Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Structural constraints for the folding, stability and function of cytochrome C Murphy, Michael E. P. 1993

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

Item Metadata


831-ubc_1993_spring_phd_murphy_michael.pdf [ 7.01MB ]
JSON: 831-1.0086377.json
JSON-LD: 831-1.0086377-ld.json
RDF/XML (Pretty): 831-1.0086377-rdf.xml
RDF/JSON: 831-1.0086377-rdf.json
Turtle: 831-1.0086377-turtle.txt
N-Triples: 831-1.0086377-rdf-ntriples.txt
Original Record: 831-1.0086377-source.json
Full Text

Full Text

STRUCTURAL CONSTRAINTS FOR THE FOLDING, STABILITYAND FUNCTION OF CYTOCHROME CByMichael Edward Patrick MurphyB.Sc., The University of Alberta, 1987A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFDOCTOR OF PHILOSOPHYinTHE FACULTY OF GRADUATE STUDIESTHE DEPARTMENT OF BIOCHEMISTRYWe accept this thesis as conformingto the required standardTHE UNIVERSITY OF BRITISH COLUMBIAJanuary 1993© Michael Edward Patrick Murphy, 1993In 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 refer-ence 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 BiochemistryThe University of British Columbia2075 Wesbrook PlaceVancouver, CanadaV6T 1Z3Date:AbstractThe relationship between structure, stability and function in yeast (Saccharomyces cerevisiae)cytochrome c was studied through the investigation of mutant proteins. Two isozymes of cyto-chrome c can be isolated from yeast, iso-1 and iso-2-cytochrome c. The structure of iso-l-cyto-chrome c had been previously determined in this laboratory using X-ray diffraction analyses. Inthe first part of this study, the structures of wild-type yeast iso-2-cytochrome c and a compositemutant protein (B-2036) composed of segments derived from the genes of both isozymes werestructurally characterized by X-ray analyses. The overall fold of the composite is similar tothe two native isozymes, however, specific intra-molecular atomic interactions are altered bythe presence of amino acid substitutions. These alterations are proposed to explain the loss ofthermodynamic stability of the composite relative to the wild-type proteins. In a second study,a-Loop A of iso-l-cytochrome c was replaced with the corresponding loop from cytochrome c2of Rhodospirillum rubrum to produce a mutant protein called RepA2 cytochrome c. Structuralanalysis of the RepA2 protein revealed that a-Loop A had folded in a conformation more similarto the original loop it replaced than the conformation of a-Loop A in cytochrome c2. Two sub-stitutions, Va120Phe and His26Asn, that result from the a-Loop A replacement were shown tocause structural perturbations. Phe20 was back substituted to Va120 to produce RepA2(Va120)cytochrome c. In vivo functional studies and reduction potential measurements of these S2-Loop A replacements demonstrate that Phe20 lowers the reduction potential by 19 mV at 25°C and diminishes cytochrome c function at higher temperatures. A third aspect of this workwas to investigate the role of Pro71. The structures of four partially functional Pro71 substitu-tions were determined: Pro7lAla, Pro7lIle, Pro7lSer and Pro71Val cytochromes c. Analysis ofthese structures revealed that loss of function could be explained by the disruption of residues82 and 83 which have been shown to be important in the binding and transfer of electrons toiicomplexed electron transfer partners. Larger side chain replacements, Pro71Val and Pro7lIle,result in additional disruptions at the site of mutation which is part of a highly conservedsegment of polypeptide chain. A comparison of all the mutant structures investigated in thiswork revealed that the overall fold of the protein was preserved at the expense of disruptingthe packing of the hydrophobic core and highly mobile surface loops. Amino acid substitutionsthat cannot be incorporated into the cytochrome c fold without structural perturbations occurin conformationally restricted regions. In addition, the displacement of groups by these re-placements indicates the limits present in specific regions in terms of conformational flexibility.The structural perturbations that result from the replacement of conformationally restrictedresidues have been shown to diminish both the reduction potential and in vivo function ofcytochrome c.iiiTable of ContentsAbstract^ iiTable of Contents^ ivList of Tables^ viiList of Figures ixList of Abbreviations and Structure Nomenclature^ xiAcknowledgements^ xiii1^Introduction1.1^The Mitochondria' Cytochromes c ^111.1.1 Structure ^ 21.1.2 Stability 101.1.3 Folding ^ 131.1.4 Function 141.2 The Yeast Cytochromes c ^ 171.3 Yeast Cytochrome c Variants 181.3.1 Pro71 Replacements ^ 221.3.2 Composite Proteins 231.3.3 Q-loop Replacements ^ 241.4 Thesis Objectives ^ 25iv2 General Experimental Methods2.1^Structure Determination ^27272.1.1^Crystal Growth and Characterization ^ 272.1.2^Data Collection ^ 292.1.3^Data Processing 312.1.4^Structure Solution and Refinement ^ 362.2 Structure Analysis ^ 372.2.1^Estimation of Coordinate Error ^ 372.2.2^Comparison of Structure Coordinates ^ 392.2.3^Comparison of Derived Structural Properties ^ 402.3 Electrochemical Properties ^ 412.3.1^Direct Electrochemistry 412.3.2^Derivation of Electrochemical Thermodynamic Properties ^ 423 Iso-2-Cytochrome c and a Iso-1/Iso-2 Composite Mutant Protein 463.1 Experimental Procedures ^ 463.2 Polypeptide Chain Conformations ^ 513.3 Comparison of Iso-2, B-2036 and Iso-1-Cytochromes c ^ 553.4 The Buried Cavity in B-2036 and Iso-l-Cytochromes c 613.5 Herne Geometry and Environment ^ 643.6 Conserved Water Molecules ^ 663.7 Comparison with Other Cytochromes c ^ 663.8 Electrochemical Properties ^ 673.9 Stability of Yeast Cytochromes c ^ 704 Replacements of Q-Loop A in Iso-1-Cytochrome c^ 724.1 Experimental Procedures ^  724.2 Results ^  774.2.1 Comparison of the Structures of RepA2, RepA2(Va120) and Iso-l-Cyto-chromes c ^  774.2.2 Electrochemistry of RepA2 and RepA2(Va120) Cytochromes c^ 824.3 Discussion ^  834.3.1 Comparison of the Wild Type Yeast Iso-1 and R. rubrum Structures^834.3.2 Structural Effects of 52-Loop A Replacements ^  864.3.3 In Vivo Functional Consequences of C2-Loop A Replacements ^ 884.3.4 The Relationship Between Herne Reduction Potential and TemperatureSensitivity ^  925 Replacements of an Invariant Proline in Iso-1-Cytochrome c^945.1 Experimental Procedures ^  945.2 Results^  965.2.1 The Environment of Pro71 in Yeast Iso-1-Cytochrome c ^ 965.2.2 Structural Differences of Pro71 Mutant Cytochromes c  1005.3 Discussion ^  1056 Summary^ 113Bibliography^ 118viList of Tables1.1 Primary sequence alignment for yeast iso-1, yeast iso-2, tuna, horse and ricecytochromes c ^  41.2 X-ray structure determinations of mitochondrial cytochromes c . ^ 51.3 Structural equivalence and primary sequence identity of yeast iso-1, tuna, horseand rice cytochromes c ^  61.4 Secondary structural elements in iso-l-cytochrome c ^  81.5 Stability of oxidized cytochromes c ^  121.6 Electrochemical properties of cytochromes c from different sources ^ 151.7 Electrochemical properties of yeast cytochrome c and mutants  161.8 Minimal unique residue positional boundaries of iso-2 polypeptide chain segmentsin each composite iso-1-cytochrome c class ^  232.9 The effect of neighboring background measurements on data quality ^ 352.10 Comparison of overall observed main chain r.m.s. deviations and the expecteddeviation errors ^  393.11 Data collection statistics of iso-2 and B-2036 cytochromes c ^ 473.12 Sequence alignment of iso-1, B-2036 and iso-2-cytochromes c  483.13 Final stereochemistry of iso-2 and B-2036 cytochromes c ^  493.14 Hydrogen bond analysis of iso-2, B-2036 and iso-l-cytochromes c ^ 523.15 Secondary structural elements in iso-2-cytochrome c ^  553.16 Heme solvent accessibility of mitochondrial cytochromes c  643.17 Heme conformation and ligand geometry in yeast iso-2 and B-2036 cytochromes c 653.18 Conserved water molecules in yeast iso-2, B-2036 and iso-l-cytochromes c . . .^67vii3.19 Electrochemical properties of the yeast iso-2, B-2036 and iso-l-cytochromes c . . 694.20 Sequence alignment of the yeast iso-1. RepA2 and RepA2(Va120) cytochromes cwith cytochrome c2 from Rhodospirillum rubrum ^  734.21 Data collection statistics of RepA2 and RepA2(Va120) cytochromes c ^ 744.22 Final stereochemistry of RepA2 and RepA2(Va120) cytochromes c ^ 754.23 Electrochemical properties of the yeast iso-1, RepA2, and RepA2(Va120) cyto-chromes c ^  834.24 Hydrogen bond interactions of Q-Loop A in yeast iso-1, RepA2, RepA2(Va120)cytochromes c and R. rubrum cytochrome c2 ^  855.25 Data collection statistics for Pro71 mutant cytochromes c ^  955.26 Final refinement statistics for Pro71 mutant cytochromes c  975.27 Final stereochemistry for Pro71 mutant cytochromes c ^  975.28 Positional deviations of groups in the vicinity of residue 71  1025.29 Average thermal factors of groups in the region of residue 71 ^ 1065.30 Functional and stability studies of Pro71 mutant cytochromes c  1086.31 Amino acid additions and replacements in the eight cytochrome c structuresstudied in this work ^  113viiiList of Figures1.1 The cytochrome c fold ^  71.2 A space-filled representation of yeast iso-l-cytochrome c  ^91.3 A stereographic representation of iso-1-cytochrome c ^  102.4 X-ray diffraction data collection geometry ^  302.5 The 0 and 8 dependence of background measurements ^  342.6 A sample SDBase information retrieval ^  422.7 Schematic of the direct electrochemistry experiment used to measure reductionpotentials ^  432.8 A sample cyclic voltammogram of iso-2-cytochrome c^  443.9 A Luzzati plot of the iso-2 and B-2036 cytochrome c structure determinations . . 503.10 Ramachandran plots of the iso-2 and B-2036 cytochrome c structures ^ 543.11 Stereo a-carbon backbone of iso-2-cytochrome c ^  563.12 Stereographic representation of iso-2-cytochrome c  563.13 The average main chain positional deviations between iso-2, B-2036 and iso-1-cytochromes c ^  573.14 The region near G1y37 in iso-2, B-2036 and iso-l-cytochromes c ^ 583.15 The region about residue 26 in iso-2, B-2036 and iso-l-cytochromes c ^ 593.16 The average main chain thermal factors of iso-2, B-2036 and iso-l-cytochromes c 603.17 The region about residues 20 and 102 in iso-2, B-2036 and iso-l-cytochromes c . 623.18 The buried cavity in B-2036 and iso-l-cytochromes c   633.19 The overall average deviations of main chain atoms of mitochondrial cytochromes c 68ix3.20 The midpoint reduction potential versus temperature of iso-2, B-2036 and iso-l-cytochromes c ^4.21 A Luzzati plot of the RepA2 and RepA2(Va120) cytochrome c structure deter-minations ^69764.22 A space-filled representation of a-Loop A of yeast iso-l-cytochrome c ^ 784.23 A stereo representation of S2-Loop A of yeast iso-l-cytochrome c ^ 794.24 The average main chain positional deviations between RepA2, RepA2(Va120)and iso-l-cytochromes c ^ 804.25 The region near residue 26 in RepA2. RepA2(Va120) and iso-l-cytochromes c • 814.26 The packing of residue 20 in RepA2, RepA2(Va120) and iso-l-cytochromes c . • 814.27 The midpoint reduction potential of the RepA2. RepA2(Va120) and iso-l-cyto-chromes c as a function of temperature. ^ 824.28 A stereographic plot of a-Loop A from yeast iso-1-cytochrome c superimposedon that of cytochrome c2 from R. rubrum ^ 844.29 Growth curves of experimental and control yeast strains in liquid lactate media 894.30 Spectra of intact yeast cells of experimental and control strains at —196 °C^. . 904.31 Average main chain thermal factors of 1-Loop A in RepA2, RepA2(Va120) andiso-l-cytochromes c ^ 915.32 A Luzzati plot of the Pro71 mutant cytochrome c structure determinations . 985.33 The region near Pro71 in yeast iso-l-cytochrome c ^ 995.34 Positional and thermal factor difference matrices for Pro71 mutant cytochromes c 1015.35 Structural changes associated with replacing Pro71 with a smaller side chain . . 1035.36 Structural changes associated with replacing Pro71 with a larger side chain . . . 1045.37 Space-filled representations of wild-type and Pro71 mutant cytochromes c . . . . 1096.38 A plot of the overall average deviation of main and side chain atoms of thestructures examined in this work from iso-1-cytochrome c ^  115List of Abbreviations and Structure NomenclatureA^hydrogen bond acceptor atomA.S.^saturated ammonium sulfateAngstrom unit; 1 A= 10 -1 ° mB crystallographic thermal factor0midpoint reduction potential under standard conditionsF^the Faraday constant: 23.06 cal mol -1 mV -1calculated structure factor magnitudeF0^observed structure factor magnitudeD hydrogen bond donor atomDTT^dithiothreitolI observed reflection intensityK absolute scale factorND^not determinedNMR^nuclear magnetic resonance spectroscopyPDB^Brookhaven Protein Data BankSHE^standard hydrogen electrodeTml^e-N-trimethyl lysineUV^ultraviolet lighta, b, c^crystallographic unit cell axesh, k, 1^Miller indicese.u.^entropy unit; 1 e.u. = cal mol —1 K -1 = 4.184 J mo1 -1 K -1r.m.s.^root-mean-squaredxiA^wavelength of the radiation used in the X-ray diffraction experiment(CuK,=1.5418 A)ionic strengthQ, X, q^Eulerian angular position of the crystal9 Bragg diffraction angle(Tx^ estimated standard deviation of the quantity x1. The conventions of the IUPAC-IUB Combined Commissions on Biochemical Nomencla-ture are followed for both three letter and one letter abbreviations for amino acids [J.Biol. Chem. 241, 527-533 (1966); J. Biol. Chem. 243, 3557-3559 (1968)]; for designat-ing atoms and for describing the conformational torsional angles of the polypeptide chain[J. Biol. Chem. 245, 6489-6497 (1970)].2. Designations for the atoms of the protoporphyrin IX heme group are according to theBrookhaven National Laboratory Protein Data Bank (Bernstein et al., 1977).3. The numbering scheme used to designate amino acid residues along the polypeptide chainof cytochrome c are according to those of the vertebrate cytochromes c. To conform tothis scheme, the amino terminal extensions of yeast iso-1 and iso-2-cytochromes c arenumbered —5 to —1 and —9 to —1, respectively.4. The designation, yeast, used herein, refers to only the Saccharomyces cerevisiae species.All other yeast species will be referred to by their taxonomic names.xi iAcknowledgementsThe many people who assisted in this work deserve my deepest appreciation. The proteinsamples were generously supplied by Profs. Barry Nall, Jacque Fetrow and Fred Sherman.Crystals and conditions for crystal growth were provided by Connie Leung, Theresa Yang andTheresa Hii. One of the Pro7lIle data sets was collected in the laboratory of Prof. LouisDelbaere. The electrochemical studies were performed in the laboratory of Prof. Grant Maukwith the assistance of Paul Barker and Steve Rafferty. Past and present members of thelaboratory of Prof. Gary Brayer contributed stimulating discussions and some of the computersoftware used in the determination and analysis of the structures. In addition, Yaoguang Luocollected the RepA2 data set with equipment provided by Prof. Daniel Yang. Over the courseof these studies, Albert Berghuis became a close friend and collaborator and critically reviewedmuch of the thesis. I would like to thank David Burk for giving me a place to stay in thefinal months of my graduate studies. I acknowledge Gary Brayer, my research advisor, for hisguidance, encouragement and the freedom to pursue my own projects. Also, I recognize theother members of my committee, Profs. Bob Molday and Steve Withers, for suggesting that theelectrochemistry experiments would compliment the structural studies. The Medical ResearchCouncil of Canada is recognized for financial support in the form of a Studentship. I appreciatethe support of my parents and family throughout my education. I am grateful to my spouse,Gail, for constant encouragement and support as well as for reviewing the thesis.xmChapter 1IntroductionA central question in the study of proteins is how a given amino acid sequence specifies theprecise three-dimensional structure of a protein and how structure determines a specific bio-chemical function. One approach to this question is to examine the structure and function ofvariants of a well characterized protein. The comparison of available mutants with wild-typeforms of a protein allows better definition of those amino acid sequences that result in a folded,stable and functional protein. One system which readily lends itself to these types of studies isthat of cytochrome c.1.1 The Mitochondrial Cytochromes cMitochondrial cytochrome c is a small soluble electron transport protein that is readily iso-lated from eukaryotic organisms and has, as a result, been extensively studied (Dickerson &Timkovich, 1975; Ferguson-Miller et al., 1979; Timkovich, 1979; Poulos & Finzel, 1984; Math-ews, 1985; Pettigrew & Moore, 1987; Moore & Pettigrew, 1990; Sherman, 1990). Cytochrome cis differentiated from other cytochromes by the nature of attachment of the heme prostheticgroup. In this protein, the heme is attached to the polypeptide chain via two thioether linkagesto Cysl4 and Cys17. The central iron atom is coordinated to four pyrrole nitrogen atoms of theheme group in a square planar arrangement. Two amino acids, His18 and Met80, provide theremaining axial ligands to form an octahedral coordination geometry. The polypeptide chaincomponent, composed of between 102 and 112 amino acids, is synthesized in the cytoplasm andis transported into the inter-membrane space of the mitochondria by heme lyase. The hemelyase also catalyzes the covalent attachment of the heme group (Dumont et al., 1988).1Chapter 1. Introduction^ 2As a component of the mitochondria' electron transport chain, cytochrome c transfers singleelectrons from Complex III to Complex IV (Pettigrew & Moore, 1987). Also known as cyto-chrome c reductase or cytochromes b ci, Complex III is a large membrane bound enzymecomplex involved in energy transduction. Cytochrome c oxidase (Complex IV) is also an integralmembrane protein and functions to conduct electrons to the terminal electron acceptor, 02.Cytochrome c oxidase is also a point of energy transduction in oxidative phosphorylation. Unlikethese two membrane bound complexes, cytochrome c is an isopotential electron carrier in thatthe electrochemical potential of the electron is conserved during transport and no biologicalenergy transduction occurs (Salemme, 1977).Historically, the cytochromes of the electron transport chain have been defined by theirunique spectroscopic properties which arise from changes in the aromatic character of the hemechromophore. The visible spectrum of cytochrome c is characterized in the reduced state bythree intense absorption Soret bands at 554 nm (a), 524 nm (13) and 416 nm (7). The a and 13bands give this protein a red color. In the oxidized state, the y Soret band shifts to 410 nm anda new absorption maxima appears at 695 nm. This band has been proposed to be linked to theiron sulfur bond of the Met80 ligand (Schechter & Saludjian, 1967). These spectral propertiesare used to assess the structural integrity of the protein.1.1.1 StructureThe first amino acid sequence determined for a cytochrome c was that from horse (Margoliashet al., 1961). Subsequently, the amino acid sequences of more than 90 cytochromes c from a widevariety of eukaryotic sources have been determined by chemical sequencing of their polypeptidechains. Several compilations and analyses of these sequences have been published (DickersonTimkovich, 1975; Dayhoff & Barker, 1976; Moore & Pettigrew, 1990). Variability plots asa function of residue position have been computed based on these compilations (Louie et al.,1988a; Moore & Pettigrew, 1990). Typically, variability is defined as the number of amino acidsat a given position divided by the frequency of the most common amino acid at that positionChapter 1. Introduction^ 3(Wu Kabat, 1970).Another approach used to present the range of amino acid residues at a given position hasbeen to create a diagram of the sequence set where the complete sequence of yeast cytochrome cis presented along with the other observed amino acids at each position (Hampsey et al., 1986;Hampsey et al., 1988). Inspection of a compilation of 94 sequences (Moore & Pettigrew, 1990)reveals that 27 out of 102 common amino acid positions are invariant and a further 16 are foundin 90 of the sequences and are highly conserved. The sequences of yeast iso-1, yeast iso-2, tuna,horse and rice cytochromes c which represent a broad range of species are presented in Table 1.1.One region of the cytochrome c sequence, residues 70 to 80, is exceptionally conserved withonly a total of 4 alternative amino acids observed in 4 out of 94 sequences in this region.In addition to the large number of amino acid sequences determined for mitochondrialcytochromes c, the detailed folds of five cytochromes c have been resolved by X-ray diffractiontechniques (see reviews: Salemme (1977), Matthews (1985) and Brayer & Murphy (1993)). Thefirst mitochondrial cytochrome c structure solved was that of oxidized horse heart (Dickersonet al., 1971). Later, the same group improved the resolution of the cytochrome c fold by deter-mining the structure of tuna cytochrome c in both the oxidized and reduced state (TakanoDickerson, 1981a,b). In the meantime, a low resolution X-ray study of bonito cytochrome c wasundertaken (Tanaka et al., 1975; Matsuura et al., 1979). Since then the structures of rice cyto-chrome c (Ochi et al., 1983) and yeast iso-l-cytochrome c (Louie et al., 1988a; Louie & Brayer,1990) have been reported. The yeast structure has also been refined in the oxidized state usingisomorphous crystals (Berghuis & Brayer, 1992). Recently, a higher resolution redeterminationof the horse structure has also been completed (Bushnell et al., 1990). Experimental detailsfor all the structural studies completed to date are listed in