein. This can be useful in determining the extent of synergisticfunctional and structural changes which arise from the interaction of individual mutations. Inthe present work, a study has been made of the effects of introducing mutations at threedistally separated sites of yeast iso-i cytoclirome c involving the invariant residues Arg38,Asn52 and Phe82. At these three residues, all four possible combinatorial mutant proteins weremade with the three single replacements R38A, N521 and F82S. It appears that the degree ofsynergism between mutation sites can be quite different depending on the particular functionalor structural aspect being assessed.6.4.1 Structural effectsThe replacement amino acids all represent substantial changes in side chain character andtherefore it is not surprising that the introduction of each of these individual mutations intoyeast iso-i cytochrome c leads to significant structural and functional changes (Louie et al.,1988b; Hickey et a!., 1991; Berghuis et a!., 1994a). The structural consequences of each mutationin the combinatorial mutant proteins are similar to those previously observed in the singlemutant proteins (Louie et cii., 1988b; Berghuis et cii., 1994a). In general, the structural effectsof each mutation are independent of the effects arising from mutations made at the other twosites. Thus for this set of mutant proteins with multiple distal substitutions, the structuraleffects observed do not have a synergistic component.Chapter 6. Multiple Distal Mutations in Cytochrome c 1806.4.2 Protein stabilityIn terms of stability to guanidine hydrochloride denaturation (Komar-Panicucci et at., 1992,1994), the effect of individual mutation sites can be understood from the structural changesobserved for each. For example, one structural role of the Arg38 side chain is to providetwo hydrogen bonding groups to interact with two conserved water molecules, Wati2i andWat 168 (Table 6.32). These hydrogen bonds are replaced in the R38A mutant protein by thesubstitution of water molecules (Wat A and B) in place of the side chain of Arg38 (Table 6.32;Figure 6.44). Apparently, the maintenance of these hydrogen bonding interactions is sufficientto preserve the structural integrity of cytochrome c, thereby accounting for the negligible changein protein stability observed upon the introduction of this mutation (Table 6.35).The N521 mutant protein is significantly more stable than wild-type yeast iso-i cytochrome c(Table 6.35). This effect appears to be due to several factors, including replacement of a polarside chain by one of a hydrophobic nature within the hydrophobic core of the protein (Hickeyet at., 1991), displacement of an internal water molecule, Wati66 (Berghuis et at., i994a), andrealignment of the hydrogen bonding network in this region (Table 6.33; Figure 6.45). Finally,the N521 mutation abolishes the structural transition between the reduced and oxidized statesof the protein (Berghuis et at., i994a) and this is likely an important feature in maintainingthe overall stability of the protein, especially in the oxidized state (Berghuis & Brayer, 1992).The F82S replacement destabilizes yeast iso-i cytochrome c (Table 6.35) by the introduction of a solvent channel and bound water molecule into the hydrophobic core of the protein(Figure 6.46; Louie et at., 1988b). This is evident in the increased thermal parameters in theimmediate vicinity of the mutation site (Figure 6.47). Such changes can also account for thelowering of the thermal stability of the heme pocket in the F82S protein relative to the wild-typeprotein (Hildebrandt et aL, 1991).An analysis of the stability data for the four multiple mutant proteins for which structureswere determined reveals the extent to which the R38A, N521 and F82S mutations interact inthis regard (Table 6.35). The combination of the N521 and F82S mutations results in a netChapter 6. Multiple Distal Mutations in Cytochrome c 181Table 6.35: Unfolding of mutant and wild-type yeast iso-i cytochromes c by guanidine hydrochlorideCm [Gdn-HC1]t LGOCytochrome c (+0.1 M) (kcal/mol)Wild-type 1.3 3.8 4.9R38A 1.4 3.4 4.8N521 2.1 4.2 9.0F82S 0.9 3.8 3.4R38A/N521 2.1 3.8 8.2R38A/F82S 0.8 3.4 2.7N521/F82S 1.7 3.4 5.6R38A/N521/F82S 1.7 3.6 6.2tMidpoints of unfolding by guanidine-hydrochioride were taken from Komar-Panicucci et al.(1992).is a measure of the cooperativity of the transition of the protein from the folded to theunfolded state and is derived from the slope of the linear region of a plot of in ([folded protein]/ [unfolded protein] ) vs. [denaturant].§G0 free energy extrapolated to 0 M [Gdn-HC].]; G° = mCmChapter 6. Multiple Distal Mutations in Cytochrome C 182stabilization of cytochrome c due to the larger stabilizing contribution of the N521 replacement.This is not as large as in the N521 single mutant protein due to the destabilizing influence ofthe F82S mutation. Since the R38A mutation has a neutral effect on stability, the R38A/N521double mutant protein has a stability similar to that of the N521 single mutant, whereas theR38A/F82S mutant protein is destabilized to the same extent as the F82S single mutant.Consistent with the stability of these double mutant proteins, the R38A/N521/F82S triplemutant has a stability which is essentially the same as that of the N521/F82S double mutantprotein. Thus with respect to stability, the effects of each of the three mutations are independentand cumulative in the multiple mutant proteins. This is a reflection of the independent natureof the structural changes induced at each individual mutation site, which are distally located.6.4.3 Reduction potential effectsEach of the mutant cytochromes c containing multiple substitution sites has a midpoint reduction potential significantly lower than that of either the wild-type protein or the related proteinswith single site substitutions (Table 6.36). For the R38A/F82S and N521/F82S mutant proteins,the observed reduction potentials are similar to what would be expected if the effect of eachof the single mutations were independent and additive. In contrast, the R38A/N521 mutanthas a reduction potential which is significantly higher than the expected value (Table 6.36).The triple mutant R38A/N521/F82S protein has the lowest reduction potential observed, witha measured value which falls approximately midway within the range of values expected bycombining the various reduction potentials observed for single and double site mutant proteins.The higher than expected reduction potential observed for the R38A/N521 double mutantcytochrome c can be explained by a consideration of the individual contributions at each siteof mutation. Arg38 likely contributes to the reduction potential of cytochrome c throughtwo mechanisms. The first mechanism is the direct electrostatic stabilization of the reducedstate of the heme group by the positively charged guanidinium side chain of this residue.The second mechanism is an electron withdrawing effect on the heme transmitted throughChapter 6. Multiple Distal Mutations in Cytochrome c 183Table 6.36: Reduction potentials for yeast iso-i cytochromes c with R38A, N521 and F82SreplacementsExpectedCytochrome C Emt (mV) IXEm (mV) iEm (mV)Wild-type 285±2 — —R38A 239±2 -46 —N521 23 1±2 -54 —F82S 247±2 -38 —R38A/N521 212±2 -73 -100R38A/F82S 203±2 -82 -84N521/F82S 189±2 -96 -92R38A/N521/F82S 162±2 -123 -111 to -142{Experimental conditions were: 25°C, pH 6.0 and j = 0.1 M. Values are listed relative to astandard hydrogen electrode reference.fElectrochemical reduction potentials were taken from Komar-Panicucci et at. (1992).1based on a simple numerical addition of the effects of individual mutations.¶range of values obtained by combining the various midpoint potentials observed for singleand double site mutant proteins.Chapter 6. Multiple Distal Mutations in Cytochrome c 184the hydrogen bonding network involving Watl2l, Wat168 and the propionate A group of theheme (Figure 6.44; Cutler et al., 1989; Davies et at., 1993). The net result of these effectsis a contribution of +45 mV to the reduction potential of the protein. The contribution ofAsn52 to the reduction potential of cytochrome c likely has three elements. The first consists ofthe maintenance of hydrogen bonding about Wat166, an internal water molecule (Figure 6.45;Berghuis & Brayer, 1992; Berghuis et a!., 1994a,b). Second is the dipole moment of the amideend of the side chain of Asn52 which is oriented to stabilize the reduced form of the heme iron(Langen et at., 1992). A third element is the hydrogen bond formed between the side chain ofAsn52 and the propionate A group of the heme which would be expected to have an electronwithdrawing effect on the heme. These three factors result in a contribution to the midpointreduction potential of +55 mV.Given that both Arg38 and Asn52 contribute to maintaining a high cytochrome c reductionpotential through interaction with a common element, namely the heme propionate A group,a strict additive calculation of the effect of the R38A and N521 mutations could overestimatethe drop in reduction potential expected. As shown experimentally (Table 6.36) this appearsto be the case and suggests that in total, interactions with heme propionate A can only counttoward raising the reduction potential of cytochrome c by some maximum value.The mechanism by which substitution of Phe82 by serine affects reduction potential is different than that of either Arg38 or Asn52 substitution, and is independent of residue replacementsat these latter two positions. The predominant role of Phe82 appears to be in forming a partof the hydrophobic heme pocket and in excluding solvent from this pocket, and it contributes+4O mV to cytochrome c reduction potential through these mechanisms (Louie et at., 1988b;Chapter 3). Phe82 does not appear to influence any structural features affected by either Arg38or Asn52. This is supported by the observation that the reduction potentials of the R38A/F82Sand N521/F82S double mutant proteins can be simply calculated based on the addition of theindividual effects of each single site substitution.More difficult to anaiyze is the effect on reduction potential when all three substitutions areChapter 6. Multiple Distal Mutations in Cytochrome c 185combined in a single mutant protein. If it is assumed that all three mutations have independentand additive effects, the theoretical reduction potential (+147 mV) would be lower than theexperimentally observed value (+162 mV; Table 6.36). As discussed previously, this discrepancy may be explained by the interdependence of the R38A and N521 mutations. However,addition of the effects of the R38A/N521 double mutant and the F82S mutant gives a theoretical reduction potential (+174 mV) which is greater than the observed experimental value(+162 mV). Therefore, the addition of the F82S mutation to the R38A/N521 double mutanthas in some manner increased the net effect on the reduction potential of the protein. Onepossibility is that the cumulative alterations in side chain packing resulting from the multiplemutations have functional consequences in addition to their structural effects. For example,it has previously been observed that the disruption of hydrophobic packing at a site which isdistant from the heme group can affect the reduction potential of this latter group (Murphyet at., 1993). A second possibility arises from the fact that only one water molecule replacesthe Arg38 side chain in the R38A/N521/F82S mutant protein whereas two water molecules arefound in the other mutant proteins with the R38A substitution (Table 6.32). The presence ofonly one hydrogen bonding group in this region of the protein may account for the discrepancy between the theoretical and observed reduction potentials. Evidence for this arises fromconsideration of the R38K mutant of yeast iso-i cytochrome c, in which a drop in reductionpotential of 23 mV is observed (Cutler et at., 1989). In this case, the substitution of Arg38by lysine results in a decrease of available hydrogen bonding groups while still maintaining thepositive charge of the side chain.To conclude, the introduction of multiple mutations within a single protein clearly affectsfunctional and structural properties in different ways. In some cases the effects of multiplemutations are strictly additive, such as global protein stability in the present work, while theeffects on other properties will be synergistic or partly so. This of course arises because withina single protein it is not uncommon that individual residues perform multiple functions andthat these functions overlap among multiple residues. This poses a dilemma in understandingChapter 6. Multiple Distal Mutations in Cytochrome c 186the roles of residues in protein function, especially in sorting out the individual contributionsof each residue to each property. It also suggests that attempts at protein engineering, whetherin modification of existing proteins or in attempts to develop new activities de novo, will befraught with difficulties in assessing and compensating for the multiple roles of individual aminoacids.SummaryThe work described in this dissertation focused on understanding the contributions of Phe82,Leu85 and associated residues (Arg38, Asn52, Leu94) in mediating reduction potential, electrontransfer, protein complex formation and protein stability in cytochrome c. Overall four specificobjectives were pursued. The first of these concerned the roles of the invariant Phe82 and thehighly conserved Leu85 in forming interactions at the complexation interface between cytochrome c and electron transfer partners. The importance of Phe82 to this process was clearlyshown in an analysis of the F82Y mutant protein (Chapter 3) where the additional hydroxylgroup of Tyr82 is in direct spatial