Table 1.2. Of these, the yeast iso-1structure has been completed to the highest resolution.As has been observed in other systems, a high degree of sequence homology is exhibitedin the conservation of the three-dimensional structures of the cytochromes c (Chothia Lesk,1986). The overall main chain structural and sequence homologies of the four available highChapter 1. Introduction^ 4Table 1.1: Primary sequence alignment for yeast iso-1, yeast iso-2, tuna, horse and rice cyto-chromes c-9Iso-1^- - - - TEFKA167 SAKK G ATL10.17KTR C L20K G GPH K V GQCHT VEIso-2 AKESTGFKP G SAKK G ATL F KTR C Q QCHT TEE GPN K V GTuna G DVAK K K T F VQK C A QCHT VEN G GKH K V GHorse ^ G DVEK G KKI F VQK C A QCHT VEK GKH K T GRice^- ASFSEAPP G NPKA G EKI F KTK C A QCHT V D K G AGH K Q G30 40^ 50 60Iso-1 PNL H G I FGR H S G Q A E G YS Y TD AN IKKNVL DENNM S EIso-2 PNL H G I FGR HS Q V K G YS Y TD AN INKNVK DEDSMS ETuna PNL G L FGR KT G QAE G YS Y TD AN KSKGIV NNDTL MEHorse PNL H G L FGR KT G QAP FT Y TD AN KNKGIT KEET L MERice PNL N G L FGR QS TTP G YS Y ST AN KDMAV I EENTL YD70 80 90 100Iso-1 YL T NP J KYIPGTKM A F G G L KK EKD R N DL IT Y LKK A C E -Iso-2 YL T NP J KYIPGTKM A F A G L KK EKD R N DL IT Y M T K A A K -Tuna YL E NPKKYIPGTKM I F A G I KK KGE R Q DL VA Y LKS A T S -Horse YL E NPKKYIPGTKM I F A G I KK KTE R E DL IA Y LKK A T N ERice YL L NP J KYIPGTKM V F P G L KJ PQE R A DL IS Y LKE A T S -The primary sequences of yeast iso-1 (Smith et al., 1979), yeast iso-2 (Montgomery et al., 1980),tuna (Kreil, 1965), horse (Margoliash et al., 1961) and rice (Mori & 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 primary sequence oftuna cytochrome c. The amidation states of residues 52 and 54 of the rice protein have beenmodified as suggested by Moore and Pettigrew (1990). Note the single-letter code J is used todenote E-N-trimethyl lysine. Those amino acid residues identical in all five protein sequencesare enclosed by boxes (residues denoted J and K are considered equivalent).Chapter 1. Introduction^ 5Table 1.2: X-ray structure determinations of mitochondrial cytochromes c.Cytochrome c Crystallization Conditions Space Unit Cell Resolution ReferencestGroup (A) (A)Yeast Iso-1(reduced)92% A.S., 0.1 M phosphate,pH 6.2, 40 mM DTTP43 212 a=b=36.46,c=137.861.2 (a,b)Yeast Iso-1(oxidized)92% A.S., 0.1 M phosphate,pH 6.2, 30 mM NaNO3P43212 a=b=36.47,c=137.241.9 (c)Tuna(reduced)85% A.S., excess ascorbate,pH 7.5P21212 a=34.44b=87.10,1.5 (d)c=37.33Tuna(oxidized)50% A.S., 15% NaNO3,1.0 M ammonium phosphate,pH 7.0P43 a=b=74.42,c=36.301.8 (e,f)Horse(oxidized)94% A.S., 0.1 M phosphate,pH 7.5P43 a=b=53.58,c=41.831.9 (g)Rice(oxidized)3.6 M A.S., pH 6.0 P61 a=b=43.78,c=110.51.5 (h)tReferences: a, Sherwood & Brayer (1985); b, Louie et al. (1988); c, Berghuis & Brayer (1992);d, Takano & Dickerson (1981a); e, Swanson et al. (1977); f, Takano & Dickerson (1981b); g,Bushnell et al. (1990); h, Ochi et al. (1983).resolution cytochrome c structures are given in Table 1.3. The sequence identity varies from57% between the yeast iso-1 and rice cytochromes c to 83% for the horse and tuna proteins.The r.m.s. main chain deviations range from 0.51 to 0.61 A.The high structural homology between the cytochromes c of known three-dimensional struc-ture allows the use of the yeast iso-l-cytochrome c structure to represent the general cyto-chrome c fold. This fold is schematically illustrated in Figure 1.1. In total, this protein canbe separated into four classes of secondary structure elements (Table 1.4). The a-helix andtwo 13-turn classes are well known (for review: Richardson, 1981); however, Q-loops have onlyrelatively recently been recognized by Leszczynski & Rose (1986). An C2-loop consists of 6to 16 residues which form a loop such that the terminal C c, atoms are less than 10 A apart.Chapter 1. Introduction^ 6Table 1.3: Structural equivalence and primary sequence identity of yeast iso-1, tuna, horse andrice cytochromes cYeast iso-1^Tuna^Horse^Rice(reduced)^(reduced)^(oxidized) (oxidized)(108)^(103)^(104)^(111)Yeast iso-1 63 (61%) 60 (58%) 62 (57%)Tuna 0.40 (0.54) 85 (83%) 61 (59%)Horse 0.48 (0.54) 0.43 (0.51) 64 (62%)Rice 0.40 (0.51) 0.45 (0.61) 0.49 (0.60) —The total number of amino acids and the oxidation state of each protein is given below itsname in the heading. The upper triangular half of the matrix contains the number of identicalresidues in common between each pair of proteins aligned in Table 1.1. The percentage identityof the smaller protein is given in parentheses. The common 412 main chain atoms (residues 1to 103) of each cytochrorne c were overlayed in a pairwise fashion by a least-squares method.The lower half of the matrix contains the average main chain distance deviations. The valuesin parentheses are the r.m.s. deviations of main chain atoms for each pair.These loops do not contain a-helices or /3-sheets as defined by Kabsch & Sander (1983). Anexamination of 270 Q-loops in 67 proteins reveals that these secondary structure elements formglobular structures as compact as complete proteins (Leszczynski & Rose, 1986).Of the 108 amino acid residues of iso-l-cytochrome c, 52 (48%) are found to be in ana-helical conformation (Louie & Brayer, 1990). A further 51 amino acids (47%) have beendefined as belonging to a-loops (Fetrow et al., 1989). The loops defined include part of Helix II(residues 49 to 54), Helix IV, all of the type II 0-turns and the 7-turn (Table 1.4). Helices IIand IV which are shorter and more poorly formed than Helices I, III and V, are not consideredhelices according to the primary Q-loop definition. After accounting for overlapping definitions,13 amino acid residues (12%) are not assigned to a secondary structure element. These includethe 6 amino terminal residues (residues —5 to 1, Table 1.1), residues 56 to 59 that form partof a distorted /3-sheet (Louie & Brayer, 1990), a short segment (residues 85 and 86) that linksChapter 1. Introduction^ 7Figure 1.1: The polypeptide chain of yeast iso-l-cytochrome c is represented as a ribbon. Theheme group, the two thioether linkages to Cys14 and Cys17, as well as the two axial hemeligands to His18 and Met80, are also depicted.Chapter 1. Introduction^ 8Table 1.4: Secondary structural elements in iso-l-cytochrome cElement class Structural element Residues involveda-helixC2-loop0-turn7-turnHelix I^ 2-14Helix II 49-55Helix III 60-70Helix IV^70-75Helix V 87-102Loop A 18-32Loop B^ 34-43Loop C 40-54Loop D 70-84Turn 1 (type I)^14-17Turn 2 (type II) 21-24Turn 3 (type II)^32-35Turn 4 (type II) 35-38Turn 5 (type II)t^43-46Turn 6 (type II) 75-7827-29The a-helices, type II /3-turns and 7-turn are as defined in Louie & Brayer (1990). The St-loopsare as defined by Fetrow et al. (1989).t Mediated through a water molecule.two secondary structure elements and the carboxy terminal residue, 103.As shown in Figures 1.2 and 1.3, the polypeptide chain of cytochrome c is folded around theheme group such that only one edge of the porphyrin ring of the heme is exposed to solvent. Toa large extent, the porphyrin ring is surrounded by hydrophobic residues and forms an integralpart of the hydrophobic core of the protein (Louie et al., 1988a). The total heme solventexposure is less than 10% of its total surface area (Bushnell et al., 1990). The two negativelycharged heme propionate groups are also buried in the protein matrix and are completelyinaccessible to solvent. These charged groups form numerous hydrogen bonds as well as a saltbridge to the polypeptide component (Dickerson et al., 1971). These studies also show that theheme group is not planar, but is distorted into a saddle shape (Louie & Brayer, 1990).Chapter 1. Introduction^ 9Figure 1.2: A space-filled representation of yeast iso-l-cytochrome c. The atoms of the centralheme group are shown as colored black spheres. All other atoms are colored white.Chapter 1. Introduction^ 10Figure 1.3: A stereographic representation of iso-1-cytochrome c with side chains drawn in thinlines and the polypeptide backbone and heme group in thick lines. Every fifth residue and thetwo termini are labeled according to the sequence alignment in Table 1.1.Four water molecules form an essential part of the structure of yeast cytochrome c. The firsttwo, Wat121 and Wat168, are hydrogen bonded to the heme propionate A and the guanidiniumgroup of Arg38. In the tuna and horse structures, the side chain of Arg38 interacts directlywith the propionate group and Wat168 is not present. A third distinct conformation of theguanidinium group in rice cytochrome c results in a hydrogen bond network that includes twowater molecules in the same position as Wat121 and Wat168 of the yeast iso-1 protein. Theother two essential waters in the yeast structure are Wat110 and Wat166. Wat110 is locatednear the His18 ligand and Wat166 is hydrogen bonded to three residues near the Met80 ligand:Asn52, Tyr67 and Thr78. Wat166 is conserved in all four known cytochrome c structures andhas been suggested to be important in the mechanism of electron transfer (Takano & Dickerson,1981b; Berghuis & Brayer, 1992).1.1.2 StabilityThe thermodynamic stability of cytochrome c has been quantitatively characterized by re-versible unfolding equilibrium studies using the denaturant guanidine hydrochloride (Knapp &Chapter 1. Introduction^ 11Pace, 1974). The presence of the folded versus unfolded protein can be monitored by trypto-phan fluoresence which is quenched by the heme in the folded state or alternatively by opticalrotation at 220 nm. The guanidine hydrochloride data may be extrapolated to zero denaturantto give the free energy of unfolding under standard conditions (Schellman, 1978). A studyof the stability of cytochrome c from horse, donkey, dog, cow, rabbit, chicken and tuna at25 °C, pH 7, revealed a similar unfolding free energy of 7 ± 1 kcal/mol (McLendon & Smith,1978). The results from this study for horse (7.38 kcal/mol) agreed well with earlier experi-ments which measured a value of 7.27 kcal/mol under similar conditions (Knapp & Pace, 1974).In contrast, the free energy of unfolding of four cytochromes c from three yeast species, genusSaccharomyces and Candida, is found to be 3 to 5 kcal/rnol less stable than horse cytochrome cunder similar experimental conditions (Table 1.5). The number of buried hydrophobic groupshas been suggested to account for the difference in stability between the yeast iso-1 and iso-2proteins and horse cytochrome c (Nall & Landers, 1981). Global thermodynamic stability hasbeen suggested to be involved in the regulation of protein turnover in some organisms (Knapp& Pace, 1974). The lower stability of the yeast proteins may be correlated with the induction oftheir expression whereas organisms such as mammals requiring a constant level of cytochrome ccorrespondingly produce a globally more stable protein.The structure of ferricytochrome c is known to be pH dependent from UV visible spectro-scopic studies (Theorell Akesson, 1941). The native form of cytochrome c is observed ina pH range of about 5 to 8 depending on the species and oxidation state. However, oxidizedcytochrome c undergoes an alkaline transition that is characterized by the loss of the 695 nmabsorbance band, having an apparent pKa of approximately 9. The pK a values of the alkalinetransition for 5 different wild-type cytochromes c and some mutant forms of the yeast iso-2protein are presented in Table 1.5. The structure of the alkaline form has remained elusive,however, the loss of the 695 nm absorbance band has been associated with the loss of the Met80ligand (Schechter & Saludjian, 1967). Further spectroscopic studies have suggested that a de-protonated lysine &amino group is the replacement ligand (Davis et al., 1974; Brautigan et al.,Chapter 1. Introduction^ 12Table 1.5: Stability of oxidized cytochromes cCytochrome c^ AG of unfolding (kcal/mol)^pKatkt Reference20 °C, 0.1 M sodium phosphatepH 7.0 — 7.2 pH 6.03.6 8.5 (a,b)3.8 8.5 (c,d)8.5 9.2 (e,e)5.9 8.9 (e,e)4.6 8.6 (e,e)2.6 (a)1.9 (a)1.9 (a)5.7 (h)3.8 (h)7.0 (h)4.7 (h)3.0 6.6 (d)2.6 6.7 (f,g)A. Nativeyeast (S. cerevisiae) iso-l-MStyeast (S. cerevisiae) iso-2horseCandida kruseiSaccharomyces oviformisB. Yeast iso-1 mutantsPro71Val-MStPro71Thr-MStPro71Ile-MStAsn52A1a/Cys102AlaAsn52Gly/Cys102AlaAsn52I1e/Cys102AlaCys102AlaC. Yeast iso-2 mutantsPro71ThrPro76GlyReferences: a, Ramdas et al. (1986); b, Pearce et al. (1989); c, Osterhout et al. (1985); d, Whiteet al. (1987); e, Saigo (1981); f, Wood et al. (1988b); g, Nall et al. (1989); h, Hickey et al. (1991)tpKa ik is the apparent pl-C a of the alkaline transition.t Residue Cys102 of these cytochromes c was modified with methyl methanethiosulfonate toprevent dimerization.Chapter 1. Introduction^ 131977). Acetimidylation of lysine residues has suggested that Lys72 or Lys79 may be the hemeiron ligand in the alkaline form (Wallace, 1984), however, mutation of these residues to alaninesdid not prevent alkaline isomerization (S. Inglis, unpublished results). Inspection of the X-raystructure of cytochrome c reveals that a drastic conformational change would be required toposition an c-amino group of a lysine residue near the heme iron. A NMR study of the alkalineform provides evidence for the presence of at least two conformations of cytochrome c at highpH (Hong & Dixon, 1989). The kinetics of the transition involve a rapid deprotonation stepfollowed by a slow conformational change (Davis et al., 1974). Folding kinetic studies of iso-2-cytochrome c at high pH have shown that the alkaline form is preceded by an intermediate withspectroscopic characteristics similar to those of the native protein (Nall, 1986). These resultssuggest that the alkaline form does possess some of the native structure observed at neutralpH.1.1.3 FoldingThe kinetic properties of unfolding and refolding of horse cytochrome c by rapid changes inguanidine hydrochloride concentration were initially characterized by monitoring absorbance(Ikai et al., 1973), and later by following tryptophan fluoresence quenching (Tsong, 1976).The structures of folding intermediates have also been characterized by two-dimensional NMRspectroscopy of pulse labeled proteins (Roder et al., 1988). These studies suggest that foldingis initiated by the formation of Helices I and V (Table 1.4) followed by contact between thesetwo helices in a manner similar to the native structure.Detailed studies of the kinetic properties of folding of yeast iso-2-cytochrome c have revealedthe presence of four kinetic phases. The fastest of these phases (T3) with a time constant of theorder of 1 ms is poorly characterized (Nall Si Landers, 1981). Two of the phases err and T2)are observed to yield a functional product as determined by absorbance at 695 nm and ascorbicacid reduction. The differences in rate between the fast (T2) and slow (Tf) absorbance detectedphases appears to arise from differences in proline isomerization in the unfolded protein. TheChapter I. Introduction^ 14product of another phase OP detected by tryptophan fluoresence may either be a folding deadend or an intermediate on a different path to the native form (Nall, 1983). Further studiesof the temperature dependence of phases rf and rt yield activation enthalpy values in therange expected for proline isomerization. Furthermore, the amplitudes of phases Ti and rt areindependent of initial pH, suggesting the involvement of proline isomerization rather than hemeligation in these kinetic phases (Osterhout et al., 1985; Osterhout & Nall, 1985).The presence of a free sulfhydryl group in yeast iso-1-cytochrome c has interfered with theaccurate characterization of folding intermediates of this protein. Cys102 of the iso-1 protein canbe blocked by treatment with iodoacetamide (Zuniga & Nall, 1983) to prevent dimer formation.An examination of the folding kinetics of iodoacetamide blocked iso-l-cytochrome c revealedsimilar kinetic phases as observed for the iso-2 protein (Zuniga & Nall, 1983).1.1.4 FunctionThe functional requirements of cytochrome c are to accept and donate single electrons withbiological redox partners. In this role, cytochrome c must be able to bind to a given redoxpartner and provide a low energy path for the electron to travel between the redox centerslocated within each protein. This last requirement also implies that the midpoint potential ofcytochrome c must be at a level that favors electron transfer thermodynamically between theredox partners (Salemme, 1972).The midpoint reduction potential is a physical property of cytochrome c that has beenmeasured both spectroscopically (Kreishman et al., 1978) and by direct electrochemistry (Arm-strong et al., 1988). Measurement of the reduction potential of several cytochromes c from dif-ferent sources reveals that this physical property is highly conserved (Table 1.6). By measuringthe temperature dependence of the midpoint potential, the enthalpic and entropic contributionsto the free energy of reduction may be determined. These thermodynamic parameters are alsoconserved in the five species studied as described in Table 1.6. Further studies have shown thatChapter I. Introduction^ 15Table 1.6: Electrochemical properties of cytochromes c from different sourcesCytochrome c Em°(mV vs SHE)0AH(kcal/mol)0ASe.u.Yeast iso-1 261 -17.1 -37Horse heart 261 -16.8 -36Tuna heart 265 -17.5 -39Turkey heart 260 -16.8 -36Candida species 260 -16.8 -36Experimental conditions were 25 °C, pH 7.0, p, = 0.01 and SHE reference. The data was takenfrom Table 3 of Margalit Schejter (1973a).the reduction potential is sensitive to ionic strength and to the presence of ions that bind pref-erentially to one oxidation state such as chloride and phosphate (Margalit Schejter, 1973a;Margalit Schejter, 1973b). Recently, direct electrochemistry has been used to measure themidpoint reduction potential of yeast iso-l-cytochrome c (Rafferty et al., 1990). The higherpotential given in Table 1.7 is due to the difference in ionic strength, the presence of differentions and the presence of small amounts of the alkaline form.An initial explanation for the high midpoint potential of cytochrome c relative to modelcompounds with the same axial ligands was given by Kassner (1972,1973). Using a simpleelectrostatic model of an ion in a sphere of constant dielectric, Kassner showed that hydrophobicgroups surrounding the heme displaced solvent molecules which stabilize the positive charge onthe heme in the oxidized state. Destabilization of this positive charge increases the reductionpotential of cytochrome c. A related structural parameter, heme solvent exposure, has also beenproposed to affect the midpoint reduction potential (Stellwagen, 1978). Schejter et al. (1982)extended the work of Kassner by separating the free energy of reduction into an electrostaticand non-electrostatic component and showed that the electrostatic term can model changesin molecular surface charge. Another electrostatic model that uses dipoles to model both theprotein and the surrounding solvent has been used to demonstrate their individual roles inChapter 1. Introduction^ 16Table 1.7: Electrochemical properties of yeast cytochrome c and mutants Cytochrome c E.m(mV)AH0(kcal/mol)ASc0 yt(e.u.)AG°(kcal/mol)ReferencetIso-1 290 ± 2 -14.0 ± 0.2 -9.1 ± 0.4 -6.7 ± 0.1 (a)0 0Decreased AH and AScytF82L 286 ± 2 -14.6 ± 0.2 -11.6 ± 0.7 -6.5 ± 0.1 (a)L94S 280 ± 2 -14.9 ± 0.1 -12.6 ± 0.4 -6.5 ± 0.1 (b)W59F 274 ± 2 -14.2 -10.7 -6.3 ± 0.1 (c)F82I 273 ± 2 -14.6 ± 0.1 -12.3 ± 0.3 -6.3 ± 0.1 (a)Y48F 262 ± 2 -14.2 -12.0 -6.0 ± 0.1 (c)0 0Increased AH and AScytF82Y 280 ± 2 -13.6 ± 0.1 -8.3 ± 0.3 -6.5 ± 0.1 (a)F82A 260 ± 2 -13.0 ± 0.2 -8.1 ± 0.5 -6.0 ± 0.1 (a)N52A 257 ± 2 -12.9 ± 0.1 -8.0 ± 0.3 -5.9 ± 0.1 (d)T78G 249 ± 2 -12.3 ± 0.2 -6.3 ± 0.2 -5.7 ± 0.1 (d)F82G 247 ± 2 -12.7 ± 0.2 -8.0 ± 0.4 -5.7 ± 0.1 (a)N52IY67F 234 ± 2 -12.6 ± 0.1 -12.5 ± 0.6 -5.4 ± 0.1 (e)Y67F 234 ± 2 -12.3 -7.5 -5.4 ± 0.1 (d)N52I 232 ± 2 -12.4 ± 0.1 -8.2 ± 0.4 -5.4 ± 0.1 (e)0 0Increased AH and decreased AScytF82S 255 ± 2 -13.8 ± 0.2 -11.0 ± 0.5 -5.9 ± 0.1 (a)I75M 249 ± 2 -13.8 ± 0.2 -11.7 ± 0.3 -5.7 ± 0.1 (d)Y48FW59F 240 ± 2 -13.0 -9.4 -5.5 (c)R38A 236 ± 2 -13.5 -11.7 -5.4 (c)Experimental conditions were 25 °C, pH 6.0, it 0.1 M and SHE reference.tReferences: a, Rafferty et al. (1990); b, Lo et al. (1993); c, Tong et al. (1993); d, Rafferty,1992; e, Berghuis (1993) .Chapter 1. Introduction^ 17modulating reduction potential (Churg^Warshel, 1983; Langen et al., 1992). The chargedheme propionates and their interaction with the solvent and polypeptide chain has also beenproposed to be important in stabilizing the heme iron in the oxidized state (Moore, 1983).Taniguchi et al. (1980) showed that the entropy of reduction of cytochrome c is more negativethan that measured for inorganic iron complexes. This decreased entropy was proposed to bedue to an ordering of the polypeptide chain and the surrounding solvent upon reduction. Theentropy and enthalpy of cytochrome c reduction has been measured in differing halide solutions(Kreishman et al., 1978) with the temperature dependence of the reduction potential observed tobe biphasic at 42 °C in 0.1 M NaC1 solutions. These results indicate that solvent structure mayplay an important role in determining the entropy of reduction of cytochrome c. Furthermore,the volume of cytochrome c has been shown to decrease upon reduction (Trewhella et al., 1988)which may result in a reorganization of the surface solvent structure (Margalit Schejter,1973a). Changes in the hydrogen bonding of surface solvent water molecules have been shownto be associated with compensating changes in enthalpy and entropy (Lumry Rajender,1970).1.2 The Yeast Cytochromes cUnlike other eukaryotic organisms, two isozymes of cytochrome c (iso-1 and iso-2) have beenisolated from the yeast Saccharomyces cerevisiae. These two proteins are believed to haveequivalent physiological