conffict with the side chain of Leu85, leading to rotation ofthe side chain of Tyr82 out toward the protein surface. This alteration of the surface contour inthis region of cytochrome c disrupts complex formation with redox partner proteins and leadsto perturbation of electron transfer kinetics. It was further shown that this structural perturbation is mitigated when a smaller side chain, such as that of alanine, is substituted at Leu85.In contrast, despite earlier predictions that Leu85 is an important determinant in cytochrome celectron transfer complexation, the current studies show this is unlikely to be the case sincethe considerable surface contour perturbations made by various substitutions at this residue donot correspondingly translate into significant changes in this property. However, it was shownfrom these results that the placement of the side chain of Argl3 is dependent on the presenceof both Phe82 and Leu85 and one role of these residues may be to restrict interactions of thisside chain to the interactive face of cytochrome c.A second objective of this thesis was to characterize the nature of an internal hydrophobicleucine cluster containing Leu85 and the nearby Leu94, as well as an adjacent internal hydrophobic cavity. Both of these features are found next to tile most buried edge of the hemegroup. As shown in Chapter 4, much of the ability of cytochrome c to absorb the introduction187Summary188of mutations at Leu85 and Leu94 appears to be a consequence of the conformational flexibilityafforded by this leucine cluster and the nearby internal cavity. Notably, these two structuralfeatures also facilitate large structural changes associated with heme rotation as observed in theF82M mutant protein. The leucine cluster and internal cavity may also play a functional role byproviding structural flexibility between oxidation states in the course of cytochrome c electrontransfer. Such conformational flexibility is not observed in an adjacent structural feature, thehighly conserved interface formed by the nearly perpendicular packing of the N and C-terminalhelices. The interaction between these helices is maintained despite structural changes in thisregion arising from substitutions at either Leu85 or Leu94.A third objective was to investigate the factors regulating the reduction potential of cytochrome c, especially with respect to the role of Phe82 and the effect of aliphatic replacements atthis site. Local structural changes in the F82L mutant protein were found to result in preservation of the hydrophobic heme environment and thus maintenance of heme reduction potential(Chapter 5). In contrast, the F821 mutant protein displays increased heme solvent exposureand a corresponding decrease in heme reduction potential. Although the F82M mutant proteinalso has an increased heme solvent exposure, no decrease in reduction potential is observed,possibly due to shielding of the heme from the solvent environment by heme substituent sidechains. This shielding effect also mitigates the impact on reduction potential when polar groupsare introduced into the heme environment as the result of substitutions at Phe82, Leu85 andLeu94.The final objective of this thesis was to place our understanding of the role of Phe82 ina broader perspective by undertaking analyses of yeast iso-i cytochromes c having multiplemutations at Phe82 and two other distally located sites. The multiple mutant proteins studiedcontained all possible combinations of the substitutions R38A, N521 and F82S, where each ofthese individual mutations by themselves would result in a significant decrease in heme reduction potential (Chapter 6). These studies showed that the structural consequences of eachof these amino acid substitutions were independent, in agreement with observations related toSummary 189protein stability. However, in terms of heme reduction potential, two results were observed. Forsubstitution of Phe82 by serine, the mechanism by which reduction potential is lowered is different than that occurring at either of the Arg38 and Asn52 sites, and is independent of residuereplacements at these latter two positions. For Arg38 and Asn52, overlapping interactions leadto higher reduction potentials in proteins with multiple mutations than would be expected froma strict additive effect of substitutions at these residues. This appears to arise from interactionof these two amino acids with a common heme element, namely the heme propionate A group.Overall, the analysis of multiple mutations in cytochrome c shows that the consequences ofindividual mutations can be completely independent or alternatively show varying degrees ofinterdependence, depending on a given protein property and the manner in which mutations arecombined. These studies and those discussed earlier point out the complex interrelationshipsthat occur between amino acids in a protein and the significant challenge this presents whendesigning experiments to understand the individual roles of these residues in protein function.BibliographyAbrams, R., Altschul, A.M. & Hogness, T.R. (1942). Cytochrome c peroxidase. II. Theperoxidase-hydrogen peroxidase complex. J. Biol. Chem. 142, 303—316.Altschul, A.M., Abrams, R. & Hogness, T.R. (1940). Cytochrome c peroxidase. J. Biol. Chem.136, 777—794.Anderson, D.E., Hurley, J.H., Nicholson, H., Baase, W.A. & Matthews, B.W. (1993). Hydrophobic core repacking and aromatic-aromatic interaction in the thermostable mutantof T4 lysozyme Serll7—÷Phe. Protein Sci. 2, 1285—1290.Arndt, T.J.W. & Wonacott, A.J., eds. (1977). The Rotation Method in Crystallography. North-Holland Publishing Company, Amsterdam.Bach, S.J., Dixon, M. & Keffin, D. (1942a). A new soluble cytochrome component from yeast.Nature (London). 149, 21.Bach, S.J., Dixon, M. & Zerfas, L.G. (1942b). Lactic dehydrogenase of yeast. Nature (London).149, 48—49.Baldwin, E.P., Hajiseyedjavadi, 0., Baase, W.A. & Matthews, B.W. (1993). The role of backbone fiexibifity in the accommodation of variants that repack the core of T4 lysozyme.Science. 262, 1715—1718.Barker, P.D., Mauk, M.R. & Mauk, A.G. (1991). Proton titration curve of yeast iso-iferricytochrome c. Electrostatic and conformational effects of point mutations. Biochemistry. 30, 2377—2383.Beratan, D.N., Onuchic, J.N., Winkler, J.R. & Gray, H.B. (1992). Electron-tunneling pathwaysin proteins. Science. 258, 1740—1741.Berghuis, A.M. & Brayer, G.D. (1992). Oxidation state-dependent conformational changes incytochrome c. J. Mat. Biol. 223, 959—976.Berghuis, A.M., Guillemette, J.G., McLendon, G., Sherman, F., Smith, M. & Brayer, G.D.(1994a). The role of a conserved internal water molecule and its associated hydrogen bondnetwork in cytochrome c. J. Mol. Riot. 236, 786—799.Berghuis, A.M., Guillemette, J.G., Smith, M. & Brayer, G.D. (1994b). Mutation of tyrosine-67to phenylalanine in cytochrome c significant1 alters the local heme environment. J. Mol.Biol. 235, 1326—1341.Bernstein, F.C., Koetzle, T.F., Williams, G.J.B., Meyer, E.F., Bruce, M.D., Rodger, J.R.,Kennard, 0., Shimanouchi, T. & Tasmui, M. (1977). The protein databank: a computer-based archival file for macromolecular structures. J. Mat. Biol. 112, 535—542.190BIBLIOGRAPHY 191Berroteran, R.W. & Hampsey, M. (1991). Genetic analysis of yeast iso-1-cytochrome c structural requirements: suppression of G1y6 replacements by an Asn52 —* Re replacement.Arch. Biochem. Biophys. 288, 261—269.Blundell, T.L. & Johnson, L.N. (1976). Protein Crystallography. Academic Press, New York.Bone, R., Silen, J.L. & Agard, D.A. (1989). Structural plasticity broadens the specificity of anengineered protease. Nature (London). 339, 191—195.Brayer, G.D. & Murphy, M.E.P. (1995). Structural studies of eukaryotic cytochroines c. In TheCytochrome c Handbook. (Mauk, A.G. & Scott, R.A., eds.). Plenum Press, New York. inpress.Brennan, R.G., Wozniak, J., Faber, R. & Matthews, B.W. (1988). Crystaffization of mutantlysozymes from bacteriophage T4. J. Crystal Growth. 90, 160—167.Brflnger, A.T. (1990). Slow-cooling protocols for crystallographic refinement by simulatedannealing. Acta Crystallogr. sect. A. 46, 585—593.Brünger, A.T. (1992). X-PLOR. Version 3.1. A System for X-ray Crystallography and NMR.Yale University Press, New Haven.Briinger, A.T., Kuriyan, J. & Karplus, M. (1987). Crystallographic R factor refinement bymolecular dynamics. Science. 235, 458—460.Bryant, C., Strottmann, J.M. & Stellwagen, E. (1985). Refolding a disuffide dimer of cytochrome c. Biochemistry. 24, 3459—3464.Burch, A.M., Rigby, S.E.J., Funk, W.D., MacGillivray, R.T.A., Mauk, M.R., Mauk, A.G. &Moore, G.R. (1990). NMR characterization of surface interactions in the cytochrome b5-cytochrome c complex. Science. 247, 831—833.Burley, S.K. & Petsko, G.A. (1985). Aromatic-aromatic interaction: a mechanism of proteinstructure stabilization. Science. 229, 23—28.Burrows, A.L., Guo, L.-H., Hill, H.A.O., Mcbendon, G. & Sherman, F. (1991). Direct electrochemistry of proteins. Investigations of yeast cytochrome c mutants and their complexeswith cytochrome b5. Eur. J. Biochem. 202, 543—549.Bushnell, G.W., Louie, G.V. & Brayer, G.D. (1990). High-resolution three-dimensional structure of horse heart cytochrome c. J. Mol. Biol. 214, 585—595.Busse, S.C., Moench, S.J. & Satterlee, J.D. (1990). One- and two-dimensional proton NMRstudies of cys-102 S-methylated yeast isozyme-1 ferricytochrome c. Biophys. J. 58, 45—51.Carter, Jr., C.W., ed. (1990). Protein and Nucleic Acid Crystallization., vol. 1 of Methods.Academic Press, San Diego.BIBLIOGRAPHY 192Churg, A.K. & Warshel, A. (1986). Control of the redox potential of cytochrome c and microscopic dielectric effects in proteins. Biochemistry. 25, 1675—1681.Cohen, H.J. & Fridovich, I. (1971). Hepatic sulilte oxidase: purification and properties. J. Biol.Chem. 246, 359—366.Connolly, M.L. (1983). Solvent-accessible surfaces of proteins and nucleic acids. Science. 221,709—713.Connolly, M.L. (1985). Computation of molecular volume. J. Am. Chem. Soc. 107, 1118—1124.Corin, A.F., Hake, R.A., McLendon, G., Hazzard, J.T. & Toffin, G. (1993). Effects of surfaceamino acid replacements in cytochrome c peroxidase on intracomplex electron transferfrom cytochrome c. Biochemistry. 32, 2756—2762.Crowther, R.A. (1972). The fast rotation function. In The Molecular Replacement Method.(Rossmari, M.G., ed.), pp. 173—178. Gordon & Breach, New York.Crowther, R.A. & Blow, D.M. (1967). A method of positioning a known molecule in an unknowncrystal structure. Acta Crystallogr. 23, 544—548.Cruickshank, D.W.J. (1949). The accuracy of electron-density maps in x-ray analysis withspecial reference to dibenzyl. Acta Crystallogr. 2, 65—82.Cruickshank, D .W.J. (1954). The accuracy of electron-density maps in x-ray analysis: correction. Acta Crystallogr. 7, 519.Cusanovich, M.A., Meyer, T.E. & Toffin, G. (1988). c-type cytochromes: oxidation-reductionproperties. In Heme Proteins. (Eichorn, G.L. & Marzilli, J.G., eds.), vol. 7 of Advances inInorganic Biochemistry, pp. 37—92. Elsevier, New York.Cutler, R.L., Davies, A.M., Creighton, S., Warshel, A., Moore, G.R., Smith, M. & Mauk,A.G. (1989). Role of arginine-38 in regulation of the cytochrome c oxidation-reductionequilibrium. Biochemistry. 28, 3188—3197.Cutler, R.L., Pielak, G.J., Mauk, A.G. & Smith, M. (1987). Replacement of cysteine-107of Saccharomyces cerevisiae iso-1-cytochrome c with threonine: improved stability of themutant protein. Protein Eng. 1, 95—99.Das, G., Hickey, D.R., McLendon, D., McLendon, G. & Sherman, F. (1989). Dramatic thermostabilization of yeast iso-1-cytochrome c by an asparagine — isoleiicine replacement atposition 57. Proc. Natl. Acad. Sci. USA. 86, 496—499.Davies, A.M., Guillemette, J.G., Smith, M., Greenwood, C., Thurgood, A.G.P., Mauk, A.G.& Moore, G.R. (1993). Redesign of the interior hydropliilic region of mitochondrial cytochrome c by site-directed mutagenesis. Biochemistry. 32, 5431—5435.Dempsey, S. (1986). The UCSD Molecular Modeling System.BIBLIOGRAPHY 193Drenth, J. (1994). Principles of Protein X-ray Crystallography. Springer-Verlag, New York.Eltis, L.D., Herbert, R.G., Barker, P.D., Mauk, A.G. & Northrup, S.H. (1991). Reductionof horse heart ferricytochrome c by bovine liver ferrocytochrome 65. Experimental andtheoretical analysis. Biochemistry. 30, 3663—3674.Eltis, L., Mauk, A.G., Hazzard, J.T., Cusanovich, M.A. & Tollin, G. (1988). Kinetics of fiavinsemiquinone reduction of the components of the cytochrome c-cytochrome 65 complex.Biochemistry. 27, 5455—5460.Eriksson, A.E., Baase, W.A., Zhang, X.-J., Heinz, D.W., Blaber, M., Baldwin, E.P. &Matthews, B.W. (1992). Response of a protein structure to cavity-creating mutationsand its relation to the hydrophobic effect. Science. 255, 178—183.Erman, J.E., Kang, D.S., Kim, K.L., Summers, F.E., Matthis, A.L. & Vitello, L.B. (1991).Electron transfer within the cytochrome c-cytochrome c peroxidase complex: dependenceof the transient-state and steady-state kinetics on ionic strength. Mol. Cryst. Liq. Cryst.194, 253—258.Everest, A.M., WaJlin, S.A., Stemp, E.D.A., Nocek, J.M., Mauk, A.G. & Hoffman, B.M. (1991).Aromatic hole superexchange through position 82 of cytochrome c is not required forintracomplex electron transfer to zinc cytochrome c peroxidase. J. Am. Chem. Soc. 113,4337—4338.Ferguson-Miller, S., Brautigan, D.L. & Margoliash, E. (1979). The electron transfer function ofcytochrome c. In The Porphyrins. Volume VII. (Dolphin, D., ed.), pp. 149—240. AcademicPress, New York.Fitzgerald, P.M.D. (1988). MERLOT: an integrated package of computer programs for thedetermination of crystal structures by molecular replacement. J. Appi. Crystallogr. 