roles (Mattoon & Sherman, 1966; Dethmers et al., 1979). The amountof each of these isozymes produced is controlled by differential regulation of the transcriptionof their respective genes in response to the presence of glucose, heme, and molecular oxygen(Laz et al., 1984). Of the cytochrome c in yeast, the relative amount of the iso-2 form presentis dependent on environmental conditions, varying from 5 to 25% (Sels et al., 1965). Theadditional stability of apoiso-2-cytochrome c to cellular degradation relative to apoiso-1 providesanother mechanism for a greater proportion of holoiso-2-cytochrome c under some conditions(Dumont et al., 1990). However, the rate of conjugation of ubiquitin to iso-2-cytochrome c inChapter I. Introduction^ 18vitro in a rabbit reticulocyte extract occurs at a rate five fold greater than for the iso-1 protein(Sokolik & Cohen, 1991; Sokolik & Cohen, 1992), suggesting that the iso-2 form may also bedegraded at a different rate in yeast cells.Early mutational mapping led to the discovery of several nuclear genes required for pro-duction of yeast cytochrome c (Sherman et al., 1965). The structural genes for iso-1 andiso-2-cytochromes c were identified and named CYC1 and CYC7, respectively. The CYC3gene has been identified as coding for heme lyase (Dumont et al., 1987) which is required forheme attachment and mitochondrial import of both yeast cytochromes c. Other genes werefound to be involved in heme synthesis. Recently, another mitochondrial protein involved incytochrome c import was identified as the product of the CYC2 gene (Sherman, 1990). TheCYC1 gene was cloned and sequenced by several groups (Montgomery et al., 1978; Smithet al., 1979). Subsequently, the iso-2-cytochrome c gene (Montgomery et al., 1980) was clonedand sequenced which confirmed that the iso-2-cytochrome c amino acid sequence (Dayhoff &Barker, 1976) differs from that of iso-1 by 17 substitutions and a four residue extension at theamino terminus (Table 1.1). The iso-1 and iso-2-cytochromes c have an overall 84% amino acidsequence identity.1.3 Yeast Cytochrorne c VariantsThe first cytochrome c mutants were produced by treatment of yeast cells with UV radiation orchemical mutagens followed by the isolation of mutants using genetic techniques (Sherman et al.,1965). Non-functional cytochrome c mutants were screened by growth on media containingchlorolactate which inhibits those yeast cells capable of metabolizing lactate (Sherman et al.,1975). The amount of cytochrome c in intact yeast cells was measured by low temperaturespectroscopy at the 550 nm a band (Sherman & Slonimski, 1964). The functional capacity andtemperature sensitivity of the mutants were measured by growth of yeast cells on lactate mediaat varying temperatures (Schweingruber et al., 1977; Schellman, 1978). These mutants werefurther analyzed by genetic deletion mapping (Sherman et al.. 1975) and amino acid or DNAChapter 1. Introduction^ 19sequencing (Hampsey et al., 1986).Using these techniques, 44 cycl non-functional missense mutants have been identified andmapped as 30 different substitutions at 16 sites (Schweingruber et al., 1979; Hampsey et al.,1988). No mutations have been isolated that abolish function without reducing the quantity ofprotein in intact cells. These results suggest that all of the mutant proteins are temperaturesensitive (Hampsey et al., 1988). Mutations at three residues, Cysl4, Cys17 and His18, resultedin a complete lack of protein in intact cells. These residues form either thioether or iron ligandlinkages to the heme and are essential either for heme attachment or mitochondrial import.Mutations at the other heme ligand, Met80, also resulted in poor production of holo-enzyme.Other mutations resulting in less than half the normal in vivo level of cytochrome c are Gly6Asp,Leu9Pro, Gly29Asp and Gly29Asn. Gly6 is part of a section of Helix I which packs againstHelix V (Table 1.4) such that a larger side chain cannot be tolerated at this position. Leu9 isalso part of Helix I and a proline at this position prevents the main chain at position 9 fromforming the hydrogen bonds necessary for adopting an a-helical conformation.Other non-functional mutations did in fact lead to the production of greater than 50% ofthe normal level of cytochrome c, however, the functional capacity of these cytochromes c wereless than 5% of the wild-type protein as judged by growth on lactate. These include mutationsof the hydrophobic residues that line the heme pocket: Leu32, Trp59, Tyr67, Leu68, Leu94 andLeu98. Mutation of either Pro30 or Pro7l to a leucine also resulted in a folded but completelynon-functional protein. The His33Pro mutant is also non-functional, however, the amount ofHis33Pro cytochrome c observed in intact yeast cells was more than twice that of the Leu9Promutant. Residue 33 appears to be less critical for folding, but directly disrupts the electrontransfer function of cytochrome c.In addition to non-functional mutants, a large number of partially to fully functional mu-tants, as measured by growth on lactate media, have been isolated (Sherman et al., 1974).For example, 85 amino acid replacements within the first 9 amino acid residues, as well as thecomplete deletion of this region, appears to be tolerated in yeast iso-1-cytochrome c (HampseyChapter 1. Introduction^ 20et al., 1988). Some of the other functional replacements are of highly conserved or invariantresidues, such as Tyr98Leu, Pro76Leu and Lys72Glu. These observations suggest that invari-ance does not necessarily imply an absolute functional requirement for these residues. However,all of the non-functional mutations with the exception of mutations to proline residues, occurat highly conserved or invariant residues (Hampsey et al., 1988).The semi-synthesis technique has been used to create mutants of horse cytochrome c (Offord,1987). In this technique, cytochrome c is cleaved into fragments with cyanogen bromide andlimited tryptic digestion. The fragments are separated, modified or replaced with synthetic sub-stitutes, reassembled and the modified protein is then characterized. A comprehensive overviewof semi-synthetic studies of cytochrome c is presented in Louie (1990). The variants producedby some semi-synthetic methods were difficult to accurately characterize due to the loose struc-ture resulting from the non-covalent interaction of multiple peptide fragments. However, whena single site cyanogen bromide cleavage between residues 65 and 66 in the horse protein wasemployed, the homoserine lactone fragment (residues 1 to 65) could be religated to the carboxyterminal fragment (residues 66 to 104) to produce a molecule with properties similar to thewild-type protein (Corradin Harbury, 1974). Using this technique, chemically synthesizedpeptides may be substituted for the carboxy terminal fragment allowing the introduction of anyamino acid, including non-natural amino acids, into the protein fold at positions 66 through 104.This technique has been exploited to replace the Met80 ligand with other functional groups toproduce cytochromes c with unusual electrochemical properties (Wallace & Clark-Lewis, 1992).Recently, the mutations Met64Leu and Ser65Met have been made in the yeast protein to makeits sequence in this region similar to that of the horse (Table 1.1) and thereby allow use of thesemi-synthetic technique in conjunction with site-directed mutagenesis (Wallace et al., 1991).The cloning and expression of the genes of the two isozymes of yeast cytochrome c andthe development of site-directed mutagenesis methods provides a means of producing specificmutants at any desired residue position (Inglis et al., 1991; Kunkel et al., 1987; Zoller & Smith,1983). Initial structure-function studies using the technique of site-directed mutagenesis wereChapter I. Introduction^ 21aimed at Phe82 which has been proposed to be critical in the mechanism of electron transfer(Pielak et al., 1985). Site-directed mutagenesis was required to produce mutants of this invariantresidue since no mutants had been isolated by classical genetics. Replacements at this positiongive proteins with significantly altered electrochemical properties (Rafferty et al., 1990) andinter-protein electron transfer kinetics (Liang et al., 1988). Two Phe82 mutants, Phe82Serand Phe82Gly cytochromes c. have been characterized structurally by X-ray crystallography(Louie et al., 1988b; Louie & Brayer, 1989). The Phe82Ser mutant structure revealed thata solvent channel had been created which increased the solvent exposed surface area of theheme. The increased heme solvent accessibility was suggested to account for the 45 mV dropin reduction potential of this mutant (Louie et al., 1988b). Structural analysis of the secondmutant, Phe82Gly, revealed a refolding of the polypeptide chain in the region of residue 82 suchthat polar peptide groups packed against the heme. As in the Phe82Ser mutant, the increasedheme polarity is proposed to account for a drop in reduction potential (Louie & Brayer, 1989).Mutants isolated by classical genetic techniques have also been used as a guide for generatingsite-directed mutants. For example, the Asn52Ile mutant of yeast cytochrome c was initiallydiscovered as a second site revertant to a number of temperature sensitive mutants (Hickeyet al., 1988). Site-directed mutagenesis was then used to create the Asn52Ile mutant of iso-l-cytochrome c and to characterize its structure and thermostability (Hickey et al., 1991). Thethermostability of Asn52Ile was measured with the additional mutation Cys102Ala to preventdimerization and a loss of reversibility upon unfolding of yeast iso-1-cytochrome c. Structuralstudies revealed that the greatly enhanced thermostability of the Asn52Ile mutation (Table 1.5)may be due to the displacement of a buried water molecule, Wat166 (Hickey et al., 1991). Thenormally resident residue Asn52 forms a hydrogen bond with Wat166 which has been proposedto be important in mediating oxidation state dependent conformational changes (TakanoDickerson, 1981b; Berghuis & Brayer, 1992).Several mutants of yeast cytochrome c have been analyzed to determine how the aminoacid sequence influences the thermodynamics of reduction (Table 1.7). These mutants couldChapter 1. Introduction^ 22potentially be divided into four categories based on whether each of the enthalpic and entropiccontributions to the free energy of the reaction has been increased or decreased. However, onlymutants in three categories are presented in Table 1.7. None of the mutants studied thus farpossesses a reduction potential greater than that of native yeast iso-l-cytochrome c at 25 °C,and thus, no mutations have increased the enthalpy while decreasing the entropy of reduction.Mutants with lower reduction potentials are observed in the other three categories. The mutantwith the lowest reduction potential is Asn52Ile cytochrome c (Table 1.7). The decrease inreduction potential of 58 mV resulted from a large increase in the enthalpy and a small increasein entropy of reduction. A similar 54 mV decrease in reduction potential was observed as aresult of the Arg38Ala mutation, however, the thermodynamic contributions differed. In thismutant, the enthalpy of reduction increased while the reaction entropy decreased. The mutantsin the third category, those in which both the enthalpy and entropy of reduction decreased,result in more modest changes in reduction potential. The Tyr48Phe cytochrome c representsthe lowest potential, a decrease of 28 mV.1.3.1 Pro71 ReplacementsGenetic fine structure mapping of four non-functional cycl mutants revealed a common lesionsite (Sherman et al., 1975). DNA sequencing of these mutants indicates that proline wasreplaced by a leucine at position 71 (Ernst et al., 1985). These mutants were further treatedwith mutagens and partially functional revertants were selected by growth on lactate at 30 °C(Sherman et al., 1983). DNA and protein sequence analyses have identified four differentrevertants: Pro7lIle, Pro7lSer, Pro7lThr and Pro71Val cytochromes c (Ernst et al., 1985).From growth rates of yeast cells on lactate media, the function of these mutant cytochromes chave been estimated to be: Pro7lVal, 90%; Pro7lThr, 60%; Pro7lSer, 30%; and Pro7lIle, 20%.The lack of growth of the original Pro7lLeu mutant suggests that this mutant is completely non-functional. Other revertants that could have been generated by single base pair substitutionsincluded: phenylalanine, tyrosine, glutamate, glutamine, lysine and arginine. Since none ofChapter I. Introduction^ 23these amino acid residues were observed out of the 29 analyzed, they are also suggested to benon-functional replacements. These results suggest that structural or functional requirementsconstrain the amino acid at position 71 to small nonpolar residues (Ernst et al., 1985).1.3.2 Composite ProteinsA composite protein is defined as a protein with segments of polypeptide chain derived frommore than one wild-type protein. Sequence analysis of a series of revertants of a nonsensemutation of iso-l-cytochrome c at position 71 revealed that some reversions resulted from anon-allelic recombination with the CY C7 gene for iso-2-cytochrome c (Ernst et al., 1981). Theresulting composite cytochromes c are composed of segments of the iso-1 and iso-2 proteins.In each of 14 composite iso-l-cytochromes c identified by peptide mapping and amino acidsequencing, a single iso-2 polypeptide chain segment representing 13% to 61% of the totalamino acid sequence was present (Ernst et al., 1982). These composites were grouped into fiveclasses based on their amino acid sequences. The minimum unique amino acid segment derivedfrom the iso-2 gene for each class of composite is detailed in Table 1.8. The composites of ClassV were most frequently observed (9 of the 14 composites isolated). This class also contains theshortest segment of the iso-2 protein, 10 residues.Table 1.8: Minimal unique residue positional boundaries of iso-2 polypeptide chain segmentsin each composite iso-l-cytochrome c classClass Positional Boundaries Yeast StrainsI —1, 26 B-2125II 15, 63 B-1155, B-2036III 26, 63 B-1585IV 26, 83 B-2080V 54, 63 B-1904, B-2037, B-2038. B-2039, B-2047B-2203, B-2204, B-2205, B-2200Chapter 1. Introduction^ 24The composite cytochromes c from three yeast strains (B-1904, B-2080, B-2036) represent-ing Classes II, IV and V have been further characterized (Dumont et al., 1990). All of thecomposites are less stable to thermal denaturation than either of the native isozymes fromwhich they are derived. The thermal transition temperatures of iso-1 and iso-2-cytochromes care 52.3 and 54.2 °C, respectively. The transition temperatures of the composites range from48.9 to 49.3 °C. To achieve reversible conditions, the composites and iso-1-cytochromes c weretreated with methyl methanethiosulfonate to specifically block the sulfhydryl group of Cys102to prevent dimerization in the unfolded state. The thermodynamics of unfolding of these com-posite proteins, as well as the two native isozymes of yeast, are being further investigated bydifferential scanning micro-calorimetry (J. Liggins and B. Nall, unpublished results). Like wild-type iso-2-cytochrome c, all composite proteins with the iso-2 sequence from positions 54 to63 are more stable to cellular degradation in their apo forms than the iso-1 protein (Dumontet al., 1990).1.3.3 Q-loop Replacementsa-loop replacements are a special class of composite protein where the replaced segment ofpolypeptide chain is an Q-loop. As described previously, four Q-loops comprising 47% of thetotal structure have been defined for the yeast iso-l-cytochrome c (Table 1.4). Oligonucleotide-directed mutagenesis has been used to delete and replace segments of these four Q-loops tostudy their role in cytochrome c biosynthesis and activity (Fetrow et al., 1989). Deletionswithin Q-Loop A. residues 22 to 28, and Q-loop D, residues 74 to 78, resulted in a completedeficiency of cytochrome c in intact yeast cells (Fetrow et al., 1989). In contrast, the deletionsof either residues 37 to 40 in a-loop B or 47 to 54 in Q-loop C produced partially functionalcytochrome c molecules.In the loop swap studies, Q-Loop A was replaced with the corresponding loops of fourcytochromes c from tuna, Rhodospirillum rubrum c2, Pseudomonas denitrificans c550, andPseudomonas aeruginosa c551, respectively (Fetrow et al., 1989). The tuna, R. rubrum andChapter 1. Introduction^ 25P. denitrificans loop replacements are of the same residue length or longer and result in cyto-chromes c that are partially functional. However, when Q-Loop A is replaced with the equivalentbut shorter loop from P. aeruginosa cytochrome c551, the composite protein did not supportgrowth on liquid lactate media although adequate levels of protein are present in intact cells(Fetrow et al., 1989). A fifth a-Loop A replacement was constructed using a larger loop un-related in amino acid sequence which had been derived from porcine pancreatic esterase. Thisa-Loop A replacement was shown to lead to a partially functional cytochrome c (Fetrow et al.,1989).1.4 Thesis ObjectivesYeast cytochrome c is a biochemically and genetically well characterized protein ideally suitedfor study of the relationship between protein structure, stability and function. The genes ofthe two isozymes of cytochrome c from yeast have been cloned, sequenced and are amenableto site-directed mutagenesis. Also, a large number of random functional and non-functionalmutants have been isolated by genetic techniques. Furthermore, the heme prosthetic group is astrong chromophore which allows for spectroscopic studies of this protein. In addition, a highresolution structure of yeast iso-1-cytochrome c is available as well as those of horse, tuna, andrice. In this work, three classes of mutants were selected for detailed investigation in order togain insight into the functional and folding properties of cytochrome c: composite proteins,1-loop replacements and single site mutations of the invariant proline at position 71.The B-2036 cytochrome c studied is a Class II yeast composite cytochrome c (Table 1.8).The wild-type yeast iso-2-cytochrome c was also investigated to allow comparison of this com-posite with both parent wild-type proteins. Both, the three-dimensional structures and electro-chemical properties were determined for iso-2 and B-2036 cytochromes c. The results of thesestudies were combined with the data available for iso-1-cytochrome c allowing for a detailedcomparison of the structure and function of all these proteins. Further analyses were initiated toChapter 1. Introduction^ 26compare these structures to the available in vivo and in vitro stability studies so that the rela-tionship between the structure, stability and function could be better understood. The B-2036and iso-2-cytochrome c proteins used for these studies were provided by Dr. B. Nall (Universityof Texas). The electrochemical properties were measured in the laboratory of Dr. G. Mauk(University of British Columbia).The second class of mutants investigated were a-Loop A replacements of yeast iso-1-cyto-chrome c. The first, RepA2 cytochrome c, contains the equivalent loop from Rhodospirillumrubrum cytochrome c2. The aim of this study was to investigate the functional role of Q-loopsin cytochrome c and to evaluate if these loops are interchangeable between related proteins. Amutation of this loop replacement, RepA2(Va120) cytochrome c, was investigated to determinethe role of residue 20 in stabilizing the loop-protein interface and in the overall function ofyeast cytochrome c. The protein samples of both a-Loop A replacements were provided byDr. J. Fetrow (State University of New York at Albany). The electrochemical properties ofthese two mutant proteins were measured in the laboratory of Dr. G. Mauk (University ofBritish Columbia).The third class of mutants was of an invariant proline residue at position 71. The structuresof four mutants, Pro7lAla, Pro7lIle, Pro7lSer and Pro7lVal cytochromes c were elucidated.The mutants were investigated to determine the role of this proline residue in directing thefold of cytochrome c and to examine the mechanism by which these amino acid replacementsalter the stability and function of cytochrome c. Protein for these studies was provided byDr. F. Sherman (University of Rochester).Chapter 2General Experimental MethodsAn overview of the techniques used to determine structures by X-ray diffraction and to measuremidpoint reduction potentials are described in this chapter. The details of each structuraldetermination are given in the following chapters.2.1 Structure Determination2.1.1 Crystal Growth and CharacterizationThe protein used in crystallization experiments was purified by established procedures (Shermanet al., 1968; Nall & Landers, 1981) and was provided by collaborators. All of the crystals used inthese studies were produced using the technique of hair seeding. The original seed crystals werederived from crystals of yeast iso-l-cytochrome c (Sherwood & Brayer, 1985). This technique,as it is applied to the hanging drop method (McPherson, 1982), is described in Leung et al.