21,273—278.Fredericks, Z.L. & Pielak, G.J. (1993). Exploring the interface between the N- and C-terminalhelices of cytoclirome c by random mutagenesis within the C-terminal helix. Biochemistry.32, 929—936.Gao, Y., Boyd, J., Pielak, G.J. & Wiffiams, R.J.P. (1991). Comparison of reduced and oxidizedyeast iso-1-cytochrome c using proton paramagnetic shifts. Biochemistry. 30, 1928—1934.Gao, Y., Boyd, J., Wiffiams, R.J.P. & Pielak, G.J. (1990). Assignment of proton resonances,identification of secondary structural elements, and analysis of backbone chemical shiftsfor the C1O2T variant of yeast iso-1-cytochrome c and horse cytochrome c. Biochemistry.29, 6994—7003.Gray, T.M. & Matthews, B.W. (1984). Intrahelical hydrogen bonding of serine, threonine andcysteine residues within o-helices and its relevance to membrane-bound proteins. J. Mol.Biol. 175, 75—81.BIBLIOGRAPHY 194Greene, R.M., Betz, S.F., Hilgen-Willis, S., Auld, D.S., Fend, J.B. & Pielak, G.J. (1993).Changes in global stability and local structure of cytochrome c upon substitutingPhenylalanine-82 with tyrosine. J. Inorg. Biochem. 51, 663—676.Gufflemette, J.G., Barker, P.D., Eltis, L.D., Lo, T.P., Smith, M., Brayer, G.D. & Mauk, A.G.(1994). Analysis of the bimolecular reduction of ferricytochrome c by ferrocytochrome b5through mutagenesis and molecular modeffing. Biochimie. 76, 592—604.Hampsey, D.M., Das, G. & Sherman, F. (1986). Amino acid replacements in yeast iso-icytochrome c. Comparison with the phylogenetic series and the tertiary structure of relatedcytochromes c. J. Biol. Chem. 261, 3259—3271.Hampsey, D.M., Das, G. & Sherman, F. (1988). Yeast iso-1-cytochrome C: genetic analysis ofstructural requirements. FEBS Lett. 231, 275—283.Harbury, H.A., Cronin, J.R., Fanger, M.W., Hettinger, T.P., Murphy, A.J., Myer, Y.P. &Vinogradov, S.N. (1965). Complex formation between methionine and a heme peptidefrom cytochrome c. Proc. Nati. Acad. Sci. USA. 54, 1658—1664.Hartshorn, R.T., Mauk, A.G., Mauk, M.R. & Moore, G.R. (1987). NMR study of the interactionbetween cytochrome b5 and cytochrome c. Observation of a ternary complex formed bythe two proteins and [Cr(en)3j.FEBS Lett. 213, 391—395.Hazzard, J.T., Mauk, A.G. & Tollin, G. (1992). Laser flash photolysis studies of electrontransfer mechanisms in cytochromes: an aromatic residue at position 82 is not required forcytochrome c reduction by flavin semiquinones or electron transfer from cytochrome c tocytochrome oxidase. Arch. Biochern. Biophys. 298, 91—95.Hendrickson, W.A. (1985). Stereochemically restrained refinement of macromolecular structures. Methods Enzymol. 115, 252—270.Hickey, D.R., Berghuis, A.M., Lafond, G., Jaeger, J.A., Cardillo, T.S., McLendon, D., Das, G.,Sherman, F., Brayer, G.D. & McLendon, G. (1991). Enhanced thermodynamic stabilitiesof yeast iso-1-cytochromes c with amino acid replacements at positions 52 and 102. J.Biol. Chem. 266, 11686—11694.Higashi, T. (1990). Auto-indexing of oscillation images. J. Appi. Crystallogr. 23, 253—257.Hildebra.ndt, P., Pielak, G.J. & Williams, R.J.P. (1991). Structural studies of yeast iso-icytochrome c mutants by resonance Raman spectroscopy. Eur. J. Biochem. 201, 211—216.Hhlgen, S.E. & Pielak, G.J. (1991). The function of the Saccharomyces cerevisiae iso-1-cytochrome c gene is independent of the codon at invariant residue Phe82 when the geneis present on a low-copy-number vector. Protein Erzg. 4, 575—578.Ho, P.S.. Sutoris, C., Liang, N., Margoliash, E. & Hoffman, B.M. (1985). Species specificityof long-range electron transfer within the complex between zinc-substituted cytochrome cperoxidase and cytochrome c. J. Am. Chem. Soc. 107, 1070—1071.BIBLIOGRAPHY195Huang, Y., Beeser, S., Guillemette, J.G., Storms, R.K. & Kornblatt, J.A. (1994). Mutationsof iso-1-cytochrome c at positions 13 and 90. Separate effects on physical and functionalproperties. Eur. J. Biochem. 223, 155—160.Inglis, S.C., Guillemette, J.G., Johnson, J.A. & Smith, M. (1991). Analysis of the invariantPhe82 residue of yeast iso-1-cytochrome c by site-directed mutagenesis using a phagemidyeast shuttle vector. Protein Eng. 4, 569—574.Ito, A. (1980a). Cytochromeb5-like hemoprotein of outer mitochondrial membrane; OM cytochrome b. I: Purification of OM cytochrome b from rat liver mitochondria and comparisonof its molecular properties with those of cytochrome b5. J. Biochem. 87, 63—71.Ito, A. (1980b). Cytochromeb5-like hemoprotein of outer mitochondrial membrane; OM cytochrome b. II: Contribution of OM cytochrome b to rotenone insensitive NADH-cyt c reductase activity. J. Biochern. 87, 73—80.Kassner, R.J. (1972). Effects of nonpolar environments on the redox potentials of heme complexes. Proc. Nati. Acad. Sci. USA. 69, 2263—2267.Kassner, R.J. (1973). A theoretical model for the effects of local nonpolar environments on theredox potentials of cytocliromes. J. Am. Chem. Soc. 95, 2674—2677.Katz, B. & Kossiakoff, A. (1990). Crystal structures of subtilisin BPN’ variants containingdisulfide bonds and cavities: concerted structural rearrangements induced by mutagenesis.Proteins: Struct. Funct. Genet. 7, 343—357.Komar-Panicucci, S., Bixier, J., Bakker, G., Sherman, F. & McLendon, G. (1992). Tuning theredox potential of cytochrome c through synergistic site replacements. J. Am. Chem. Soc.114, 5443—5445.Komar-Panicucci, S., Weis, D., Bakker, G., Qiao, T., Sherman, F. & Mcbendon, G. (1994).Thermodynamics of the equilibrium unfolding of oxidized and reduced Sacchczromyces cerevisiae iso-1-cytochromes c. Biochemistry. 33, 10556—10560.Kreil, G. (1965). Die C-terminale aminosauresequenz des tunfischcytochroms c. Z. Physiol.Chemie. 340, 86—87.Ladd, M.F.C. & Palmer, R.A. (1985). Structure Determination by X-Ray Crystallography.Plenum Press, New York.Langen, R., Brayer, G.D., Berghuis, A.M., McLendon, G., Sherman, F. & Warshel, A. (1992).Effect of the Asn52—*lle mutation on the redox potential of yeast cytochrome c. Theoryand experiment. J. Mol. Biol. 224, 589—600.Lattman, E.E. & Love, W.E. (1972). A rotational search procedure for detecting a knownmolecule in a crystal. Acta Crystallogr. sect. B. 26, 1854—1857.Lederer, F., Ghrir, R., Guiard, B., Cortail, S. & Ito, A. (1983). Two homologous cytochromes b5in a single cell. Eur. J. Biochem. 132, 95—102.BIBLIOGRAPHY196Leung, C.J., Nail, B.T. & Brayer, G.D. (1989). Crystaffization of yeast iso-2-cytochromeusing a novel hair seeding technique. J. Mol. Biol. 206, 783—785.Lo, T.P. & Brayer, G.D. (1995). Structural analyses of aliphatic residue replacements of theinvariant phenylalanine 82 in cytochrome c. Biochemistry, submitted.Lo, T.P., Guillemette, J.G., Louie, G.V., Smith, M. & Brayer, G.D. (1995a). Structural studiesof the roles of residues 82 and 85 at the interactive face of cytochrome c. Biochemistry.34, 163—171.Lo, T.P., Komar-Panicucci, S., Sherman, F., McLendon, G. & Brayer, G.D. (1995b). Thestructural and functional effects of multiple mutations at distal sites in cytochrome c.Biochemistry, in press.Lo, T.P., Murphy, M.E.P., Guillemette, J.G., Smith, M. & Brayer, G.D. (1995c). Replacementsin a conserved leucine cluster in the hydrophobic heme pocket of cytochrome c. ProteinSci. 