(1989). For the current work, this hair seeding technique was also applied to the liquid-liquiddiffusion capillary crystallization method which was first described by Salemme (1972). Inthis method, a capillary tube (about 1.5 x 100 mm) is sealed at one end with a Bunsenburner. Solutions are then added to the tube with a syringe and forced to the bottom by handcentrifugationl. In a typical yeast cytochrome c crystallization experiment, 30 pl of saturated(NH4)•SO4 buffered to pH 6.0 to 6.5 with 0.1 M sodium phosphate and 10 to 80 mM reducingagent (dithiothreitol or sodium dithionite) is placed at the bottom of the capillary tube. Asecond less dense solution (10 id.) containing 80 mg/ml cytochrome c protein, 60% saturated(NH4)2SO4 with the same buffer and reducing agent is placed near the tube opening. Seed'The capillary tube is placed into a disposable centrifuge tube attached to a string and swung.27Chapter 2. General Experimental Methods^ 28crystals are then introduced by drawing a hair through a solution containing a crushed crystalof a wild-type or variant yeast cytochrome c and then by dipping the hair into the proteinsolution layered at the top of the capillary tube. This solution is then forced down the tube byhand centrifugation. To prevent mixing during centrifugation, the solution of highest densitymust always be at the bottom of the tube. Restricted by the small diameter of the capillarytube the two layered solutions mix over a period of days. Crystals of cytochrome c first appearafter 12 to 36 hours and reach maximal size in 10 days.Crystals used in diffraction analyses were first transferred into a freshly prepared solutionof the same buffer to maintain the reduced state of the heme group. This solution consistedof a slightly higher concentration of ammonium sulfate than the calculated final concentrationused in the crystallization experiment. Dithiothreitol was substituted for Na2S2O4 in mountingsolutions used for crystals grown with this reducing agent since Na2S2O4 forms a SO2 radical bya monomerization reaction which may damage the crystalline protein (Ferguson-Miller et al.,1979). The crystals were then mounted for X-ray diffraction analyses in thin wall glass capillarieswith diameters of 0.7 to 1.0 mm depending on the size of the crystal.Precession photography was used to initially characterize cytochrome c crystals (Buerger,1964). In this way, the unit cell dimensions of iso-2-cytochrome c were determined to bea = b = 36.43 A, c = 137.84 A and to belong to either of the two enantiomorphic space groupsP43212 or P41212 (Leung et al., 1989). Earlier studies had shown crystals of yeast iso-l-cyto-chrome c to be of the space group P43212 and to have unit cell parameters of a = b = 36.46 A,c = 136.86 A (Louie et al., 1988a). Based on a comparison of precession photographs andunit cell parameters of the two isozymic crystals, it was assumed that these crystals wereisomorphous. This assumption was subsequently confirmed by successful refinement of theiso-2 structure. The mutant crystals used were assumed to be isomorphous based on unitcell dimensions and diffraction symmetry. The assumption of isomorphism was verified by thesuccessful refinement of these mutant structures. The volume per unit mass, V ii,, for spaceChapter 2. General Experimental Methods^ 29group P43212 assuming a single molecule per asymmetric unit, was computed from:abcV =-In 8M(2.1)where a, b and c are the unit cell lengths and M is molecular weight of cytochrome c. Thefactor of 8 is the number of molecules in the unit cell. The V ii, for these yeast cytochrome ccrystals is approximately 1.7 A 3 mol/g, indicating a solvent content of about 31% (Matthews,1968).2.1.2 Data CollectionData sets were collected for six of the eight forms of cytochrome c studied using an Enraf NoniusCAD4-F11 diffractometer. The CuK, radiation used was Ni filtered and generated from an X-ray tube operated at 26 mA and 40 kV. The initial orientation of each crystal was found by smallangle precession photographs. Crystals were then transferred to the diffractometer and an initialorientation matrix was derived from the angular position of one reflection and the alignment ofa crystallographic axis parallel to the instrument cb axis (see Figure 2.4 for a description of datacollection geometry). The orientation matrix and the cell parameters were then refined againsta minimum of 16 strong centered reflections. The maximal resolution and peak scan widthwere selected based on the diffraction quality and size of the crystal. During data collection,each reflection was measured by a continuous SZ peak scan of 0.4 to 0.6° at 0.55°/min at anambient temperature of 15 °C. In addition, a selection of 3 to 18 reflections were measuredevery 2 to 8 hours to monitor crystal decay and slippage. At least one cb independent reflection(x = 90°) was measured at 5° intervals in for use in absorption correction. Approximately700 reflections were collected in a 24 hour period and a complete data set to 1.9 A was collectedin about 12 days. Each data set was complete to the chosen resolution and was obtained froma single crystal.Only small crystals with volumes less than 0.01 mrn 3 were obtained for the RepA2 andPro7lIle mutants. These crystals were too small to collect diffractometer data with a standardsealed tube X-ray source. However, the combination of greater flux X-ray sources and areaDiffractedbeamDirectbeamChapter 2. General Experimental Methods^ 30Figure 2.4: A schematic of the three circles which determine the crystal position inthree-dimensional Eulerian space is presented. The crystal position is determined by progressiverotation about three angles, fl, x and 0. The diffraction angle, 20, is also indicated.detector technology allowed the collection of good quality data sets from these crystals. Thetwo types of area detectors used to collect data were the Enraf Nonius FAST system andthe Rigaku R-AXIS II imaging plate system. The oscillation method was employed in bothinstruments (Arndt & Wonacott, 1977). In this method, the crystal is oscillated through anangle 0 and the diffracted X-rays are collected on a two-dimensional detector in a fashionsimilar to using X-ray film. The rotating anode X-ray sources were operated at between 70 to100 mA and 50 to 60 kV. In both systems, the crystal orientation and diffraction spot indexingwas found by analysis of a minimum of two still frames collected at widely different 0 angles.Hundreds of frames are typically collected on the FAST system with a small oscillation angle of0.10 to 0.15° and a X-ray exposure time of 20 to 40 seconds per frame. The goniometer of theFAST system allows repositioning of the crystal to collect a complete data set. An oscillationChapter 2. General Experimental Methods^ 31angle of 0.7 to 1.5° and an exposure of 10 to 20 minutes per frame is used with the R-AXISdetector. The crystal must be manually repositioned and a new orientation found to collect themissing data cusp.2.1.3 Data ProcessingA diffractometer reflection scan was divided and stored in three parts. The first 1/6 of a scanwas taken to be one background intensity measurement, the middle 2/3 of the scan as thepeak intensity (4), and the last 1/6 of the scan as the second background measurement. Inthe simplest case, the two background measurements are assumed to be representative of thebackground radiation of the peak and are used as the background correction. The expectedstatistical error in these measurements is the square root of the number of counts measured.For weak reflections with relatively few counts, this error is large. This is especially true forbackground measurements which are only 1/3 of the total scan.This statistical error may be decreased if one assumes that the background measurements ofneighboring reflections are also representative of the current peak background. In the presentstudy, this approach was taken and a list of candidate background measurements was madefrom reflections found within a sphere of reciprocal space surrounding the current reflection.An average background was then computed. The list of candidates (Ii) was then compared tothe standard deviation of the average, and those measurements that differed by more than twodeviations from the average were discarded. The final average background intensity (4) andthe variance based on counting statistics (0-1) was calculated as follows:T  =^2 4at) = 77-1-(2.2)(2.3)where n is the number of background measurements included in the average after discardingthose more than two standard deviations from the initial average. Note that the variancedecreases as the number of background measurements increases, however, the assumption thatChapter 2. General Experimental Methods^ 32the neighboring background measurements are representative of the peak is weakened. If onlythe background measurements made adjacent to the peak are included, as in the individualbackground case, the variance is a special case of Equations 2.2 and 2.3 with n = 2. Thisindividual variance was compared to the variance of the averaging calculation to determine ifbackground averaging was used instead of the original individual backgrounds. Finally, thecorrected peak intensity (io) and error estimate (crio ) was calculated by:/0 = /p — 4/b (2.4)0.10 420.b2 (2.5)The factor of four arises since only a fourth of the time was spent to measure a single backgroundcompared to the peak.For non-spherical crystals, the average path of the incident and diffracted X-ray beams variesfor each reflection measured. This average path difference results in a change in intensity due toabsorption. This effect is strongly dependent on the crystal morphology and the data collectiongeometry. The method used to correct for absorption was that of North et al. (1968). Theintensity of 0 independent reflections were measured at 5° rotation intervals of 0. The resulting0 curve was then normalized and interpolated by a Fourier function to give an absorption curve(P4). This curve was then used to estimate the relative absorption (A) of a general reflection:1A = —2 [PA(cb -I- 0 cos(x)) PA(0 — 0 cos(X))1 (2.6)The intensity and sigma of the reflection were corrected by dividing by the absorption factor.The above method does not account for the 0 dependence of absorbance. A linear inter-polation on 0 between absorption curves computed from several / independent reflections wasused to account for the 0 dependence. The absorption curves were normalized by assumingthat total absorption over 0 is constant.The decay of diffracted intensities due to radiation damage was monitored periodicallythroughout data collection by a set of intensity control reflections. To account for the 8 de-pendence of crystal decay, intensity controls were collected in groups of similar resolution. AChapter 2. General Experimental Methods^ 33polynomial was derived from a group of intensity controls of similar resolution by a least-squaresfit. The B dependence was then calculated by normalizing the polynomials to the first intensitycontrol group measured.The Lorentz effect arises from the different angles through which a reflection passes throughthe Ewald sphere during a scan. The equation used for both the Lorentz and polarizationcorrection for diffractometer geometry is:sin(20) LP = (1 — sin2 (26)/2)The intensity and sigma were multiplied by this factor.Symmetry related reflections and repeat measurements were merged by averaging to improvethe accuracy of the data set. The estimated standard deviation in the averaged intensity (as)was computed by:'22=1 2= E n3/2The final intensity was converted to a structure factor amplitude by taking the square root.The final estimate of the error in the structure factor F was calculated by:QI^0-IaF = — = (2.9)A computer program, ICP 2 , was constructed to implement the data correction techniques out-lined.In Figure 2.5, the 0 and 0 dependence of the background measurements obtained from adata set collected from the Pro7lSer mutant are plotted. The greatest dependence is on 0,however there is also a correlation with 0, particularly at low resolution. In the past, individualbackgrounds or a 0 dependent background function have been used to correct protein data setsfor background radiation (Wyckoff, 1985). In these methods, the statistical quality of the dataset is improved, however, systematic errors were introduced due to the dependence of the back-ground radiation level on angle 0 (Figure 2.5). The background averaging method described in'Intensity Correction Program(2.7)(2.8)80.0Cz060.00L-0)0a) 40.0-vo20.0 t^I^[^1^I^1^[^1^[0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0 80.0 90.0 100.0PhiChapter 2. General Experimental Methods^ 34Figure 2.5: A plot of average background measurements as a function of the angles 0 and 0for data collected from the Pro7lSer mutant. Each curve represents one value of sin(e)/A:^, 0.073; A, 0.128; 0, 0.180; 7, 0.227; 0, 0.264. The background radiation level is strongly 0dependent and 0 dependent at lower 0 values.this work provides the needed improvement in the statistical precision of the background mea-surements and minimizes the introduction of systematic errors since only neighboring reflectionsare included in the averaging. In Table 2.9, the merging R-factor, the percentage of negativeintensity measurements after background correction and the percentage of intensities that aregreater than their estimated standard deviations based on counting statistics are given for datacorrections including a varying number of neighboring background measurements. Clearly, in-cluding a modest number of neighboring background measurements significantly decreases themerging R-factor and reduces the number of negative intensities, indicating an improvement inthe statistical consistency of the data.Chapter 2. General Experimental Methods^ 35Table 2.9: The effect of neighboring background measurements on data qualityPro7lSer Leu85PhetNo. of Pf-merget T^n T^T n•-merge .10 < 0 -10^°V())4 (%) (%) (%) (%) (%) (%)2 43.7 20.0 54.1 34.7 24.7 45.56 34.1 18.8 58.8 28.8 23.3 50.310 31.8 18.0 60.4 27.4 23.1 52.114 30.7 17.8 60.9 26.8 23.3 52.622 29.4 17.3 61.5 25.7 22.7 53.038 28.6 17.0 61.8 24.8 23.0 53.474 27.8 17.1 62.4102 27.8 17.2 62.6The number of reflections included in the Rni„ge calculation was 12087 and 9241 for thePro7lSer and Leu85Phe data sets, respectively.tData for the Leu85Phe mutant was provided by T.P. Lo.t The merging R-factor over all intensities is defined as:^Ehkl^o jIz (hkl)Rmerge =^Ehki E7-0 ii (hkoThe software used for the processing of data frames collected on the FAST instrument,MADNES, was written by Messerschmidt Pflugrath (1987). Due to the small oscillationangle, no single frame collected on the FAST instrument contains a complete diffraction spot.Adjacent frames are grouped together and a three-dimensional integration box is centered on anexpected peak position. The R-AXIS data processing software is based on techniques describedby Rossman et al. (1979). This method uses a two-dimensional integration box and combinespartial peaks spread over two frames after intensity integration. In both methods, two scalingfactors, K and B, are computed for each frame using symmetry related and duplicate intensitymeasurements by a least-squares fit of the expression:Is = Im x-e-B(sV- ) 2^ (2.10)Chapter 2. General Experimental Methods^ 36where Is and 'm are the scaled and measured intensities, respectively. These scaling factors alsocorrect for crystal decay and absorption effects. The scaled intensities are merged, correctedfor Lorentz and polarization effects, and are then converted to structure factor amplitudes.2.1.4 Structure Solution and RefinementAs discussed above, all crystal forms studied in this work were isomorphous. This isomorphismallowed the creation of an initial phase model for each data set through the construction of amodel for the new form of cytochrome c based on the previous structural results obtained foriso-1-cytochrome c (Louie & Brayer, 1990). In the case of structures with mutations at only afew sites, the model was constructed by removing the side chain atoms at the sites of mutationexcept for the /3-carbon, to form alanine residues. The amino acid substitutions in the iso-2and B-2036 proteins were modeled explicitly as described in Chapter 3. In the case of iso-2-cytochrome c, no solvent was initially included in the refinement model, however, for the othermutant proteins, the sulfate ion and a selection of water molecules from the iso-1 structure wereincluded in the initial refinement model. The solvent molecules were chosen based on how wellthey were conserved in the mutant structures previously determined and their thermal factorsin the refined iso-1 structure.Difference maps were computed using the coefficients, Foil"' — Fiov, where Fiovi and F iav arethe observed structure factor magnitudes of the mutant and native structures, respectively.The resulting electron density maps were inspected with the program FRODO (Jones, 1978)as modified by S. Oatley and S. Evans. For some mutations, this type of map was used todetermine the initial orientation of newly modified side chains. If the map did not suggest anunambiguous side chain placement, the side chain involved was modeled as an alanine residueand placed in the unit cell in the same orientation and position as the refined iso-1 structure.Structure refinement was then carried out by a restrained parameter least-squares procedurewith the program PROLSQ (Hendrickson & Konnert, 1981) using data with F > 2o-F. AFo — F, electron density map was then computed to determine the conformation of the sideChapter 2. General Experimental Methods^ 37chain of the mutant residue. Omit, Fo — Fc , 2F0 — F, and 3F0 — 211, maps were also examinedat intervals during the course of refinement as a guide for the manual adjustment of side chainsin the structure. Water molecules, modeled as oxygen atoms, were added to the model bysearching for peaks in Fo — F, maps. A water molecule was included in the refinement if atleast one hydrogen bond was formed to the existing structure and the oxygen atom refined toa reasonable thermal factor (less than 55 A2 ).2.2 Structure Analysis2.2.1 Estimation of Coordinate ErrorIn the analysis of structures of mutant proteins, atomic coordinates are compared to those ofthe native enzyme and other available mutant structures. A prerequisite to such a comparisonis some estimate of the accuracy and reliability of the coordinate information. Three methodswere used to estimate the coordinate accuracy of structure determinations. The first method,as described by Luzzati (1952), assumes that a complete data set was used in the structuredetermination and that the only error is the coordinate error. In this method, the conventionalR-factor as a function of resolution is plotted, along with theoretical curves calculated byassuming various r.m.s. coordinate errors. Inspection of these plots for the structures describedin this work suggests radial coordinate errors in the range of 0.15 to 0.20 A. This value is astatistical estimate of the overall r.m.s. coordinate error, whereas the actual positional errorof an individual atomic coordinate will depend on the location of the atom in the structure.However, neither of the intrinsic assumptions of this method was fully met. The data usedfor refinement was selected by the criteria that F > 2o-F resulting in the use of 50 to 80% ofthe theoretically accessible data. In addition, the refinement method includes only an isotropicthermal factor term for each atom. The effect of this term on the coordinate accuracy isunknown.The second method uses an individual atom approach assuming a complete model and onlycoordinate error (Cruickshank. 1949; Cruickshank, 1954; Chambers & Stroud, 1979). The errorChapter 2. General Experimental Methods^ 38in the coordinates of a given atomic position are dependent on the fit between the observed andcomputed structure factors, the type of atom and the thermal factor of the atom. The overallr.m.s. error of the protein atoms (no solvent atoms included) of the proteins studied rangesfrom 0.15 to 0.27 A. As in the Luzzati method, the relatively incomplete solvent structuremodels and the presence of disordered side chains in these structures violates the assumptionsof the method. Nonetheless, the Cruickshank and Luzzati methods do give comparable globalcoordinate error estimates.A third indication of error in structure determinations may be derived by a comparison ofthe mutant structures to the wild-type yeast iso-1 structure. As a guideline, the expected overallr.m.s. deviation may be assumed to be equal to the overall r.m.s. deviation error computed fromthe error estimates provided by the Cruickshank method and the equation:CA-B 7-7 1ICrA2 + 4 (2.11)where A and B are equivalent atoms in two different structures being compared, o-A_B is theerror in the positional deviation between atoms A and B, and uA and 5B are the coordinateerrors of atoms A and B from the Cruickshank method. The actual observed overall r.m.s. mainchain deviations and the calculated overall r.m.s. deviation errors using the equation above forthe Pro71Ala, Pro71Ile, Pro7lSer and Pro71Val mutants are presented in Table 2.10. Asexpected, the computed r.m.s. deviation errors are smaller than the observed r.m.s. deviationsbecause the calculations do not account for actual structural perturbations between the mutantand native structures.The formation of crystal contacts is another potential source of coordinate variability. A listof crystal contacts in the yeast iso-1-cytochrome c structure is presented in Louie and Brayer(1990). Since the structures studied in this work form crystals which are isomorphous withthose of wild type yeast iso-l-cytochrome c, the regions of these proteins that form contacts inthe crystal are very similar. Therefore, differential crystal contacts are unlikely to be a factorin the observed differences in comparisons between wild type yeast iso-l-cytochrome c and thecytochrome c variant structures studied herein.Chapter 2. General Experimental Methods^ 39Table 2.10: Comparison of overall observed main chain r.m.s. deviations and the expecteddeviation errorsMutant Structure^ Pro7lAla Pro7lIle Pro71Ser Pro71Valobserved r.m.s. deviation (A) 0.18 0.16 0.19 0.22calculated r.m.s. deviation error (A) 0.14 0.10 0.13 0.14The observed and calculated r.m.s. deviation of each mutant structure was calculated for themain chain atoms of residues 1 to Comparison of Structure CoordinatesIn the past, when a new protein structure became available the basis of protein structuralanalysis consisted of a detailed description of the observed structure. More recently, the avail-ability of many closely related three-dimensional structures has required the development ofnew techniques for their analysis. This is particularly evident in the case of multiple structuresof mutant proteins where a direct comparison to the native structure is desired.Direct visual comparison of related protein structures with computer graphic representationsrequires that the structures be superimposed. The structures depicted in all of the figures in thisthesis were superimposed as follows. An orientation matrix was computed by a least-squaresfit of the target structure to the parent structure of the common main chain atoms of residues—3 to 103 and the atoms of the heme group (43 atoms). The residues from the amino terminusto residue —4 were excluded since they are positionally disordered. The orientation matrix wasthen applied to the entire coordinate set of the target structure to produce a superimposedcoordinate set. The iso-2 structure was used as the parent structure in the analysis describedin Chapter 3, whereas the iso-1 structure is the parent structure in Chapters 4 and 5.An alternative to direct visual inspection of the atomic coordinates is the constructionof difference matrices. A C a positional difference matrix is constructed by computing a setof intra-molecular C a — Ca vectors for each structure. The matrix element m ii contains thedifference in the length between the equivalent C a — Ca, vectors in the two structures for residuesChapter 2. General Experimental Methods^ 40i and j. A similar matrix may be constructed by comparing the difference in the averageresidue main chain thermal factors of two structures (Berghuis & Brayer, 1992). The matricesare symmetrical about the diagonal and thus the two types of matrices may be combinedto eliminate the redundant information. In this work, the C, and thermal factor plots werecombined by computing the absolute value of the matrix elements. The matrix was thencontoured and plotted to provide a direct view of the main chain positional and thermal factorchanges between two structures. One advantage of the difference matrix method of structuralcomparison is that it does not require previous superposition of the structures which may biasthe comparison.2.2.3 Comparison of Derived Structural PropertiesIn addition to directly comparing the coordinates of two structures, a comparison of derivedstructural properties such as hydrogen bonds is useful in determining how structure is relatedto stability and function. In this study, hydrogen bonds were defined by the following criteria:a H...A distance < 2.60 A, 2.70 A or 3.05 A where A is an oxygen, nitrogen or sulfur atomrespectively; a D-H...A angle > 120°; and, a C-A...H angle > 90°. The hydrogen atom waspositioned as close to the hydrogen bond acceptor as possible within the constraints of knownhydrogen-heteroatom bond lengths and angles. The program Hbond created by S. Evans wasused to define hydrogen bonds according to this scheme. The solvent exposed surface area isanother frequently examined structural property of protein structures (Lee & Richards, 1971).In this study, the surface area exposed to solvent was computed by rolling a probe with a 1.4 Aradius over the molecular surface as implemented by Connolly (1983). All of the internal cavityvolume measurements were computed by the algorithm described in Connolly (1985) using aprobe radius of 1.1 A. Other derived structural properties of interest include average main chainand side chain thermal factors, residue torsional angles and lists of residue neighbors.Chapter 2. General Experimental Methods^ 41To date, few methods have been described that combine and compare these derived proper-ties between many closely related structures. To address this issue for the cytochrome c struc-tures discussed in this work, an information retrieval system was developed named SDBase.Filter programs were constructed to convert the results of several programs that produced listsof hydrogen bonds, torsional angles, solvent exposed surface area, neighboring contacts andaverage thermal factors to the format used by SDBase. In SDBase, both structurally derivedand other related information such as literature references are stored and keyed by the nameof the structure, the residue number and the type of information. The search program allowsretrieval of specific combinations of information by the creation of regular expressions (Denninget al., 1978) to match the information keys. For example, a regular expression consisting of astring of ordinary characters would match any key which contained the same string. A sampleSDBase search using this type of regular expression is described in Figure 2.6.The SDBase system was used in conjunction with molecular graphics programs to examineunusual structural features observed in mutant structures. In addition, this system was instru-mental in preparing the tables of conserved and altered hydrogen bond interactions found inthis thesis. Derived parameters from other native structures of mitochondrial cytochrome ccomputed from coordinates obtained from the Protein Data Bank (Bernstein et al., 1977) werealso added to the database, allowing a broader comparison of the structural features of cyto-chrome c. A database containing more than 10000 separately keyed pieces of information onover 20 structures has been constructed and is routinely queried.2.3 Electrochemical Properties2.3.1 Direct ElectrochemistryA schematic diagram of the direct electrochemistry experiment is presented in Figure 2.7. Theexperimental methods follow those described by Barker et al. (1989), Rafferty et al. (1990) andRafferty (1992). A 25 mm diameter gold disk electrode was polished with increasingly finegrades of alumina from 0.3 pan to 0.05 pm and sonicated in deionized water. The electrodeChapter 2. General Experimental Methods^ 42SDBase> position 26SDBase> field bond nameSDBase> structure isol iso2 b2036SDBase> search26 hbond B2036,iso2 ASN 31 N . ASN 26 OD126 hbond isol HIS 26 NE2 . GLU 44 026 name B2036,iso2 ASN26 name isol HISSDBase>Figure 2.6: In this SDBase information search, the key word "position" is used to limit thesearch to residue 26. The next two lines further limit the search by the use of regular expressionsto information concerning hydrogen bonds, residue names and to the structures of iso-1, iso-2and B-2036 cytochrome c. The keyword "search" initiates the search, the results of which areprinted below.was electrochemically cleaned in a solution of 100 mM NaC104 and 20 mM phosphate buffer,pH 6, by cycling through a potential range of --1 to 1.2 V. The gold surface was then modifiedby immersion into a saturated solution of 4,4'-dithiodipyridine to sufficiently increase the rateof electron transfer to cytochrome c. A stream of water saturated argon was passed over theprotein solutions to removed dissolved oxygen. A potential range of —44 to 544 mV (versusSHE) was scanned at sweep rates of 10 to 100 mV s -1 .All protein samples were oxidized with NH4[Co(dipicolinate)] and excess oxidant was re-moved with a Biorad P6-D6 desalting gel filtration column (1 x 20 cm). The oxidation stateand concentrations of the samples were determined by UV-visible spectroscopy, assuming anextinction coefficient of 106.1 mM —l cm -1 at 409.5 nm. Experiments were performed on ap-proximately 0.5 ml of a 0.4 mM protein solution in f.t = 0.1 M buffer (50 mM KC1 with theremaining ionic strength provided by sodium phosphate), at pH Derivation of Electrochemical Thermodynamic PropertiesA cyclic voltammogram of iso-2-cytochrome c indicating the peak oxidizing and reducing cur-rents is presented in Figure 2.8. The midpoint reduction potential was computed as the averagePotentiostatIWVV■11111■1• 1■11■11.111.• •Side armAuxiliary electrodeCapillaryProteinsolutionWorking electrode(Gold)Chapter 2. General Experimental Methods^ 43ReferenceelectrodePlotterFigure 2.7: Schematic of the direct electrochemistry experimental setup used to measure reduc-tion potentials. The three electrodes of the electrochemical cell are attached to a potentiostat.The buffer solution in the side arm is prevented from bulk mixing with the protein solution bya capillary junction. The potentiostat cycles a selected ramp voltage relative to the referenceelectrode through the cell. This voltage is recorded on the horizontal axis of the plotter. Theresulting Faradaic current at the working electrode is recorded on the vertical axis of the plotter.Epo = 314 mVChapter 2. General Experimental Methods^ 443 —2—0-1)L-1-—2—Epp = 260 mV100^200^300^400^500^600Eapp (mV versus SHE)Figure 2.8: A sample cyclic voltammogram of iso-2-cytochrome c at 25 °C at a sweep rate of50 mV/s is plotted. The broken lines were drawn tangentially to the baselines of the voltammo-gram. The applied voltage when current peaked at both the positive anode (Epa ) and positivecathode (Epc ) are indicated by dotted lines.(287 mV) of the applied voltage at the two peak currents (25 °C, pH 6.0 and p=0.1 M). Thepeak voltage difference (54 mV) is similar to the theoretical voltage (57 mV) expected for afully reversible system (Greef et al., 1985). The method used to calculate the enthalpy and en-tropy from the temperature dependence of the midpoint reduction potentials was obtained fromTaniguchi et al. (1980) and Rafferty, (1992). Since only the temperature of the cytochrome creaction couple was changed while the temperature of the reference was held constant, the en-0tropy (ASeyt ) of the cytochrome c half reaction was computed directly from the slope of a plotof the midpoint potential (E7 ) versus temperature (T) using the equation:—30Chapter 2. General Experimental Methods^ 4500^DE„,ASeyt nF aT (2.12)where n is the number of electrons transferred in the reaction and F is the Faraday constant.The standard conditions were /1 = 0.1 M and pH 6. Since n = 1 for cytochrome c, the value0^0of nF is 23.06 cal mot -1 mV-1 . By definition the AGH and AHH of the standard hydrogenelectrode (SHE) are zero, however, the practical entropy (ASH ) of the SHE is 15.6 e.u. (Latimeret al., 1938; Latimer, 1952) and the entropy of the complete reaction was computed by:0^0AS = .AScyt — ASH^ (2.13)0= AScyt 15.6 e.u. (2.14)The free energy and enthalpy of the complete reaction was then derived from the equations:0^0AG = —n.FEn,^ (2.15)0^0^0OH = AG + T.AS (2.16)Chapter 3Structure - Function Analyses of Yeast Iso-2-Cytochrome c and the CompositeMutant Protein B-20363.1 Experimental ProceduresCrystals of yeast iso-2-cytochrome c were initially grown by hair seeding hanging drops withcrushed yeast iso-l-cytochrome c crystals as previously described (Leung et al., 1989). Crystalsused in diffraction analyses were transferred into a freshly prepared solution of 97% saturated(NH4)2SO4, 0.3 M NaC1, 0.1 M sodium phosphate buffer (pH 6.0), and 0.03 M dithiothreitol,to maintain the reduced state of the heme group. A complete diffraction data set to 1.9 Awas collected using a single crystal on an Enraf Nonius diffractometer using continuous Qscans of 0.4° at 0.55°/min, at an ambient temperature of 15 °C. Three standard reflectionswere measured every 2.2 hours of X-ray exposure time to monitor crystal decay and slippage.Crystal specific data collection statistics are presented in Table 3.11.An initial model for the structure of iso-2-cytochrome c was constructed by using the coor-dinates of the refined iso-1 structure as a template (Louie & Brayer, 1990). An alignment of theiso-1 and iso-2 primary sequences is presented in Table 3.12. Amino acid sequence differenceswere modeled with the computer program MUTATE (R,.J. Read, unpublished results). In thisprocedure, at positions where amino acid substitutions have occurred, the template side chainwas replaced by a new side chain with a conformation that follows the path of the templateside chain as far as possible. Where this is not possible, the conformation is determined froma library of preferred side chain conformations (Bhat et al., 1978: Janin et al., 1978; JamesSielecki, 1983; Moult & James, 1986). All of the new amino acid side chains were accom-modated into the iso-2-cytochrome c fold without the formation of unacceptably close van der46Chapter 3. Iso-2-Cytochronle c and a Iso-1/Iso-2 Composite Mutant Protein^47Table 3.11: Data collection statistics of iso-2 and B-2036 cytochromes cParameter Iso-2 B-2036Crystal size 0.2 x 0.4 x 0.4 mm 0.2 x 0.25 x 0.25 mmUnit cell parameters (A):a, b 36.44(1) 36.39(1)c 137.86(4) 137.28(4)Resolution (A) 1.9 1.95No. of reflections measured 8159 10866No. of unique reflections 7946 7352Max. decay correction 1.3 1.5Max. absorption correction 3.0 1.4R-scale with iso-1t 30.4% 20.9%t This R-scale is defined as:^2 E I Foi —^IE (Fit, Fj)where Foi and F'of are the observed structure factor amplitudes of wild-type iso-l-cytochrome cand the data set being compared, respectively.Waals contacts. As a result, no manual adjustments were applied to the iso-2 starting model.The four extra amino terminal residues of iso-2-cytochrome c were not included in the startingmodel.The resulting iso-2 starting model was placed in the unit cell in the same position andorientation as the refined iso-1 structure. The model was refined by a restrained parameterleast-squares procedure with the program PROLSQ (Hendrickson & Konnert, 1981) using datawith F > 2o- (F). Fragment deleted, Fo — Fc , 2Fo — Fe and 3F0 — 2F, maps were examinedat intervals during the course of refinement as a guide for manual adjustment of side chains inthe structure. The four additional amino terminal residues were added to the model in groupsof two. Water molecules, modeled as oxygen atoms, were added to the model periodically bysearching for peaks in Fo — F, maps. A water molecule was included in the refinement if at leastone hydrogen bond was formed to the existing structure, and if 2F, — F, or 3F, — 2F, mapshad significant electron density at the water position. In total, 58 bound water molecules withChapter 3. Iso-2-Cytochrome c and a Iso-1/Iso-2 Composite Mutant Protein^48Table 3.12: Sequence alignment of iso - 1, B -2036 and iso -2 -cytochromes c-9 1^10 20Iso-1^-^-^-^- T E FK A GSAKKGATLFKTRC L QCHT V E K G G P H K V GB2036 -^-^- - E FK A GSAKKGATLFKTRC QCHT I E E GGP N KVGIso-2^AK ES T G FK P GSAKKGATLFKTRC Q QCHT E E G G P N K V G30 40 50 60Iso-1 PNLHGIFGRHSGQ A E GYSYTDANI K K N V L W D E NN MB2036 PNLHGIFGRHSGQ VK GYSYTDANI N KNV K WDE DS MIso-2 PNLHGIFGRHSGQ V K GYSYTDANI N K N V K W DE DS M70^80 90 100Iso-1 SEYLTNPKKYIPGTKMAF G GLKKEKDRNDLITY L K KA C E-B2036 SEYLTNPKKYIPGTKMAF G GLKKEKDRNDLITY L K KA C E-'so-2 SEYLTNPKKYIPGTKMAF A GLKKEKDRNDLITY MT KA AK-The primary sequences of iso-1 (Smith et al., 1979), B-2036 (Ernst et al., 1982) and iso-2(Montgomery et al., 1980) cytochromes c have been aligned. Boxes enclose regions of sequenceidentity. For the composite B-2036 mutant protein, the bold sequence represents the regionderived from the yeast iso-2-cytochrome c average thermal factor of 25.5 A2 were identified. An exceptionally large peak of densityappeared at the amino terminal end of an a-helical segment, involving residues 2 to 4, and wasmodeled by a SO — ion as in the iso-1 structure (Louie Si Brayer. 1990). Over the course ofrefinement, the conventional crystallographic 11-factor was reduced from 34.8 to 19.0% (22.4%for Fo > a(F0))•Crystals of B-2036 were obtained by techniques similar to those used in the growth ofiso-2 crystals. These crystals were grown in a solution of 94% saturated (NH4)2SO4, 0.1 Msodium phosphate buffer (pH 6.2), and 0.04 M dithiothreitol. Free interface diffusion capillarieswere used to obtain the largest crystals (Saleinme, 1972). Shortly before data collection, thecrystals were transferred into a freshly prepared solution of mother liquor. Data collectionand subsequent corrections were performed as described in Chapter 2 and the resulting datastatistics are also presented in Table 3.11. A starting model for B-2036 was prepared fromthe coordinate sets of the iso-1 and iso-2 structures. The sulfate ion and 15 water moleculesChapter 3. Iso-2-Cytochrome c and a Iso-1/Iso-2 Composite Mutant Protein^49conserved in the iso-1 and iso-2 structures were also included in the starting model. Duringthe course of refinement, 7 of the original 15 water molecules were later removed from thestructure and 47 additional water molecules were identified for a total of 55 bound watermolecules (average thermal factor of 28.5 A 2 ) in the final structure. The conventional R-factorwas reduced from 35.3 to 17.5% (18.8% for Fo > F0)) using the same techniques as describedfor the iso-2-cytochrome c protein.The agreement of the iso-2 and B-2036 structures with ideal stereochemical geometry and thecorresponding refinement weights used in the final refinement cycle are outlined in Table 3.13.For both structures, the deviations from ideal stereochemistry are well within the estimatedcoordinate error. Figure 3.9 shows a plot of the conventional R-factor as a function of resolution,Table 3.13: Final stereochemistry of iso-2 and B-2036 cytochromes cRestraint r.m.s. deviation from ideal values(restraint weighting values)Iso-2^B-2036Distances (A)Bond 0.020 (0.020) 0.019 (0.020)Angle 0.041 (0.030) 0.037 (0.030)Planar 1-4 0.050 (0.050) 0.047 (0.045)Plane (A) 0.016 (0.020) 0.016 (0.020)Chiral volume (A 3 ) 0.205 (0.160) 0.196 (0.150)Torsional angle (deg.)Planar 2.4 (2.5) 2.6 (3.0)Staggered 25.7 (20.0) 22.8 (20.0)Orthonormal 14.5 (15.0) 12.0 (15.0)Non-bonded contact (A)tSingle torsion 0.213 (0.250) 0.218 (0.250)Multiple torsion 0.202 (0.250) 0.186 (0.250)Possible hydrogen bonds 0.249 (0.250) 0.227 (0.250)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.0.05 0.10Resolution3.3 A^2.5 A 1.7 A5.0 A10.0 A-oChapter 3. Iso-2-Cytochrome c and a Iso-1/Iso-2 Composite Mutant Protein^50Figure 3.9: A plot of the conventional R-factor as a function of resolution at the end of refine-ment. The fit to the data of iso-2 (0) is represented by the thick line and the fit to the B-2036data (^) is represented by a thin line. The theoretical dependence of R-factor on resolution,assuming various levels of r.m.s. error in the atomic positions of the model (Luzzati, 1952), aredrawn in broken lines. The fraction of data used in refinement is presented above, using thesame data point and curve scheme.along with theoretical curves calculated by assuming various r.m.s. coordinate errors (Luzzati,1952). Inspection of this plot suggests a coordinate error for both structures of approximately0.2 A. This value is a statistical estimate of the overall coordinate error, whereas the actualpositional error of an individual atomic coordinate will depend on the location of the atom inthe structure. The errors of atoms in the core of the protein are likely to be less than 0.2 A,whereas highly mobile atoms on the protein surface are less accurately determined. The r.m.s.coordinate errors of residues —5 to 103 and the heme group by an individual atom method(Cruickshank, 1949; Chambers & Stroud, 1979) are 0.19 A and 0.16 A for iso-2 and B-2036,respectively. Also included in Figure 3.9 is a plot of the fraction of structure factors includedin the refinement as a function of resolution.Chapter 3. Iso-2-Cytochrome c and a Iso-1/Iso-2 Composite Mutant Protein^513.2 Polypeptide Chain ConformationsFrom the coordinates of the iso-2 and B-2036 cytochrome c structures, hydrogen bonds weredefined by geometrical criteria (Table 3.14). The majority of the hydrogen bonds are conservedbetween the iso-2 and B-2036 structures. Many of the observed differences in main chainhydrogen bonds result from the formation of bifurcated interactions within a-helical segments,where the amide of a residue hydrogen bonds with both of the carbonyls of the i + 3 andi + 4 residues. Changes in hydrogen bonding involving side chain atoms result from amino acidsubstitutions and alternative placement of amino acid side chains. The latter is particularlyevident for poorly ordered residues such as G1u21, Lys58, Glu61 and Lys86.Plots of the main chain 0 and 0 angles for both iso-2 and B-2036 cytochrome c are presentedin Figure 3.10. In total, ten non-glycyl residues have values that fall outside the permissibleregions of these two plots. The residue with angles the furthest from a permissible region isLys27 (0, = —158°, —138° in iso-2; —137°. —127° in B-2036) which is part of a 7-turn as iniso-1-cytochrome c (Louie & Brayer, 1990). The other eight residues are not far from permissibleregions of the plot. Asn56 is the only non-glycyl residue found in the aL conformation. Thisresidue is part of a 310 turn that terminates an a-helix and has a comparable conformation iniso-l-cytochrome c (Louie & Brayer, 1990).Main chain hydrogen bonds and 0, 0 angles have been used to classify sections of the iso-2polypeptide chain into secondary structure elements (Table 3.15). In total, five a-helices, threetype I, three type II and one 7 turn were identified. The i + 2 position of each type II /3-turn is a conserved glycine residue (Table 3.12). The secondary structure assignments for theB-2036 structure are identical with the exception of Helix III, where the last turn is in a 310conformation and Helix V which ends at residue Cys102.An a-carbon tracing of yeast iso-2-cytochrome c with the heme iron ligands, His18 andMet80, as well as the heme thioether linkages to residues Cys14 and Cysl7, is illustrated inFigure 3.11. In Figure 3.12, all atoms for the iso-2 protein are presented in the same orientationas in Figure 3.11. The amino