4, 198—208.Louie, G.V. & Brayer, G.D. (1989). A polypeptide chain-refolding event occurs in the G1y82variant of yeast iso-1-cytochrome c. J. Mo!. Biol. 210, 313—322.Louie, G.V. & Brayer, G.D. (1990). The high resolution refinement of yeast iso-1-cytochrome cand comparison with other eukaryotic cytochromes c. J. Mol. Biol. 214, 527—555.Louie, G.V., Hutcheon, W.L.B. & Brayer, G.D. (1988a). Yeast iso-1-cytochrome c: a 2.8 Aresolution three-dimensional structure determination. J. Mol. Biol. 199, 295—314.Louie, G.V., Pielak, G.J., Smith, M. & Brayer, G.D. (1988b). The role of Phe-82 in yeastiso- 1-cytochrome c and remote conformational changes induced by a serine residue at thisposition. Biochemistry. 27, 7870—7876.Lum, V.R., Brayer, G.D., Louie, G.V., Smith, M. & Mauk, A.G. (1987). Computer modellingof yeast iso-1-cytochrome c-yeast cytochrome c peroxidase complexes. Protein Structure,Folding and Design. 2, 143—150.Luzzati, V. (1952). Traitement statistique des erreurs dans la determination des structurescristallines. Acta Crystallogr. 5, 802—810.Marchon, J.C., Mashiko, T. & Reed, C.A. (1982). How does nature control cytochrome redox potentials? In Electron Transport and Oxygen Utilization. (Ho, C., ed.), pp. 67—72.Elsevier, North, Amsterdam.Margoliash, E. & Bosshard, H.R. (1983). Guided by electrostatics, a textbook protein comesof age. Trends Biochem. Sci. 8, 316—320.Margoliash, E., Smith, E.L., Kreil, G. & Tuppy, H. (1961). Amino acid sequence of horse heartcytochrome c. Nature (London). 192, 1125—1127.BIBLIOGRAPHY 197Mathews, F.S. (1985). The structure, function and evolution of cytochromes. Progr. Biophys.Mol. Biol. 45, 1—56.Matthew, J.B., Weber, P.C., Salemme, F.R. & Richards, F.M. (1983). Electrostatic orientationduring electron transfer between flavodoxin and cytochrome c. Nature (London,). 301,169—171.Matthews, B.W. (1968). Solvent content of protein crystals. J. Mo!. Biol. 33, 491—497.Mauk, M.R., Barker, P.D. & Mauk, A.G. (1991). Proton linkage of complex formation betweencytochrome c and cytochrome b5: electrostatic consequences of protein-protein interactions. Biochemistry. 30, 9873—9881.Mauk, M.R., Ferrer, J.C. & Mauk, A.G. (1994). Proton linkage in formation of the cytochrome c-cytochrome c peroxidase complex: electrostatic properties of the high- and low-affinity cytochrome binding sites on the peroxidase. Biochemistry. 33, 12609—12614.Mauk, M.R., Mauk, A.G., Weber, P.C. Li Matthew, J.B. (1986). Electrostatic analysis of the interaction of cytochrome c with native and dimethyl ester heme substituted cytochrome b5.Biochemistry. 25, 7085—7091.McLendon, G., Hickey, D., Berghuis, A., Sherman, F. & Brayer, G. (1991). Effects of reactionfree energy in biological electron transfer in vitro and in vivo. In Electron Transfer inInorganic, Organic, and Biological Systems. (Bolton, J.R., Mataga, N. & McLendon, G.,eds.), vol. 228 of Advances in Chemistry Series, pp. 179—189. American Chemical Society,Washington.McLeod, R.M., Farkas, W., Fridovich, I. Li Handler, P. (1961). Purification and properties ofhepatic sulfite oxidase. J. Biol. Chem. 236, 1841—1846.McPherson, A. (1982). Preparation and Analysis of Protein Crystals. John Wiley & Sons, NewYork.McPherson, A. (1990). Current approaches to macromolecular crystallization. Eur. J. Biochem.189, 1—23.McRee, D.E. (1993). Practical Protein Crystallography. Academic Press, San Diego.McRee, D.E., Redford, S.M., Getzoff, E.D., Lepock, J.R., Hallewell, R.A. & Tamer, J.A. (1990).Changes in crystallographic structure and thermostability of a Cu,Zn superoxide dismutasemutant resulting from the removal of a buried cysteine. J. Biol. Chem. 265, 14234—14241.Meyer, T.E., Rivera, M., Walker, F.A., Mauk, M.R., Mauk, A.G., Cusanovich, M.A. & Toffin,G. (1993). Laser flash photolysis studies of electron transfer to the cytochrome b5-cytochrome c complex. Biochemistry. 32, 622—627.Michel, B., Mauk, A.G. Li Bosshard, H.R. (1989). Binding and oxidation of mutant cytochromes c by cytochrome-c oxidase. FEBS Lett. 243, 149—152.BIBLIOGRAPHY 198Moench, S.J., Chroni, S., Lou, B.-S., Erman, J.E. & Satterlee, J.D. (1992). Proton NMR comparison of noncovalent and covalently cross-linked complexes of cytochrome c peroxidasewith horse, tuna, and yeast ferricytochromes c. Biochemistry. 31, 3661—3670.Moench, S.J., Erman, J.E. & Satterlee, J.D. (1993). Species-specific differences in covalentlycrosslinked complexes of yeast cytochrome c peroxidase with horse and yeast iso-i ferricytochromes c. mt. J. Biochem. 25, 1335—1342.Moench, S.J. & Satterlee, J.D. (1989). Proton NMR comparison of the Saccharomyces cerevisiaeferricytochrome c isozyme- 1 monomer and covalent disulfide dimer. J. Biol. Chem. 264,9923—9931.Montgomery, D., Leung, D., Smith, M., Shalit, P., Faye, G. & Hall, B. (1980). Isolation andsequence of a gene for iso-2-cytochrome c in Saccharomyces cerevisiae. Proc. Natl. Acad.Sci. USA. 77, 541—545.Moore, G.R., Harris, D.E., Leitch, F.A. & Pettigrew, G.W. (1984). Characterisation of ionisations that influence the redox potential of mitochondrial cytochrome c and photosyntheticbacterial cytochromes c2. Biochim. Biophys. Acta. 764, 331—342.Moore, G.R. & Pettigrew, G.W. (1990). Cytochromes C: Evolutionary, Structural and Physicochemical Aspects. Springer-Verlag, Berlin.Moore, G.R., Pettigrew, G.W. & Rogers, N.K. (1986). Factors influencing redox potentials ofelectron transfer proteins. Proc. Nati. Acad. Sci. USA. 83, 4998—4999.Moore, G.R. & Williams, R.J.P. (1977). Structural basis for the variation in redox potential ofcytochromes. FEBS Lett. 79, 229—232.Mori, E. & Morita, Y. (1980). Amino acid sequence of cytochrome c from rice. J. Biochem.87, 249—266.Murphy, M.E.P., Fetrow, J.S., Burton, R.S. & Brayer, G.D. (1993). The structure and functionof omega loop A replacements in cytochrome c. Protein Sci. 2, 1429—1440.Murphy, M.E.P., Nail, B.T. & Brayer, G.D. (1992). Structure determination and analysis ofyeast iso-2-cytochrome c and a composite mutant protein. J. Mol. Biol. 227, 160—176.Nair, S.K., Calderone, T.L., Christianson, D.W. & Fierke, C.A. (1991). Altering the mouth ofa hydrophobic pocket. Structure and kinetics of human carbonic anhydrase II mutants atresidue Val-121. J. Biol. Chem. 266, 17320—17325.Nair, S.K. & Christianson, D.W. (1993). Structural consequences of hydrophilic amino acidsubstitutions in the hydrophobic pocket of human carbonic anhydrase II. Biochemistry.32, 4506—4514.BIBLIOGRAPHY 199Nocek, J.M., Stemp, E.D.A., Finnegan, M.G., Koshy, T.I., Johnson, M.K., Margoliash, E.,Mauk, A.G., Smith, M. & Hoffman, B.M. (1991). Low-temperature, cooperative conformational transition within [Zn-cytochrome c peroxidase, cytochrome c] complexes: variationwith cytochrome. J. Am. Chem. Soc. 113, 6822—6831.North, A.C.T., Phillips, D.C. & Mathews, F.S. (1968). A semi-empirical method of absorptioncorrection. Acta Crystallogr. sect. A. 24, 351—359.Northrup, S.H., Boles, J.O. & Reynolds, J.C.L. (1988). Brownian dynamics of association ofcytochrome c and cytochrome c peroxidase electron transfer proteins. Science. 