acid sequence numbering scheme is as described in Table 3.12.Chapter 3. Iso-2-Cytochrome c and a Iso-1/1 -so-2 Composite Mutant Protein^52Table 3.14: Hydrogen bond analysis of iso-2, B-2036 and iso-1-cytochromes cI. Common hydrogen bondst (iso-2/13-2036/iso-1)A. Main - main6 N - 2 0 24 N - 21 0 64 N - 60 0 92 N - 88 07 N - 3 0 27 N - 29 0 65 N - 61 0 93 N - 89 08 N - 4 0 29 N - 17 0 67 N - 63 0 94 N - 90 09 N - 5 0 32 N - 19 0 68 N - 64 0 95 N - 91 010 N - 6 0 34 N - 102 0 69 N - 66 0 96 N - 92 011 N - 7 0 35 N - 32 0 70 N - 67 0 97 N - 93 012 N - 8 0 38 N - 35 0 74 N - 70 0 98 N - 94 014 N - 10 0 40 N - 57 0 75 N - 71 0 99 N - 95 015 N - 10 0 53 N - 49 0 78N-750 100 N - 96 017N-140 54 N - 50 0 85 N - 68 0 101 N - 97 018 N - 14 0 59 N - 38 0 91 N - 87 0 102 N - 98 0B. Main -side1N - Thr96 OG12 N - Asp93 OD15 N - Ser2 OGThr8 0G1 - 5 0Thrl2 OG1 - 8 0Thrl2 OG1 - 9 0Hisl8 ND1 - 30 0Thr19 OG1 - 25 0Asn31 ND2 - 21 0^Tyr46 OH - 28 0^73 N - Asn70 OD133 N - Asn31 OD1^49 N - Hem 02D Lys79 NZ - 47 0His33 ND1 - 20 0^52 N - Thr49 OG1^79 N - Hem OlDArg38 NH1 - 33 0^Lys55 NZ - 74 0 80 N - Thr78 OG1Arg38 NH2 - 33 0^57 N - Ser40 OG^Lys86 NZ - 69 0Ser40 OG - 52 0 63 N - Asp60 OD1^Arg91 NH2 - 85 041 N - Hem 02A^Thr69 OG1 - 65 0^Thr96 OG1 - 92 043 N - Tyr48 OHC. Side -sideThr19 OG1 - Asn31 ND2 Thr49 OG1 - Hem OlD Tyr67 OH - Met80 SD Arg91 NE - Ser65 OGTyr48 OH - Hem 01A Asn52 ND2 - Hem 02A Thr78 OG1 - Hem OlDII. Unique hydrogen bonds to iso-2 and B-2036A. Main -main13 N - 10 0^55 N - 51 037 N - 34 0B. Main -side31 N - Asn26 OD1^Ser63 OG - 58 0III. Unique hydrogen bonds to iso-2 and iso-1A. Main -sideLys27 NZ - 15 0^62 N - Asp60 OD1B. Side -sideTrp59 NE1 - Hem 02.,A,56 N - 53 0Lys86 NZ - 83 0Arg91 NH2 - 86 066 N - 62 0ContinuedChapter 3. Iso-2-Cytochrome c and a Iso-1/Iso-2 Composite Mutant Protein^53IV. Unique hydrogen bonds to B - 2036 and iso - 1A. Main -main13 N - 9 0^82 N - 80 0B. Main -sideCys102 SG - 98 0V. Hydrogen bonds present only in iso-2A. Main - main(-5) N - (-9) 0 34 N - 103 0 92 N - 89 0 100 N - 97 0(-4) N - (-8) 0 70 N - 66 0 96 N - 93 0 103 N - 100 010 N - 7 0B. Main -side(-9) N - G1u66 0E2 Thr(-5) 0G1 - (-9) 0 Ser47 OG - 47 0 Thr69 OG1 - 66 0Thr(-5) OG1 - (-8) 0 Lys27 NZ - 16 0C. Side -sideSer63 OG - Tyr74 OH Tyr74 OH - Ser63 OG Asn92 ND2 - G1u88 0E2VI. Hydrogen bonds present only in B-2036A. Main -side46 N - 43 0B. Main -side(-5) N - Thr(-5) 0G1^Lys58 NZ - 36 0^60 N - Ser63 OG^Ser63 OG - 60 0Asn26 ND2 - 44 0 Lys58 NZ - 59 0 62 N - Asp60 OD2C. Side -SideThr(-5)0G1 - Asp62 OD1 Lys5 NZ - Asp93 OD2 Ser63 OG - Asp60 OD1 Lys99 NZ - Glu61 0E2VII. Hydrogen bonds present only in iso-1A. Main - main37 N - 59 0^55 N - 52 0^69 N - 65 0^73 N - 70 048 N - 46 0B. Main-side(-3) N - Thr(-5) OG1^His26 NE2 - 44 0^31 N - His26 ND1^Lys99 NZ - 96 020 N G1u21 OE1C. Side -sideThr(-5)OG1-G1u61 OE1 Thr(-5)0G1-Asp62 OD1 Asn63 ND2-Asp60 OD2 Tyr74 OH-Asn63 OD1 Where main chain atoms are listed, only the residue number is given. The abbreviations iso-1, iso-2 andB-2036 refer to yeast iso-1, yeast iso-2, and B-2036 cytochromes c, respectively.tHydrogen bonds were defined by the following criteria: a H...A distance < 2.60 A, 2.70 A or 3.05 Awhere A is an oxygen, nitrogen or sulfur atom respectively; a D-H...A angle > 120°; and, a C-A...Hangle > 90°. The hydrogen bonds listed are defined by geometry only and do not account for disorderin side chain positions in the structures.o000^0 S°°^OD..•.0 .•'',,0(000 0D•0J /r:ICI^*,...*IC0^.....rI'^*Sx• ••0oa^r,^\0 ,0^-.0 0C8 )^o0^•0*•-120^-60^0^60^120^180cca 180120601p-60-120-180-180b 180-180-180-120gfr 0-60120600)o ° 6 °*°621carii)®^o CS0^o^0 Q0 v0 00.....',.o4,..^I* *0^,D^....."'...-"j/4^I0^\ix &^0\ s0^6?^o° 0 ",0,^. •^,,**I^ •t0*-120^-60^0SO60^120 180Chapter 3. Iso-2-Cytochrome c and a Iso-1/Iso-2 Composite Mutant Protein^54Figure 3.10: Ramachandran plots of the and angles of (a) iso-2 and (b) B-2036 cytochromes c(Ramakrishnan Ramachandran, 1965). Glycine residues are denoted by an asterix (*) andall other residues by (0). Main chain conformations for an alanine residue in poly(L-alanine)which are fully allowed, and allowed assuming shortened permissible minimum interatomicdistances are enclosed by solid and dashed lines, respectively.Chapter 3. Iso-2-Cytochrome c and a Iso-1/Iso-2 Composite Mutant Protein^55Table 3.15: Secondary structural elements in iso-2-cytochrome cSecondary structuralelementResiduesinvolvedMain chain torsionalangles (deg.)tA. Average anglesa-Helix I 2-14 (-70, -38),Q-Turn (type I) 14-17 (-62, -38), (-46, -44)/3-Turn (type II) 21-24 (-59, 131), (71, 17)-y- Turn 27-29 (-158. -138), (-59, -46), (-106, 170)P-Turn (type II) 32-35 (-54, 136), (81, 5)/3-Turn (type I) 34-37 (-50, -56), (-51, -33)/3-Turn (type I) 35-38 (-51, -33), (-93, 33)a-Helix IIt 49-56 (-63, -39)a-Helix III 60-70 (-67, -39)a-Helix IV 70-75 (-65, -40)P-Turn (type II) 75-78 (-59, 130), (105, -25)a-Helix Vt 87-103 (-61, -43)B. Overall averagesa-Helix 45 residues (-63, -40)P-Turn (type I) 3 turns (-54, -42), (-63, -15)/-Turn (type II) 3 turns (-57, 132), (86, -1)t The main chain torsional angles refer to the average (0,0) angles for residues in each a-helicalsegment, excluding the 2 terminal residues. For /3-turns, the angles listed correspond to (0,0)2and (0,0)3, respectively. All three of (00/0 1 , (0,0)2 and (0,0)3 are given for the ry-turn.Overall averages for each class of structures are also listed. The a-helices have been numberedfrom the N-terminal end of the polypeptide chain.t The last turn of these a-helices is in a 310 conformation.The heme component is almost completely buried by the surrounding polypeptide chain, withonly part of one edge exposed to bulk solvent.3.3 Comparison of Iso-2, B-2036 and Iso-1-Cytochromes cA comparison of the superimposed structures of iso-2. B-2036 and iso-l-cytochromes c revealsthe strong conservation of the main chain fold in these closely related proteins. A detailedChapter 3. Iso-2-Cytochrome c and a Iso-1/Iso-2 Composite Mutant Protein^56K 5T -5ittD 90F 10Q 15K 100A -91#4twa 1 20*AD 50K5Figure 3.11: Stereo-drawing of the a-carbon backbone of yeast iso-2-cytochrome c. Also drawnin thick lines are the heme group, the two covalent thioether linkages from the heme to Cysl4and Cys17, as well as the side chains of the axial iron ligands, Hisl8 and Met80. The hememoiety is viewed edge on in this orientation. Every 5th residue and the two termini are labeledaccording to the sequence alignment in Table 3.12.Figure 3.12: Stereographic representation of iso-2-cytochrome c with side chains drawn in thinlines and the polypeptide backbone and heme group in thick lines. The orientation and labelingare similar to that in Figure 3. Iso-2-Cytochrome c and a Iso-.I/Iso-2 Composite Mutant Protein^572.5111.0-0.5-kt k-rwit vwf er^ViSfry0.0-6^5^15^25^35^45^55^65^75^85^95^105Residue NumberFigure 3.13: The average positional deviations of main chain atoms between iso-2 and B-2036(thick line), iso-2 and iso-1 (thin line), and B-2036 and iso-1 (dotted line) cytochromes c alongthe course of the polypeptide chain. The horizontal lines represent the overall average deviationsfor the main chain atoms of residues —3 to 103 for each comparison. Vertical arrows denotepositions of amino acid substitutions and the horizontal bar indicates the region of B-2036derived from the iso-2 gene.analysis of main chain positional differences between the three structures is presented in Fig-ure 3.13. The average displacement of main chain atoms between iso-1 and iso-2-cytochromes cfor residues —3 to 103 is 0.30 A. The region of greatest structural difference is at the amino ter-minus, where the iso-2 protein contains an extension of four extra amino acid residues. Theseadditional residues are placed on the surface away from the heme crevice (Figures 3.11 and3.12) and as Figure 3.13 shows, cause the first two residues in common with the iso-1 protein(-5 and —4, Table 3.12) to be in an altered conformation.The next largest main chain displacement is centered around Gly37. The conformations ofiso-1, iso-2 and B-2036 in this region are depicted in Figure 3.14. Two distinct main chain con-formations are observed in electron density maps of the iso-2 structure. One of these is similarChapter 3. Iso-2-Cytochrome c and a Iso-1/Iso-2 Composite Mutant Protein^58Figure 3.14: Stereographic plot of iso-2 (thick lines), B-2036 (medium lines), and iso-1 (thinlines) cytochromes c in the region near G1y37. The a-carbon of G1y37 is displaced 2.4 A in iso-2relative to iso-l-cytochrome c. In iso-l-cytochrome c, G1y37 N forms a hydrogen bond (brokenlines) with Trp59 0; however, in the iso-2 protein, G1y37 N hydrogen bonds with Gly34 0.The hydrogen bond between Arg38 N and 11e35 0 is present in both structures. The B-2036structure is analogous to the iso-2 protein in this that observed in iso-l-cytochrome c where a hydrogen bond is formed between G1y37 N andTrp59 0 (Table 3.14). The second conformation is generated by rotating of Phe36 by 180°,and the formation of a new hydrogen bond between G1y37 N and Gly34 0, generating a type I0-turn (Table 3.15). As part of this movement, the geometry of the nearby G1y38 N to G1y35hydrogen bond is improved and the 0-turn from residues 35 to 38 is converted from type IIin iso-l-cytochrome c to type I in iso-2-cytochrome c. This second conformation was chosenfor structure refinement, since in this conformation the thermal factors of the atoms involvedrefined to lower values. The conformation of Gly37 in the B-2036 structure is similar to thatin the preferred iso-2 conformation. An amino acid substitution in this region of the structuremay provide an explanation for the observed structural differences. This occurs at position58, where there is a leucine in iso-1 and a lysine residue in iso-2 and B-2036 cytochromes c.The Leu58 side chain in the iso-1 structure stabilizes the conformation of Gly37 by hydrophobicpacking interactions. The side chain of Lys58 does not form similar interactions in the iso-2 andB-2036 structures. In B-2036, the alternative conformation is further stabilized by Lys58 NZChapter 3. Iso-2-Cytochrome c and a Iso-1/Iso-2 Composite Mutant Protein^59interacting with Phe36 0 and Trp59 0 (Table 3.14).Another region that displays significant structural heterogeneity is near positions 26 and 43,which are neighbors in the folded structure (Figure 3.15). In the iso-1 structure, His26 formstwo hydrogen bonds through its NE2 and ND1 atoms with G1u44 0 and Asn31 N respectively(Table 3.14). The iso-2 and B-2036 structures have an asparagine at position 26 that alsointeracts with Asn31 N through its OD1 atom. The Asn26 ND1 atom forms a further weakinteraction to Lys44 0 in these two latter structures. The decreased volume of an asparagineside chain relative to that of histidine at position 26 is compensated for by the substitution ofA1a43 for a valine in the iso-2 and B-2036 structures. Small shifts in the overall atomic positionsof residues 23 to 26 and 42 to 45 allow the optimal packing of the different amino acid sidechains in this region. The influence of these changes on the average main chain thermal factorsis illustrated in Figure 3.16.At position 63, there is an asparagine in iso-1, but a serine in B-2036 and iso-2-cytochromes c.The hydrogen bond between Asn63 OD1 and Tyr74 OH in iso-1 is maintained in the iso-2structure with a hydrogen bond between Ser63 OG and Tyr74 OH. Alternatively, in the B-2036Figure 3.15: Stereo-diagram of the region about residue 26 in iso-2 (thick lines) and iso-1 (thinlines) cytochromes c. Hydrogen bonds are represented by broken lines. His26 in the iso-1structure forms hydrogen bonds that bridge the main chain at G1u44 and Asn31. Only one ofthese hydrogen bonds is present in iso-2-cytochrome c as a result of the His26Asn substitution.The nearby substitution of an alanine for a valine at position 43 in iso-2-cytochrome c is alsoillustrated.Chapter 3. Iso-2-Cytochrome c and a Iso-1/Iso-2 Composite Mutant Protein^60e-NaQ 40L00306EL- 20a>0)0a)>- 10-6^5^15^25^35^45^55^65^75^85^95^105Residue NumberFigure 3.16: The average thermal factor of the four main chain atoms of each amino acid iniso-2 (thick line), B-2036 (thin line) and iso-1 (dotted line) cytochromes c are plotted as afunction of residue number. The vertical arrows indicate positions of amino acid substitutionsbetween iso-1 and iso-2-cytochromes c and the horizontal bar represents that region of B-2036that is derived from the iso-2 gene. Differences in the thermal factor profiles between iso-2 andiso-1 (thick line), iso-2 and B-2036 (thin line) and B-2036 and iso-1 (dotted line) cytochromes care plotted above (scale to the upper right). Data for residues —9 to —6, which occur only iniso-2-cytochrome c, are not shown.structure a hydrogen bond is made from Ser63 OG to Asp60 OD1 replacing the interactionbetween Asn63 ND2 and Asp60 OD1 in iso-l-cytochrome c. Since a serine residue is smallerthan an asparagine, small changes in main chain conformation are required to maintain theobserved hydrogen bonds (Figure 3.13).There is also a small peak in the region of Phe82 in the positional deviation plot of Fig-ure 3.13. The nearest sequence substitution is the replacement of Gly83 for an alanine residuein iso-2-cytochrome c. The phenyl ring of Phe82 is observed to shift 0.5 A outwards toward theChapter 3. Iso-2-Cytochrome c and a Iso-1/Iso-2 Composite Mutant Protein^61protein surface in iso-2 and B-2036. As a consequence, this phenyl ring also packs less tightlywith the adjacent side chain of Leu68. The polypeptide main chain shifts with the phenyl ringin the case of iso-2-cytochrome c, and as a result, the hydrogen bond between Phe82 N andMet80 0 is not present in the iso-2 structure (Table 3.14). The conformation of the main chainis more conserved between B-2036 and iso-1-cytochrome c.An examination of the B-2036 starting model revealed a steric clash between 11e20 CD1 andCys102 SG, two residues differing between iso-1 and iso-2-cytochromes c (Table 3.12). Cys102is near the end of Helix V (Table 3.15) and 11e20 is located between the His18 ligand and atype II 0-turn (residues 21-24). Iso-1 has valine and cysteine at positions 20 and 102, whereasin iso-2, these residues are isoleucine and alanine, respectively. This steric clash in the B-2036protein is relieved by displacements of up to 0.8 A in atoms of the side chains of residues 20 and102. There is little displacement in the a-carbon atoms of the residues involved (Figure 3.17),although the shift in the side chain atoms of Cys102 in B-2036 forces a rotation of 15° in themain chain b angle of this residue leading to some distortion at the end of Helix V. Presumably,the His18 ligand and the /3-turn composed of residues 21 to 24 constrain the position of themain chain atoms of 11e20. The side chain atoms of 11e35 and Leu98, two nearby residues thatpack against Cys102 and 11e20, shift up to 0.7 A to accommodate the extra volume of the twosubstituted side chains.3.4 The Buried Cavity in B-2036 and Iso-l-Cytochromes cLeu98 is one of the residues that line a small buried cavity located behind the heme groupin the iso-1 structure (Figure 3.18). The substitution of this residue by a methionine in iso-2eliminates this cavity. As in iso-1, B-2036 also has a leucine at position 98 and an internalcavity. The displacements described previously in the region of Cys102 and 11e20 increase thevolume of this cavity from about 30 A3 in iso-1, to 50 A 3 in B-2036. This enlarged cavity isillustrated in Figure 3.18(b). Both cavities are large enough to accommodate a water molecule;however, no significant electron density appears in this region and the lack of hydrogen bondIle 35he 36Phe 10^Leu 98 Phe 10^Leu 98Chapter 3. Iso-2-Cytochrome c and a Iso-1/Iso-2 Composite Mutant Protein^62Figure 3.17: The structures of iso-2 (thick lines), B-2036 (medium lines) and iso-1 (thin lines)cytochromes c are plotted in stereo in the region about residues 20 and 102. The substitutionof a cysteine for an alanine at position 102 in B-2036 relative to iso-2-cytochrome c results insmall shifts in neighboring residues to accommodate the larger cysteine side prevents favorable enthalpic interactions to offset the large entropic energy requiredto isolate a water molecule from the bulk solvent.A larger cavity could be expected to decrease the stability of the B-2036 protein (Erikssonet al., 1992); however, it appears to be preferable to the steric clash between 11e20 and Cys102.Conversely the substitution Leu98Met in the iso-2 structure completely fills this cavity. Themethionine side chain adopts the same conformation as the leucine side chain it replaces and theX3 torsional angle of 73° is close to a staggered conformation (Figure 3.18). Two heme atoms,CMA and CMB, form part of the internal cavity of both iso-1 and B-2036 cytochromes c. Iniso-2, however, Met98 packs directly against the rear of the heme. The cavity in iso-2-cyto-chrome c seems to be filled without the addition of significant structural strain and, therefore,would be expected to be a stabilizing mutation (Karpusas et al., 1989). Substituting methio-nine for leucine in iso-l-cytochrome c by site-directed mutagenesis could be used to test thedestabilizing effect of the internal cavity in the iso-1 structure. It is of interest that only leucineand methionine are observed at position 98 in mitochondrial cytochromes c (Hampsey et al.,1986).bLeu98 CD1HemLeu98 CD1HemChapter 3. Iso-2-Cytochrome c and a Iso-1/1"so-2 Composite Mutant Protein^63aFigure 3.18: Both the B-2036 and iso-l-cytochromes c contain a buried cavity behind the hemeas viewed in the standard orientation (Figure 3.11). In (a) the structures of iso-2 (thick lines),B-2036 (medium lines), and iso-1 (thin lines) cytochromes c are superimposed, with the cavityfound in iso-1 depicted as a dotted surface. The atoms of the iso-1 structure that border thecavity are labeled. In (b) the surface representation and labels are those of B-2036. In bothplots, the Met98 CE atom of the iso-2 structure fills this cavity.Chapter 3. Iso-2-Cytochrome c and a Iso-1/Iso-2 Composite Mutant Protein^643.5 Heme Geometry and EnvironmentThe overall environment of the heme in iso-2, B-2036 and iso-1-cytochromes c is hydrophobicand highly conserved, with only the heme edge facing the viewer in Figures 3.11 and 3.12being exposed to the bulk solvent. The solvent exposure of the heme is similar in all knownmitochondrial cytochrome c structures, as documented in Table 3.16. In addition, all hemepropionate hydrogen bonds are conserved with the exception of the Trp59 NE1 hydrogen bondto the 02A atom of the heme, which is absent in B-2036 according to the definitions outlined inTable 3.14. In this protein, the indole ring of Trp59 has moved away from the heme propionatepossibly in response to the changes centred at G1y37 described above.The conjugated heme porphyrin ring of yeast iso-2-cytochrome c is distorted from planarity.However, each of the five-membered pyrrole rings, as well as the pyrrole N ligands (pyrrole Nplane) are nearly planar. In Table 3.17, the angular deviations of the pyrrole ring planes froma plane fit to the entire porphyrin ring and the pyrrole N plane are presented along with ironligand distances. The angular deviations and ligand distances are similar to those observedTable 3.16: Heme solvent accessibility of mitochondrial cytochromes cCytochrome c structure Iso-2 B-2036 Iso-1 Tuna Horse RiceSolvent accessible heme atomsand surface area exposed (A 2 )CHD 0.0 2.2 2.9 4.6 0.0 0.0CMC 8.5 8.3 9.4 8.2 8.3 10.4CAC 0.0 3.0 3.7 1.9 5.2 0.0CBC 20.9 18.6 18.0 23.5 17.3 12.8CMD 11.7 10.7 10.4 10.1 3.9 8.1Total heme exposure (A 2 ) 41.4 42.9 44.4 48.3 34.7 31.3% Heme surface area exposed 8.9 9.2 9.5 10.5 7.4 6.8Computations were done using the method of Connolly (1983) with a probe radius of 1.4 A.Chapter 3. Iso-2-Cytochrome c and a Iso-1/Iso-2 Composite Mutant Protein^65Table 3.17: Heme conformation and ligand geometry in yeast iso-2 and B-2036 cytochromes cA. Angular deviations between both plane normals of individual pyrrole rings andthe heme co-ordinate bonds, and both the pyrrole nitrogen plane and theporphyrin ring planetPyrrole ring Pyrrole N plane Porphyrin ring planeA 10.3° (9.1°) 3.8° (4.8°)B 11.5° (11.2°) 10.5° (10.7°)C 12.3° (8.8°) 10.6° (10.7°)D 12.3° (10.9°) 5.8° (6.6°)Heme coordinate bondsFe-His18 NE2 3.2° (3.8°) 9.8° (2.8°)Fe-Met80 SD 2.0° (4.3°) 6.7° (7.1°)B. Henze iron coordinate bond distances (A)Hisl8 NE2 1.86 (2.06)Met80 SD 2.42 (2.33)Hem NA 2.02 (1.99)Hem NB 2.06 (2.03)Hem NC 1.97 (2.02)Hem ND 2.02 (1.99)The values for B-2036 cytochrome c are in parentheses.t Each pyrrole ring plane is defined by 9 atoms, which include the 5 ring atoms plus the firstcarbon atom bonded to each ring carbon. The porphyrin ring plane is defined by all the atomsin the 4 pyrrole ring planes and the iron atom. The pyrrole N plane is defined by only the 4pyrrole nitrogens. These latter two planes differ in orientation by an angle of 6.8°.Chapter 3. Iso-2-Cytochrome c and a Iso-1/Iso-2 Composite Mutant Protein^66in other cytochrome c structures (Bushnell et al., 1990; Louie & Brayer, 1990). The greatestplanar distortion in yeast iso-2, iso-1, B-2036 and horse cytochromes c occurs at pyrrole ringsB and C which are attached to Cys14 and Cys17 through thioether linkages.3.6 Conserved Water MoleculesAn analysis of bound water molecules was performed for the iso-2, B-2036 and iso-1 structures.In this procedure a water molecule was considered structurally conserved if its atomic displace-ment from a counterpart in iso-2-cytochrome c was less than 2 A and conserved water-proteinhydrogen bonds were present. In total, 15 such conserved water molecules were observed (Ta-ble 3.18). Many of these interact with main chain amide and carbonyl groups. Further analysisshows only three waters are conserved between all known cytochrome c structures completedto date. In the yeast proteins, only four waters are completely buried, and as such, form an in-tegral part of the protein structure. Three of these (Wat106, 118 and 131) may play functionalroles in stabilizing the alternative oxidation states of these proteins (Berghuis & Brayer, 1992).3.7 Comparison with Other Cytochromes cThe structure of yeast iso-2-cytochrome c can also be compared to the high resolution structuresof horse (Bushnell et al., 1990), tuna (Takano & Dickerson, 1981a) and rice cytochromes c (Ochiet al., 1983). A plot of the overall average deviations of these cytochromes c from that of theiso-2 protein is presented in Figure 3.19. At each individual residue in the polypeptide chain afurther indication of the range of pairwise deviations between these cytochrome c is given byvertical bars. This comparison shows the structurally most conserved region of cytochrome cstretches from residues 61 to 101. As reference to Figure 3.11 shows, this segment of polypeptidechain forms most of the Met80 ligand side of the molecule. Despite the overall high degree ofstructural homology in this and other regions of the polypeptide chain, there are points ofsubstantial structural difference between these cytochromes c. These include residues 21 to28, 36 to 38 and 55 to 57. All three of these regions occupy positions on the protein surfaceChapter 3. Iso-2-Cytochrome c and a Iso-1/Iso-2 Composite Mutant Protein^67Table 3.18: Conserved water molecules in yeast iso-2, B-2036 and iso-l-cytochromes cWater NumberIso-2^B-2036^Iso-1Hydrogen bonds106tt 106 166 Asn52 ND2, Tyr67 OH, Thr78 0G1107 107 154 100 0, 103 OXT1091: 108 110 19 N, 29 0112 139 197 His39 ND1115 1 110 122 79 0, 81 N117 118 207 66 0, 70 0Mitt 122 121 39 0, 42 N, Hem O1A123 149 133 23 N130 158 138 50 N131t 109 168 Arg38 NH1, Asn31 0 Hem O1A144 145 215 77 N145 126 142 G1u88 0E1153 119 158 86 N, 87 N156 123 208 Asn92 OD1, Ser65 OG158 156 140 46 0Indicates water and hydrogen bonds observed in all of the structures of yeast iso-1, yeast iso-2,rice, tuna and horse cytochromes c.tIndicates a water molecule buried in the protein matrix.