241, 67—70.Northrup, S.H., Thomasson, K.A., Mifier, C.M., Barker, P.D., Eltis, L.D., Guillemette, J.G.,Inglis, S.C. & Mauk, A.G. (1993). Effects of charged amino acid mutations on the bimolecular kinetics of reduction of yeast iso- 1-ferricytochrome c by bovine ferrocytochrome b5.Biochemistry. 32, 6613—6623.Ochi, H., Hata, Y., Tanaka, N., Kakudo, M., Sakurai, T., Aihara, S. & Morita, Y. (1983).Structure of rice ferricytochrome c at 2.0 A resolution. J. Mol. Biol. 166, 407—418.Pelletier, H. & Kraut, J. (1992). Crystal structuie of a complex between electron transferpartners, cytochrome c peroxidase and cytochrome c. Science. 258, 1748—1755.Pettigrew, G.W. & Moore, G.R. (1987). Cytochromes c: Biological Aspects. Springer-Verlag,Berlin.Pielak, G.J., Atkinson, R.A., Boyd, J. & Wiffiams, R.J.P. (1988). Two-dimensional NMR as aprobe of structural similarity applied to mutants of cytochrome c. Eur. J. Biochem. 177,179—185.Pielak, G.J., Mauk, A.G. & Smith, M. (1985). Site-directed mutagenesis of cytochrome c showsthat an invariant Phe is not essential for function. Nature (London). 313, 152—154.Poulos, T.L. & Finzel, B.C. (1984). Heme enzyme structure and function. In Peptide andProtein Reviews. (Hear M.T.W., ed.), vol. 4, pp. 115—171. Dekker, New York.Poulos, T.L. & Kraut, J. (1980). A hypothetical model of the cytochrome c peroxidasecytochrome c electron transfer complex. J. Biol. Chem. 255, 10322—10330.Qi, P.X., De Stefaiio, D.L. & Wand, A.J. (1994a). Solution structure of horse heart ferrocytochrome c determined by high-resolution NMR and restrained simulated annealing.Biochemistry. 33, 6408—6417.Qi, P.X., Urbauer, J.L., Fuentes, E.J., Leopold, M.F. & Wand, A.J. (1994b). Structural waterin oxidized and reduced horse heart cytochrome c. Nat. Struct. Biol. 1, 378—382.Qin, L., Rodgers, K.K. & Sugar, S.G. (1991). Electron transfer between cytochrome b5 surfacemutants and cytochrorne c. Mo!. Cryst. Liq. Cryst. 194, 311—316.BIBLIOGRAPHY 200Rafferty, S.P., Pearce, L.L., Barker, P.D., Guillemette, J.G., Kay, C.M., Smith, M. & Mauk,A.G. (1990). Electrochemical, kinetic, and CD consequences of mutations at position-82of yeast iso-1-cytochrome c. Biochemistry. 29, 9365—9369.Ramdas, L., Sherman, F. & Nail, B.T. (1986). Guanidine hydrochloride induced equilibriumunfolding of mutant forms of iso-1-cytochrome c with replacement of proline-71. Biochemistry. 25, 6952—6958.Rees, D.C. (1985). Experimental evaluation of the effective dielectric constant of proteins. J.Mol. Biol. 141, 323—326.Reid, K.S.C., Lindley, P.F. & Thornton, J.M. (1985). Sulphur-aromatic interactions in proteins.FEBS Lett. 190, 209—213.Rieder, R. & Bosshard, H.R. (1980). Comparison of the binding sites on cytochrome c forcytochrome c oxidase, cytochrome be1 and cytochrome c1. Differential acetylation of lysylresidues in free and complexed cytochrome c. J. Biol. Chem. 255, 4732—4739.Roberts, V.A., Freeman, H.C., Olson, A.J., Tamer, J.A. & Getzoff, E.D. (1991). Electrostaticorientation of the electron-transfer complex between plastocyanin and cytochrome c. J.Biol. Chem. 266, 13431—13441.Roder, H., Elove, G.A. & Englander, S.W. (1988). Structural characterization of folding intermediates in cytochrome c by H-exchange labeffing and proton NMR. Nature (London).335, 700—704.Rossman, M.G., ed. (1972). The Molecular Replacement Method. Gordon & Breach, New York.Rossman, M.G., Leslie, A.G.W., Abdel-Meguid, S.S. & Tsukihara, T. (1979). Processing andpost-refinement of oscillation camera data. J. AppI. Crystallogr. 12, 570—581.Salemme, F.R. (1972). A free interface diffusion technique for the crystallization of proteins forX-ray crystallography. Arch. Biochem. Biophys. 151, 533—539.Salemme, F.R. (1976). An hypothetical structure for an intermolecular electron transfer complex of cytochromes c and b5. J. Mol. Biol. 102, 563—568.Sato, M., Yamamoto, M., Imada, K., Katsube, Y., Tanaka, N. & Higashi, T. (1992). A high-speed data-collection system for large-unit-cell crystals using an imaging plate as a detector.J. Appi. Crystallogr. 25, 348—357.Schejter, A., Aviram, I. & Goldkorn, T. (1982). The contribution of electrostatic factors to theoxidation-reduction potentials of c-type cytochromes. In Electron Transport and OxygenUtilization. (Ho, C., ed.), pp. 95—99. Elsevier, North, Amsterdam.Schoenborn, B.P. (1988). Solvent effect in protein crystals. A neutron diffraction analysis ofsolvent and ion density. J. Mol. Biol. 201, 741—749.BIBLIOGRAPHY 201Sherman, F. (1990). Studies of yeast cytochrome c: how and why they started and why theycontinued. Genetics. 125, 9—12.Singh, J. & Thornton, J.M. (1985). The interaction between phenylalanine rings in proteins.FEBSLett. 191, 1—6.Smalas, A.O. & Hordvik, A. (1993). Structure determination and refinement of benzamidineinhibited trypsin from the North Atlantic salmon (Salmo salar) at 1.82 A resolution. ActaCrystallogr. sect. D. 49, 318—330.Smith, M. (1986). Site-directed mutagenesis. Phil. Trans. Royal Soc. London ser. A. 317,295—304.Smith, M., Leung, D.W., Gillam, S., Astell, C.R., Montgomery, D.L. & Hall, B.D. (1979).Sequence of the gene for iso-i cytochrome c in Saccharomyces cerevisiae. Cell. 16, 753—761.Stellwagen, E. (1978). Haem exposure as the determinate of oxidation-reduction potential ofhaem proteins. Nature (London). 330, 86—88.Stemp, E.D.A. & Hoffman, B.M. (1993). Cytochrome c peroxidase binds two molecules ofcytochrome c: evidence for a low-affinity, electron-transfer-active site on cytoclirome cperoxidase. Biochemistry. 32, 10848—10865.Stenkamp, R.E. & Jensen, L.H. (1975). Effects of data set thresholds. Acta Crystallogr. sect.B. 31, 1507—1509.Stout, G.H. & Jensen, L.H. (1989). X-ray Structure Determination. A Practical Guide. SecondEdition. John Wiley & Sons, New York.Takano, T. & Dickerson, R.E. (i98ia). Conformation change of cytochroxne C: ferricytochrome crefinement at 1.8 A resolution and comparison with the ferrocytochrome structure. J. Mol.Biol. 153, 95—115.Takano, T. & Dickerson, R.E. (i981b). Conformation change of cytochrome C: ferrocytochrome c structure refined at 1.5 A resolution. J. Mol. Biol. 153, 79—94.Tanaka, N., Yamane, T., Tsukihara, T., Ashida, T. & Kakudo, M. (1975). The crystal structureof bonito ferrocytochrome c at 2.3 A resolution. J. Biochem. 77, 147—162.Taniguchi, V.T., Sailasuta-Scott, N., Anson, F.C. & Gray, H.B. (1980). Thermodynamics ofmetalloprotein electron transfer reactions. Pure Appl. Chem. 52, 2275—2281.Tegoni, M., White, S.A., Roussel, A., Mathews, F.S. & Cambfflau, C. (1993). A hypotheticalcomplex between crystalline flavocytochrome b2 and cytochrome c. Proteins: Struct. Funct.Genet. 