(Figure 3.11) and their positioning is strongly influenced by nearby amino acid substitutionswhich occur between the compared cytochromes c (Louie & Brayer, 1990).3.8 Electrochemical PropertiesDirect electrochemistry of iso-2 and B-2036 cytochromes c was reversible from 10 to 35 °C(Figure 3.20). The midpoint reduction potential and the derived thermodynamic parametersat 25 °C are presented in Table 3.19 along with data for iso-1-cytochrome c. Although themidpoint potentials at 25 °C are similar, the enthalpic and entropic contributions to the freeenergy change of reduction of iso-2 and B-2036 cytochromes c have increased. The enthalpicterm has been associated with the polarity of the heme environment through electrostatic effects8040^50^60Residue Number7010^20^30 90^1003 . 0 ^ferrtrrerrirrilrr■r tIrslIttlItItellterstIrrreir esti,  Ir2.52.0-C0•-0 1.5 -(Zt0^-(3)0.5-0.0- 'a`>Chapter 3. Iso-2-Cytochrome c and a Iso-1/Iso-2 Composite Mutant Protein^68Figure 3.19: A plot of the overall average deviations of main chain atoms of reduced yeast iso-1(Louie & Brayer, 1990: PDB file lYCC), reduced tuna (Takano & Dickerson, 1981a: PDB file5CYT), oxidized horse (Bushnell et al., 1990) and oxidized rice (Ochi et al., 1983: PDB file1CCR) cytochromes c from those of yeast iso-2-cytochrome c along the course of the polypeptidechain. Only residues 1 to 103, which are common to all five cytochromes c, are represented.The vertical bars represent the range of individual pairwise average deviations of main chainatoms between yeast iso-2 and the other cytochromes c. Structures with pairwise deviationsboth larger than 0.7 A and 30% greater than the overall average deviation of all structures arelabeled at the particular residue involved (T, tuna; H, horse; R, rice cytochromes c). The peakat position 37 is not labeled because all of the structures deviate significantly from iso-2-cyto-chrome c.(Kassner, 1973). The increase in reduction enthalpy in the iso-2 and B-2036 proteins may bedue to a decrease in the polarity of the heme environment as a result of the structural changesresulting from the substitutions Va120Ile and Leu98Met. The structural features that influencethe entropy of reduction are less well defined. Previous studies have shown that the reducedcytochrome c structure is more compact and thermally less mobile (Eden et al., 1982; Trewhellaet al., 1988; Berghuis & Brayer, 1992). Mutations that alter the packing of the hydrophobiccore may also increase the entropy of reduction.Chapter 3. Iso-2-Cytochrome c and a Iso-1/1 -so-2 Composite Mutant Protein^69300.0295.0290.00C0Oa_C0 285.0280.0275.010.0^15.0^20.0^25.0^30.0^35.0^40.0Temperature (°C)Figure 3.20: The midpoint reduction potential of the iso-2 (0), B-2036 (0) and wild-type iso-1(A) cytochromes c as a function of temperature. Data for iso-1-cytochrome c is from Rafferty,(1992).Table 3.19: Electrochemical properties of the yeast iso-2, B-2036 and iso-l-cytochromes c0 0 0Protein En, OH AScyt AG(mV) (kcal/mol) (e.u.) (kcal/mol)Iso-2 286 ± 2 -15.9 ± 0.4 -16 ± 1 -6.6 ± 0.1B-2036 288 ± 2 -16.5 ± 0.4 -17 ± 1 -6.7 ± 0.1Iso-1 290 ± 2 -14.0 ± 0.2 -9.1 ± 0.4 -6.7 ± 0.150Experimental conditions were 25 °C, pH 6.0, it = 0.1 M and SHE reference. Data for iso-l-cytochrome c is from Rafferty et al., (1990).Chapter 3. Iso-2-Cytochrorne c and a Iso-1/Iso-2 Composite Mutant Protein^703.9 Stability of Yeast Cytochromes cThe thermal transition temperature (T in ) of iso-l-cytochrome c has been measured to be 52.3 ±0.3 °C (Dumont et al., 1990). Iso-2-cytochrome c is slightly more stable with a Tni of 54.2 ±0.3 °C, while the Tin of the B-2036 protein is 48.9 ± 0.3 °C. Recently, a study of cavity creatingmutations in lysozyme found that the free energy of folding increases with the size of a residentcavity and this can be expressed as 24 to 33 cal mot -1 A -3 (Eriksson et al., 1992). Thus, thefilling of the buried cavity by Met98 without additional strain in the iso-2 structure would beexpected to increase this protein's stability relative to iso-1 by about 0.8 kcal mot -1 and maypartially account for the increase in Tin observed. The steric clash between residues 11e20 andCys102 in B-2036 cytochrome c and the resulting enlargement of the buried cavity may accountfor the decrease in thermostability of this protein.The structural studies described here also identify further individual amino acid substi-tutions resulting in alterations in the folding and packing of the three-dimensional structureof cytochrome c. Of the 16 amino acid substitutions between iso-1 and iso-2-cytochromes c,five of these cause changes in interactions between well ordered components of the structure:Va120Ile, His26Asn, Ala43Val, Leu98Met and Cys102Ala, and may significantly alter proteinstability. Two other substitutions, Leu58Lys and Gly83Ala, affect less well ordered sections ofpolypeptide chain. The remaining 11 substitutions do not appear to alter how the polypeptidechain is folded or those interactions that maintain that fold. All 11 of these latter substitutionsoccur at the protein surface and are not expected to alter protein stability.The B-2036 composite protein may be an example of an intermediate of protein evolutionsince it is the product of a non-allelic recombination. Although the overall fold is stronglypreserved, the composite is less stable than either of the wild-type proteins. This loss of ther-modynamic stability appears to be a result of incongruous amino acid substitutions. Oneexample described here is the substitution Va120Ile which appears to destabilize the proteinif the structurally complementary substitution Cys102Ala is not also present, as in the iso-2Chapter 3. Iso-2-Cytochrome c and a Iso-1/Iso-2 Composite Mutant Protein^71structure. Systematic substitution of amino acid residues that differ between iso-1 and iso-2-cytochromes c will better define structural determinants of both thermal stability and themidpoint reduction potential. Those amino acid substitutions that cause structural perturba-tions of the well ordered sections of the polypeptide chain, as described here, are particularlygood candidates for study by site-directed mutagenesis. Another direction of study is to examinethe structural effects of replacing secondary structural elements of cytochrome c (eg. Q-loops)with those of related proteins to further understand how large components of polypeptide chainare assembled and packed on the core of the protein.Chapter 4Structure - Function Analyses of Omega Loop A Replacements in YeastIso-1-Cytochrome c4.1 Experimental ProceduresQ-loop swap mutagenesis and subsequent protein purifications were performed by Dr. J. Fetrow(State University of New York). An alignment defining the primary sequence of the Q-loop swapproteins RepA2 and RepA2(Va120) with respect to those of yeast iso-l-cytochrome c and Rhodo-spirillum rubrum cytochrome c2 is given in Table 4.20. RepA2 crystals were produced by hairseeding liquid-liquid diffusion capillaries with crushed iso-1-cytochrome c crystals. Crystalsof yeast RepA2(Va120) cytochrome c were initially grown by hair seeding hanging drops withcrushed yeast iso-1-cytochrome c crystals as previously described (Leung et al., 1989). Crystalsused for diffraction analyses were subsequently produced by macro-seeding hanging drops withsmall RepA2(Va120) crystals. For both proteins, the final crystallization conditions were 95%(NH4)2504, 0.07 M Na2S204 and 0.1 M sodium phosphate buffer adjusted to pH 5.5. Crys-tals used in diffraction analyses were first transferred into a freshly prepared solution of 97%(NH4)2SO4, 0.1 M sodium phosphate buffer at pH 5.5, and 0.04 M dithiothreitol, to maintainthe reduced state of the heme group. These were then mounted in thin wall glass capillaries.A single RepA2 crystal was utilized to collect diffraction data to 2.25 A resolution (91%complete) on a Rigaku R-AXIS II imaging plate detector system. Each frame had a 0 oscillationof 1.0° and was exposed for 15 minutes to CuKa X-rays from a rotating anode generatoroperating at 80 mA and 50 kV. A total of 99 frames were collected after 25 hours of X-rayexposure time. The resulting images were processed to structure factor amplitudes and themerging R-factor for both full and partial intensities was 7.3%. Other data collection parameters72Chapter 4. Replacements of Q-Loop A in Iso-l-Cytochrome c^ 73Table 4.20: Sequence alignment of the yeast iso-1, RepA2 and RepA2(Va120) cytochromes cwith cytochrome c2 from Rhodospirillum rubrum-5^1^10Iso-i^TEFKAGSAKKGATLFKTRCLQCRepA2RepA2(Va120)R. rubrum - - - - EGDAAAGEKVSK - KCLAC20^30H TVEKGGPHKVGPNL HG IFDQ^ANDQ ANH TFDQGGANKVGPNL HGV40^50^ 60^70Iso-1^FGRHSGQAEGYSYTDAN I K K NV - - - LWDENNMSEYLTNPR. rabrumFENTAAHKDNYAYS ESYTEMKAKGLTWTEAN L AAYVKNP80^90^100Iso-1^KKYIP^ GTKMAFGGLKKEKDRNDLITYLKKACER. rubrum KAFVLEKSGDPKAKSKMTF - KLTKDDE I ENVIAYLKTLK -The primary sequences of yeast iso-1 (Smith et al., 1979) cytochrome c and R. rubrum (Duset al., 1968) cytochrome c2 have been aligned based on a comparison of their superimposed three-dimensional structures. The amino acid residues of the RepA2 and RepA2(Va120) mutants aregiven only where they differ from the yeast iso-1 sequence. The box encloses the sequence of1k-Loop A.are listed in Table 4.21. The absolute scale of the resultant structure factors was derived by alinear rescale against a wild-type iso-l-cytochrome c data set having the same resolution range.The overall thermal factor of 20 A 2 was derived by inspection of a Wilson (1942) plot.From a single RepA2(Va120) crystal, a complete diffraction data set to 1.9 A was collectedon an Enraf Nonius CAD4-F11 diffractometer using continuous Q scans of 0.6° at 0.55°/min.The incident CuKc radiation was Ni filtered and generated from a X-ray tube operated at26 mA and 40 kV. Intensity backgrounds were measured by extending the scans by 25% oneither side. Three standard reflections were measured every 8 hours of X-ray exposure timeto monitor crystal decay and slippage. Diffraction intensities were corrected for backgroundradiation, absorption, decay, Lorentz and polarization effects as described in Section 2.1.3.Crystal specific data collection statistics are presented in Table 4.21. The initial scale andoverall thermal factor (16 A 2 ) were determined as described for RepA2 cytochrome c.Chapter 4. Replacements of Q-Loop A in Iso-l-Cytochrome c^ 74Table 4.21: Data collection statistics of RepA2 and RepA2(Va120) cytochromes cParameter RepA2 RepA2(Va120) Iso-ltCrystal size (mm) 0.1 x 0.1 x 0.05 0.48 x 0.38 x 0.2Space group P43212 P43212 P43212Unit cell parameters (A):a, b 36.38 36.37 36.46c 137.63 137.64 136.86Resolution (A) 2.25 1.9No. of reflections measured 25731 9290No. of unique reflections 4498 7918t The cell dimensions and space group of wild-type yeast iso-l-cytochrome c are provided onlyfor comparison purposes (Louie & Brayer, 1990).A starting model of RepA2(Va120) cytochrome c was constructed from the coordinates ofyeast iso-l-cytochrome c (Louie & Brayer, 1990). At the position of each differing amino acid,the residue present was initially modeled as an alanine (Table 4.20). A sulfate ion and a selectionof 52 water molecules of the iso-l-cytochrome c structure were included in the initial model.The model was then refined by a restrained parameter least-squares method (Hendrickson &Konnert, 1981) using data with F > 2o- (F). Fragment deleted, F0 — Fe , 2F,— F, and 3F, — 2Femaps were examined at intervals during the course of refinement as a guide for the additionand manual adjustment of side chains in the structure. Water molecules were added to therefinement model if at least one hydrogen bond was formed to the existing structure and if2F, — I', or 3F, — 2F, maps had significant electron density at the water position. At the endof refinement, 74 water molecules and one sulfate ion (average solvent thermal factor of 34 A 2 )were included into the model and the conventional R-factor was 19.1%. The final refined overallthermal factor was 22 A 2 .The final RepA2(Va120) structure was used as an initial model for the refinement of theRepA2 structure, with residue Va120 converted to an alanine by deletion of the coordinatesof the two 7-carbons. The same selection of 52 water molecules and a sulfate ion from theChapter 4. Replacements of a-Loop A in Iso-l-Cytochrome c^ 75iso-1 structure were used as the initial solvent model. This model was refined as described forRepA2(Va120) with the eventual inclusion of the Phe20 side chain to a final R-factor of 18.3%.The final RepA2 structure contained 51 water molecules and a comparable sulfate ion to thatfound in the iso-1 and RepA2(Va120) cytochromes c (overall solvent thermal factor of 27 A 2 ).The final refined overall thermal factor was 17 A2 .The agreement of the RepA2 and RepA2(Va120) structures with ideal stereochemical ge-ometry and the corresponding refinement weights used in the final refinement cycles of eachare outlined in Table 4.22. To estimate coordinate errors, a plot of the conventional R-factoras a function of resolution along with theoretical curves calculated by assuming various r.m.s.coordinate errors (Luzzati, 1952) was constructed (Figure 4.21). Inspection of this plot suggestsTable 4.22: Final stereochemistry of RepA2 and RepA2(Va120) cytochromes cStereochemicalparameterr.m.s. deviation fromideal valuesRepA2^RepA2(Va120)RestraintweightDistances (A)Bond 0.018 0.018 0.020Angle 0.042 0.039 0.030Planar 1-4 0.053 0.048 0.050Plane (A) 0.013 0.015 0.020Torsional (deg.)Planar 2.3 2.5 2.5Staggered 24.5 23.9 20.0Orthonormal 16.5 13.8 15.0Chiral volume (A 3 ) 0.210 0.213 0.160Non-bonded contact (A)tSingle torsion 0.217 0.209 0.250Multiple torsion 0.195 0.195 0.250Possible hydrogen bonds 0.224 0.192 0.250t 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. Replacements of C1-Loop A in Iso-l-Cytochrome c^ 76Resolution10.0 A^5.o A^3.3 A^2.5 A^2.0 A^1.7 A1.00.9 s...0O^ IL^0.8 joC0.700^^^ 47-00.6 au_0.30-^ 0.24 A0.25-0"0 0.20-I0.15-0.100.20 A0.16 A0.05^0.10 0.15^0.20SIN(0) / A0.25^0.30Figure 4.21: A plot of the conventional R-factor as a function of resolution at the end ofrefinement. The fit to the data of RepA2 (Q) is represented by the thick line and the fit to theRepA2(Va120) data (^) is represented by a thin line. The theoretical dependence of R-factor onresolution, assuming various levels of r.m.s. error in the atomic positions of the model (Luzzati,1952), are drawn in dashed lines. The fraction of data used in refinement is presented above,using the same data point and curve scheme.a coordinate error for both structures of approximately 0.20 to 0.24 A. This value is a statisticalestimate of the overall coordinate error, whereas the actual positional error of an individualatomic coordinate will depend on the location of the atom in the structure. The error of atomsin the core of the protein is likely to be less, whereas highly mobile atoms on the protein surfaceare less accurately determined. A further analysis by an individual atom method (Cruickshank,1949; Chambers & Stroud, 1979) suggests r.m.s. coordinate errors of 0.16 A for both RepA2Chapter 4. Replacements of Q-Loop A in Iso-l-Cytochrome c^ 77and RepA2(Va120) cytochrome c.4.2 Results4.2.1 Comparison of the Structures of RepA2, RepA2(Va120) and Iso-1-Cyto-chromes ca-Loop A is located adjacent to the heme prosthetic group and is an integral component of theheme pocket (Figures 4.22 and 4.23). Part of this loop is adjacent to the solvent exposed edgeof the heme and may be involved in the binding of redox partners. The structural differencespresent were examined by the superposition of the main chain and heme atom coordinates ofRepA2 and RepA2(Va120) cytochromes c onto those of iso-1-cytochrome c by a least-squaresfit. The two amino terminal residues (Thr-5 and Glu-4) were not included in this processsince they are substantially disordered. The average main chain positional deviation betweenthe three superimposed structures was computed on a per residue basis and plotted as shown inFigure 4.24. The average deviations of the 11-Loop A main chain atoms from the iso-1 structureare 0.21 A and 0.16 A for the RepA2 and RepA2(Va120) cytochromes c, respectively. Both ofthese average deviations are similar to the overall averages for all main chain atoms which are0.22 A and 0.17 A for the two mutant proteins, respectively (Figure 4.24).Other than the disordered amino terminal residues —5 and —4, the region of Figure 4.24with the most significant deviation is centered at position 44. The structures of RepA2 andiso-l-cytochromes c in this region are plotted in Figure 4.25. Note that as a result of S2-Loop Asubstitution, residue 26 is changed from a histidine to an asparagine in the RepA2 protein. Thehistidine side chain in iso-l-cytochrome c forms two hydrogen bonds to Asn31 N and Glu44 0bridging these two sections of polypeptide chain. The OD1 atom of asparagine 26 in RepA2is able to occupy the same location as His26 ND1 in iso-1 and forms the first hydrogen bond.However, due to a shorter side chain length, Asn26 ND2 cannot occupy the same location asthe original His26 NE2 atom of the iso-1 protein. As a result, the polypeptide chain from Gln42to G1y45 shifts so that G1u44 0 is positioned within hydrogen bonding distance.Chapter 4. Replacements of 52-Loop A in Iso-l-Cytochrome c^ 78Figure 4.22: A space-filled representation of yeast iso-l-cytochrome c. The atoms of the centralheme group are shown as black spheres. All other atoms are colored white except those ofa-Loop A (residues 18 to 32) which are shown as grey, with Va120 and His26 highlighted inblack.Chapter 4. Replacements of CI-Loop A in Iso-l-Cytochrome c^ 79Figure 4.23: A stereo plot of iso-l-cytochrome c with C2-Loop A drawn with thick lines. Everytenth residue is labeled with the one letter amino acid code. The orientation of this plot issimilar to that of Figure 4.22.A further analysis of the positional deviation of well ordered side chains with average thermalfactors less than 25 A2 reveals three side chains with different conformations between RepA2and iso-1-cytochromes c. These residues are Tyr97, Leu98 and A1a101 all of which pack againstthe side chain at position 20. As can be seen in Figure 4.26, the side chain of Phe20 in theRepA2 structure is positioned in a direction defined by the vector between atoms Va120 CBand Va120 CG2 of the wild-type iso-l-cytochrome c. This conformation has the phenyl ringdirected towards the protein surface. While the presence of a phenylalanine at position 20does not markedly perturb the main chain atoms of C/-Loop A (Figure 4.24), the side chainatoms of Tyr97, Leu98 and Ala101 are displaced on average by 0.8, 0.7 and 0.7 A, respectively(Figure 4.26).In the RepA2(Va120) structure, Leu98 is one of the residues that line an internal hydrophobicburied cavity large enough to contain a solvent molecule. A similar cavity is observed in theiso-l-cytochrome c structure (Louie & Brayer, 1990). The structural and functional role of thiscavity is unknown, but amino acid substitutions in the hydrophobic core of the protein havebeen shown to affect the size of this cavity (see Section 3.4). The RepA2 structure does notChapter 4. Replacements of a-Loop A in Iso-l-Cytochrome c^ 80,^AAIlliMPLOINSISEVAINMEMMI IMERMANIMINIIMEMT '^fit^Vim0.015^25^35^45^55^65^75^85^95Residue NumberFigure 4.24: The average positional deviations of main chain atoms between RepA2 andwild-type yeast iso-1 (thick line), RepA2(Va120) and wild-type yeast iso-1 (thin line), andRepA2 and RepA2(Va120) (dotted line) cytochromes c along the course of the polypeptidechain. The horizontal lines represent the