16, 408—422.Thaller, C., Eichele, G., Weaver, L.H., Wilson, E., Karisson, R. & Jansonius, J.N. (1985). Seedenlargement and repeated seeding. Methods Enzyinol. 114, 132—135.BIBLIOGRAPHY 202Thurgood, A.G.P., Pielak, G.J., Cutler, R.L., Davies, A.M., Greenwood, C., Mauk, A.G.,Smith, M., Williamson, D.J. & Moore, G.R. (1991). Change in charge of an unvaried hemecontact residue does not cause a major change of conformation in cytochrome c. FEBSLett. 284, 173—177.Timkovich, R. (1979). Cytochrome c: the architecture of a protein-porphyrin complex. In ThePorphyrins. Volume VII. (Dolphin, D., ed.), pp. 241—294. Academic Press, New York.Tong, H., Berghuis, A.M., Chen, J., Luo, Y., Guss, J.M., Freeman, H.C. & Brayer, G.D. (1994).ASIR: an automatic procedure for determining solvent structure in protein crystallography.J. Appi. Crystallogr. 27, 421—426.Varadarajan, R. & Richards, F.M. (1992). Crystallographic structures of ribonuclease S variantswith nonpolar substitutions at position 13: packing and cavities. Biochemistry. 31, 12315—12327.Watkin, D. (1994). The control of difficult refinements. Acta Crystallogr. sect. A. 50, 411—437.Wendoloski, J.J., Matthew, J.B., Weber, P.C. & Salemme, F.R. (1987). Molecular dynamics ofa cytochrome c-cytochrome b5 electron transfer complex. Science. 238, 794—796.Whitford, D., Gao, Y., Peilak, G.J., Williams, R.J.P., McLendon, G.L. & Sherman, F. (1991).The role of the internal hydrogen bond network in first-order protein electron transfer between Saccharomyces cerevisiae iso-1-cytochrome c and bovine microsomal cytochrome b5.Eur. J. Biochem. 200, 359—367.Willie, A., McLean, lvi., Liu, R., Hilgen-Wiffis, S., Saunders, A.J., Pielak, G.J., Sligar, S.G.,Durham, B. & Millett, F. (1993). Intracomplex electron transfer between ruthenium-65—cytochrome b5 and position-82 variants of yeast iso-1-cytochrome c. Biochemistry. 32,7519—7525.Willie, A., Stayton, P.S., Sligar, S.G., Durham, B. & Millett, F. (1992). Genetic engineering ofredox donor sites: measurement of intracomplex electron transfer between ruthenium-65—cytoclirome b5 and cytochrome c. Biochemistry. 31, 7237—7242.Wilson, A.J.C. (1942). Determination of absolute scale from relative X-ray intensity data.Nature (London). 150, 151—152.Zhou, J.S. & Hoffman, B.M. (1993). Cytochrome c peroxidase simultaneously binds cytochrome c at two different sites with strikingly different reactivities: titrating a “substrate”with an enzyme. J. Am. Chem. Soc. 115, 11008—11009.Zhou, J.S. & Hoffman, B.M. (1994). Stern-Volmer in reverse: 2:1 stoichiometry of the cytochrome c-cytochrome c peroxidase complex. Science. 265, 1693—1696.Zoller, IVI.J. & Smith, M. (1983). Oligonucleotide-directed mutagenesis of DNA fragmentscloned into M13 vectors. Methods Enzymol. 100, 468—500.BIBLIOGRAPHY 203Zuniga, E.H. & Nail, B.T. (1983). Folding of yeast iso-i-AM cytochrome c. Biochemistry. 22,1430—1437.AddendumI. Publications in peer reviewed journals arising from the work described in thisthesis:1. The experiments described in Chapter 3, dealing with the determination of the structuresof the F82Y, L85A and F82Y/L85A mutants of yeast iso-i cytochrome c and the roles ofPhe82 and Leu85 at the interactive face of this protein, have been published in:Lo, T.P., Guillemette, J.G., Louie, G.V., Smith, M. & Brayer, G.D. (1995).Structural studies of the roles of residues 82 and 85 at the interactive face ofcytochrome c. Biochemistry 34, 163—171.2. The determination of the structures of the L85C, L8SF, L85M and L94S mutant proteins and an interpretation of the structural and functional effects of these amino acidsubstitutions as described in Chapter 4 have been published in:Lo, T.P., Murphy, M.E.P., Gufflemette, J.G., Smith, M. & Brayer, G.D. (1995).Replacements in a conserved leucine cluster in the hydrophobic heme pocket ofcytochrome c. Protein Science 4, 198—208.3. In conjunction with the structural studies described in Chapters 3 and 4, experiments involving the analysis of the kinetics of electron transfer between yeast iso-i cytochrome cmutant proteins and bovine cytochrome b5 were performed. My contribution to thesestudies was the modeling of two complexes formed between wild-type yeast iso-i cytochrome c and bovine cytochrome b5 by computational techniques. The interpretation ofthe observed electron transfer kinetics using these models of the protein-protein complexhas been described in:204Addendum205Guillemette, J.G., Barker, P.D., Eltis, L.D., Lo, T.P., Smith, M., Brayer,G.D. & Mauk, A.G. (1994). Analysis of the bimolecular reduction of ferncytochrome c by ferrocytochrome b5 through mutagenesis and molecular modeling. Biochirnie 76, 592—604.4. The studies described in Chapter 5 regarding the structural and functional analyses ofthe F821, F82L and F82M mutants of yeast iso-i cytochrome c have been prepared forpublication and submitted as:Lo, T.P. & Brayer, G.D. (1995). Structural analysis of the replacement of theinvariant phenylalanine 82 in cytochrome c by aliphatic residues. Biochemistry,submitted.5. The studies presented in Chapter 6 regarding the elucidation of the structures of yeast iso-1 cytochrome c proteins having multiple site mutations at Arg38, Asn52 and Phe82 andthe analysis of the effects of these mutations on reduction potential and protein stabilityhave been published in:Lo, T.P., Komar-Panicucci, S., Sherman, F., McLendon, G. & Brayer, G.D.(1995). The structural and functional effects of multiple mutations at distalsites in cytochrome c. Biochemistry, in press.II. Publications arising from collaborative work on related topics:These studies dealt with an analysis of cytochrome b5 and mapping antibody sites with syntheticpeptides. As such, these results deal with topics that fall outside the scope of the present workand therefore are not discussed herein.1. In this study, the structure of a triple mutant of cytochrome b5 was determined usingmolecular replacement methods (Section 2.4.3.2). These analyses revealed the manner inwhich the addition of a negative charge adjacent to the heme moiety results in a decreasein heme reduction potential for this protein.Addendum 206Funk, W.D., Lo, T.P., Mauk, M.R., Brayer, G.D., MacGiffivray, R.T.A. &Mauk, A.G. (1990). Mutagenic, electrochemical, and crystallographic investigation of the cytochrome b5 oxidation-reduction equilibrium: Involvement ofasparagine-57, serine-64, and heme propionate-7. Biochemistry 29, 5500—5508.2. This work determined the exposed surface areas of cytochrome c and the correlation ofthe surface exposure of short stretches of polypeptide chain with the antigenicity of thecorresponding peptides from a peptide library.Schrab, C., Twardek, A., Lo, T.P., Brayer, G.D. & Bosshard, H.R. (1993).Mapping antibody binding sites on cytochrome c with synthetic peptides: areresults representative of the antigenic structure of proteins? Protein Science 2,175—182.

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0088865/manifest

Comment

Related Items