overall average deviations for the main chain atomsof residues (-3) to 103 for each comparison: RepA2 and isol, 0.22 A; RepA2(Va120) and iso-1,0.17 A; RepA2(Va120) and RepA2, 0.18 A.contain an internal cavity observable with a 1.4 A radius probe. The loss of this well definedcavity appears to be a result of the repositioning of the Leu98 side chain. None of the atoms ofthe Phe20 side chain occupy the space of the original Va120 CG1 methyl group which is buriedin the iso-1 structure. Movement of the side chain of Leu98 in the RepA2 structure, however,partially fills the space vacated by Va120 CG1.The substitution of valine for phenylalanine at position 20 in the RepA2(Va120) proteineliminates most of the structural changes described above. Residues Va120, Tyr97, Leu98and Ala101 are all positioned in a manner similar to that observed in the iso-1 structure(Figure 4.26). Furthermore, in the RepA2(Va120) structure, an internal cavity comparable tothe wild-type iso-l-cytochrome c structure is present. However, as observed in the RepA21.00.8O• 0.60a)0(1)-)O 0.4N0.2Pro 30Pro 30Gln 42-'•••—iGlu 44Ala 43Asn. 31^ Asn 31Gly 45His 26Phe 20Val 20Phe 20Val 20Chapter 4. Replacements of 11-Loop A in Iso-l-Cytochrome c^ 81Figure 4.25: The structures of wild-type yeast iso-1 (thin lines), RepA2 (thick lines) andRepA2(Va120) (medium lines) cytochromes c are plotted in the region of residue 26. In theiso-1 structure, His26 bridges these two segments of polypeptide chain by forming two hydrogenbonds (dashed lines). In the RepA2 and RepA2(Va120) structures, the main chain surroundingG1u44 shifts to form the same two hydrogen bonds to the shorter replacement Asn26 side chain.Leu 98Figure 4.26: Superimposed structures showing the packing of residue 20 against the nearbycarboxy terminal a-helix in wild-type yeast iso-1 (thin lines), RepA2 (thick lines) andRepA2(Va120) (medium lines) cytochromes c. In the RepA2 structure, three side chains (Tyr97,Leu98, A1a101) are repositioned to accommodate the larger phenylalanine side chain at position20.Chapter 4. Replacements of Q-Loop A in Iso-l-Cytochrome c^ 82structure, the substitution of asparagine for histidine at position 26 results in similar shifts ofthe main chain atoms of residues G1n42 through G1y45 (Figure 4.24).4.2.2 Electrochemistry of RepA2 and RepA2(Va120) Cytochromes cDirect electrochemistry of RepA2 and RepA2(Va120) cytochromes c appears to be reversiblewith a peak to peak separation of 62 ± 7 mV for all measurements, over the temperature rangefrom 10 to 25°C. However, measurements above 25 °C could not be made since protein solutionsbecame turbid resulting in greatly increased peak to peak separation in the voltammogramsand a loss of reversibility. A plot of the temperature dependence of the midpoint potentialof these two Q-loop replacement proteins and iso-1-cytochrome c is presented in Figure 4.27.Thermodynamic parameters derived from this data are listed in Table 4.23 along with midpoint300.05 290.0-EaC(1)00_0a280.0-270.050 10.0^15.0^20.0^25.0Temperature (°C)30.0Figure 4.27: The midpoint reduction potential of the RepA2 (0), RepA2(Va120) (0) andwild-type iso-1 (,A) cytochromes c are plotted as a function of temperature. Data for iso-l-cyto-chrome c is from Rafferty (1992).Chapter 4. Replacements of 12-Loop A in Iso-1-Cytochronle c^ 83Table 4.23: Electrochemical properties of the yeast iso-1, RepA2, and RepA2(Va120) cyto-chromes c^0^0^0Protein^Em^OH AScyt AG(mV)^(kcal/mol)^(e.u.)^(kcal/mol)Iso-1 290 2 -14.0 ± 0.2 -9.1 ± 0.4 -6.7 ± 0.1RepA2 271 ± 2 -18.2 ± 0.7 -24 ± 2 -6.3 ± 0.1RepA2(Va120) 290 ± 2 -15.2 ± 0.1 -13.0 ± 0.2 -6.7 ± 0.1Experimental conditions were 25 °C, pH 6.0, it = 0.1 M and SHE reference. Data for iso-1-cytochrome c is from Rafferty et al., 1990.potentials under standard conditions. The midpoint potential for the RepA2(Va120) proteinat 25 °C is the same as that observed for wild-type iso-l-cytochrome c within experimentalerror. In contrast, the midpoint potential of RepA2 differs markedly and is 19 mV lower thanthat of the wild-type cytochrome c. The enthalpy and entropy of reduction of the RepA2and RepA2(Va120) proteins are less than that of wild-type iso-1-cytochrome c and partiallycompensate for each other.4.3 Discussion4.3.1 Comparison of the Wild Type Yeast Iso-1 and It. rubrum StructuresOverall, the amino acid sequence identity between yeast iso-l-cytochrome c and R. rubrum cyto-chrome c2 is 38% as shown by the sequence alignment in Table 4.20. A comparison of the aminoacid sequences of their respective 12-Loop A segments (residues 18 to 32) reveals that 10 of the15 residues present are identical (67% of the total). There are 18 additional residues from otherpolypeptide chain segments that contact S2-Loop A in iso-l-cytochrome c. Nine of these (50%)are identical with those of R. rubrum cytochrome c2. The main chain and heme coordinatesof R. rubrum cytochrome c2 can be superimposed onto the iso-1 structure with an averagedeviation of 1.1 A, excluding residues 57 and 76 which differ greatly in conformation betweenHis 26 His 26Val 20His 18 Leu 32Lys 22^ Lys 22Val 28^ Val 28Chapter 4. Replacements of Q-Loop A in Iso-l-Cytochrome c^ 84Figure 4.28: A stereographic plot of a-Loop A from yeast iso-l-cytochrome c (thick lines)superimposed on that of cytochrome c9 from R. rubrum (thin lines). The overall path of thepolypeptide chain is similar in both loops, but the precise conformation differs. The termini ofthis loop, His18 and Leu32, and every fifth residue of yeast iso-1-cytochrome c are labeled.these two proteins. The Q-Loop A portion of these superimposed structures is illustrated inFigure 4.28. If only the main chain atoms for Q-Loop A are considered, the average deviationobserved is 0.53 A.Many of the residues of a-Loop A conserved between the yeast iso-1 and R. rubrum cy-tochromes form intra-loop hydrogen bonds which stabilize this folded structure (Table 4.24).Note that two of the four intra-loop hydrogen bond differences listed in Table 4.24 are equiv-alent since the hydrogen bond between the side chain of residue 26 and the main chain amideof residue 31 is conserved despite a change in amino acid sequence. The other two hydrogenbond differences are a result of an amino acid difference at position 21 and the absence of ary-turn in the R. rubrum structure. A further understanding of the structural heterogeneity inC2-Loop A can be found by examining the interaction of this loop with the remainder of theprotein molecule. Hydrogen bond interactions involving only one loop atom are more varieddue to amino acid substitutions occurring at positions outside of a-Loop A (Table 4.24). Sub-stitutions at positions outside of S2-Loop A particularly affect the packing of buried loop sidechains.One prominent example of packing differences in a side chain group between the yeast iso-1and R. rubrum cytochromes involves residue 20 which forms part of the hydrophobic core ofChapter 4. Replacements of a-Loop A in Iso-l-Cytochrome c^ 85Table 4.24: Hydrogen bond interactions of a-Loop A in yeast iso-1, RepA2, RepA2(Va120)cytochromes c and R. rubrum cytochrome c2Cytochrome^Intra-a-Loop A interactions^Involving other protein residuesYeast iso-1His18 ND1 - 30 0Thr19 0G1 - 25 0Thr19 OG1 - Asn31 ND220 N - G1u21 OE118 N - 14 0His26 NE2 - 44 0Lys27 NZ - 15 029 N - 17 024 N - 21 027 N - 29 031 N - His26 ND1Asn31 ND2 - Thr19 OG1Asn31 ND2 - 21 032 N - 19 033 N - Asn31 OD1His33 ND1 - 20 035 N - 32 0Tyr46 OH - 28 0Yeast RepA2e 20 N - Glu21 OE1 Asn26 ND2 - 44 0e31 N - Asn26 OD131 N - His26 ND1e His26 NE2 - 44 0Yeast RepA2(Va120)e 20 N - Glu21 OE1 Asn26 ND2 - 44 031 N - Asn26 OD1 e His26 NE2 - 44 0e 31 N - His26 ND1 e 33 N - Asn31 OD1R. rubrume 20 N - Glu21 OE1 ED Asn26 ND2 - Asn45 ND2te 27 N - 29 0 e His26 NE2 - 44 031 N - Asn26 OD1 28 N - 17 0e 31 N - His26 ND1 e His33 ND1 - Asn31 OD1Lys43 NZ - 31 0ED Lys43 NZ - Asn31 OD1C2-Loop A is comprised of the amino acids from His18 to Leu32. Where main chain atomsare involved only the sequence number and atom name are given. All hydrogen bonds arelisted for wild-type iso-l-cytochrome c. For the other three structures, the symbol 6 indicatesan additional hydrogen bond in a particular structure whereas e indicates the absence of ahydrogen bond relative to the iso-1 structure. Hydrogen bonds are defined by the followingcriteria: a H...A distance < 2.60 A. 2.70 A or 3.05 A where A is an oxygen, nitrogen or sulfuratom respectively; a D-H...A angle > 120°; and, a C-A...H angle > 90°.tIn the structure determined for R. rubrum cytochrome c2 (Salemme et al., 1973), if the sidechain amide group of Asn45 is rotated 180°, the OD1 atom would be positioned within hydrogenbonding distance.Chapter 4. Replacements of a-Loop A in Iso-l-Cytochrome c^ 86both molecules. This residue is a valine in the iso-1 structure and a much larger phenylalanine inR. rubrum cytochrome c2. The larger side chain of the R. rubrum protein is accommodated inthe hydrophobic core by the complementary substitution of PhelOSer (Table 4.20). In addition,the presence of Phe33 eliminates a hydrogen bond to the carbonyl oxygen of residue 20 formedin the iso-1 structure by His33 ND1. These factors serve to optimize the fit of the phenyl groupof residue 20 into the hydrophobic core of the R. rubrum protein.Another major packing difference involves residue 26. In one instance, the hydrogen bondbetween the side chain of residue 26 and the main chain of residue 31 is in fact conservedbetween the two proteins. However, the interaction of this residue with the nearby polypeptidechain segment (involving primarily residues 44 and 45) is substantially different. In the iso-1structure, a hydrogen bond is formed between His26 NE2 and Glu44 0. In the R. rubrumstructure, Asn26 ND2 interacts with the side chain of Asn45 leading to a different folding ofthe polypeptide chain in this area.4.3.2 Structural Effects of Q-Loop A ReplacementsIn RepA2 cytochrome c, Q-Loop A from R. rubrum has been spliced into the yeast iso-l-cytochrome structure in place of the normally resident residues. An examination of this struc-ture reveals that the new Q-Loop A has an overall main chain conformation very similar tothat observed for the original loop in the iso-1 structure (Figure 4.24). In some cases, suchas the Q-Loop A amino acid differences at positions 21 and 22 on the surface of the protein,these substitutions are easily accommodated into the yeast iso-1 structure without any ob-vious structural perturbations. However, in other cases, such as the side chain substitutionsat positions 20 and 26, which occur in the molecular interior where these residues form closemolecular interactions, the a-Loop A substitution has had a significant structural effect. Forexample, if the R. rubrum loop is inserted without adjustment of the new phenylalanine sidechain at position 20, a steric clash with the side chain of Phe10 results. This conflict is relievedin the RepA2 structure by redirecting the Phe20 side chain towards the molecular surface byChapter 4. Replacements of a-Loop A in Iso-1-Cytochrome c^ 87a 120° rotation about the xi torsional angle of this residue. Nonetheless, this reorientation isnot entirely sufficient to avoid disruption of the packing in the hydrophobic core of the yeastiso-l-cytochrome c structure and as a consequence the side chain of Leu98 shifts to fill some ofthe internal empty space generated by the Va120Phe substitution. At the molecular surface, thelarge phenyl group at position 20 in the RepA2 mutant also displaces the side chains of Tyr97and Ala101 which pack against the original Va120 side chain (Figure 4.26). The end resultis that residues not part of a-Loop A, along with the Phe20 side chain, undertake positionaladjustments which allow the swapped loop to have a main chain conformation nearly identicalto the loop it replaces.In the RepA2(Va120) mutant, residue 20 has been substituted back to a valine as in thewild-type iso-l-cytochrome c. The side chain of Va120 in the RepA2(Va120) structure is in aconformation comparable to that of the iso-1 protein (Figure 4.26). As a result, this substitutionrestores the packing of the hydrophobic core and does not lead to the displacement of Tyr92,Leu98 or Ala101.Although the substitution of asparagine for histidine at position 26 in both loop replacementproteins is conservative, as Figures 4.24 and 4.25 show, it does result in alterations in the mainchain fold centered at the adjacent residue G1u44. An analysis shows that the main chainatomic positions of residue 26 appear to be more tightly fixed by hydrogen bond interactionswith other groups than residues 42 through 45 which occupy a more solvent exposed position.As a result, this latter segment of polypeptide chain has greater flexibility to shift position inorder to maintain a hydrogen bond between the side chain of residue 26 and the carbonyl groupof residue 44.It is of interest that in yeast iso-2-cytochrome c there is also an asparagine at position 26(see Section 3.3). Although the polypeptide chain in the region of residue 44 is slightly displacedrelative to that of the iso-1 structure, the hydrogen bond between Asn26 ND2 and G1u44 0 isnot formed (Table 3.14). This appears to be a consequence of the replacement of alanine atthe adjacent residue 43 by a valine which fills the space vacated as a result of the substitutionChapter 4. Replacements of a-Loop A in Iso-l-Cytochrome c^ 88of the smaller asparagine side chain for the histidine side chain at position 26. This secondsubstitution therefore appears to prevent the type of displacement observed in the polypeptidechain of RepA2 and RepA2(Va120) cytochromes c which is required to form a hydrogen bondto the carbonyl group of G1u44.4.3.3 In Vivo Functional Consequences of a-Loop A ReplacementsData concerning the in vivo function and further measurements of cytochrome c levels in intactcells were completed and kindly provided by Dr. J. Fetrow (State University of New York). Invivo function was assessed by measuring the rate of growth of yeast strains containing wild-type, RepA2 and RepA2(Va120) as well as a strain containing no cytochrome c (Figure 4.29).The substantive difference in functional activity of RepA2 at various temperatures could becaused by a decreased level of intact cytochrome c present within the cell. This possibilitywas examined by determining the levels of holo-cytochrome c present within intact cells bylow temperature spectroscopy (Figure 4.30). Note that this spectroscopic technique can onlydetermine levels of holoprotein, that is, protein with a heme group covalently attached. Levelsof apoprotein are not determined using this technique.In vivo, low temperature spectroscopy shows that the amount of RepA2(Va120) holoproteinpresent in yeast cells is slightly less than wild-type protein at 25, 30 and 37 °C; however, thefunction of these proteins is very similar at all three temperatures (Figures 4.29 and 4.30). Intotal. RepA2(Va120) cytochrome c differs from the wild-type yeast iso-1 protein by four aminoacids (Table 4.20). The RepA2 mutant, on the other hand, which differs from RepA2(Va120)cytochrome c by a single amino acid, displays very different functional behavior and appearsto be a temperature sensitive mutant. This protein is functional at levels comparable to thewild-type at 25 °C, somewhat less at 30 °C (Fetrow et al., 1989), and only at very low levelsat 37 °C. Low temperature spectroscopy indicates that there is less of this protein present inintact yeast cells at all three temperatures, but that the relative ratios are similar. Howeverthe RepA2 mutant does contain holoprotein in quantities sufficient to allow growth of cells in1000^A. 25°C100----100_00^7:30^15.00^22:30■,--e— e--Chapter 4. Replacements of C2-Loop A in Iso-l-Cytochronie c^ 89^0:00^7:30^15:00^22:30^30:00nf^co 1000 —^ C. 37°C100• 1 0120^0:00^7:30^15:00^22:30^30:00Culture Time (Hours)Figure 4.29: Growth curves of yeast strains in liquid lactate media containing RepA2 (•)RepA2(Va120) ), wild-type iso-1 (0) and no (A) cytochromes c are plotted. Growth curveswere determined by a variation of the method of Schweingruber et al. (1979). Data and figurewere kindly provided by of Dr. J. Fetrow, State University of New York.0.130.11870.10750.096250.085500 527.5 555 582.5 610O(0-2 0.285O610Chapter 4. Replacements of 12-Loop A in Iso-l-Cytochrome c^ 90CO0.3050.2950.28503cri^527.5^555^582.5^610Wavelength (nm)Figure 4.30: Spectra of intact yeast cells, both experimental and control strains, at —196 °C.Spectra of RepA2 (thick line), RepA2(Va120) (dotted line), wild-type iso-1 (thin line) and no(dashed line) cytochromes c are plotted. The peak at 550 nm (indicated with an arrow) is oneof the herne absorbance peaks in reduced cytochrome c. Spectra were recorded as described inFetrow et al. (1989). Data and figure were kindly provided by of Dr. J. Fetrow. State Universityof New York.oQL0U0 20-a)0La)0EH. 10a)00")La)Chapter 4. Replacements of a-Loop A in Iso-l-Cytochrome c^ 91liquid lactate media.The presence of significant amounts of RepA2 protein in intact cells suggests that themutation has not interfered with protein production. Cytochrome c in yeast is transcribed inthe nucleus, translated on a ribosome in the cytosol, and transported across the mitochondrialmembrane where the heme is covalently attached by heme lyase (Pettigrew & Moore, 1987).The messenger RNA, the apoprotein and the holoprotein of RepA2 cytochrome c appear to benearly as stable as those of iso-1-cytochrome c to degradation at higher temperature. The morelikely explanation for the reduced function of RepA2 cytochrome c at higher temperatures isthat it is a functionally less efficient protein.The average main chain thermal factors of a-Loop A in the RepA2 structure are significantlyhigher than in the RepA2(Va120) and iso-1 structures (Figure 4.31). The replacement of aminoacids at the protein surface combined with the increase in mobility of Q-Loop A in the RepA230020^22^24^26^28^30^32Residue NumberFigure 4.31: The average thermal factor of the four main chain atoms of each amino acid inRepA2 (thin line). RepA2(Va120) (dotted line) and wild-type iso-1 (thick line) cytochromes c areplotted as a function of residue number for Q-Loop A. The thermal factors have been normalizedover all protein atoms by an additive factor. The greatest main chain thermal factor differencebetween the RepA2 structure and the other two structures is within the 1-Loop A region.Chapter 4. Replacements of C2-Loop A in Iso-l-Cytochrome c^ 92structure may interfere with the binding of yeast cytochrome c with physiologically importantelectron transfer partners thereby diminishing cytochrome c function. This loss of function isultimately expressed in terms of slower cell growth.4.3.4 The Relationship Between Heme Reduction Potential and Temperature Sen-sitivityThe RepA2(Va120) cytochrome c has a midpoint potential comparable to that of wild-type yeastiso-1-cytochrome c (Table 4.23). It therefore appears that the amino acid substitutions presentin comparison to the original iso-1 Il-Loop A have had no effect on this functional parameter.This includes the Lys22Gln substitution which has caused a change in the surface charge ofthe molecule. However, the single substitution of valine for a phenylalanine at position 20 inthe RepA2(Va120) cytochrome c did restore the midpoint potential of the RepA2 cytochrome cto comparable values found for the wild-type iso-1 protein at 25 °C (Figure 4.27). Clearly,Va120 is important for the maintenance of the heme reduction potential in wild-type iso-1-cyto-chrome c. Nonetheless, the mechanism by which residue 20 modulates midpoint potential is notreadily apparent. Previous studies have linked a change in the solvent accessibility of the heme(Stellwagen, 1978; Louie et al., 1988b) or polarity of the heme environment (Kassner, 1973;Louie Sc Brayer, 1989) to a change in reduction potential. However, the solvent accessibilitiesof the porphyrin ring atoms of the heme are the same for RepA2, RepA2(Va120) and iso-l-cyto-chromes c. Furthermore, no other changes in the position of polar groups near the heme areobserved in either the RepA2 or RepA2(Va120) structures relative to the wild-type protein.The thermodynamic parameters derived from the temperature dependence of the midpointreduction potential may provide insight into the mechanism of reduction potential modulationby residue 20 (Table 4.23). The larger decrease in the entropy of reduction is partially offsetby a decrease in the enthalpic contribution. A change in the polypeptide contribution to thethermodynamic properties of cytochrome c reduction has been suggested to be primarily dueto electrostatic effects (Schejter et al., 1982), however, the substitution at position 20 does notChapter 4. Replacements of a-Loop A in Iso-l-Cytochrome c^ 93involve a change in net molecular charge. The change in enthalpy could be accounted for by theincrease in hydrophobic bulk of the protein (Kassner, 1973). Taniguchi et al. (1980) have sug-gested that polypeptide chain ordering and protein solvent interactions are also likely importantin determining the entropy of reduction. The Phe20 side chain is largely solvent accessible inthe RepA2 structure which may alter solvent ordering upon reduction. The increased mobilityof a-Loop A as seen in the temperature factor values (Figure 4.31) may also contribute to thiseffect. In addition, the increase in thermal factors may result in an increase in dynamic hemesolvent exposure.Chapter 5Structural Analysis of Mutants of an Invariant Proline in YeastIso-1-Cytochrome c5.1 Experimental ProceduresSamples of the Pro7lAla, Pro7lIle, Pro7lSer and Pro71Val cytochromes c were providedby Dr. F. Sherman (University of Rochester). Crystals of these mutant proteins were pro-duced by hair seeding liquid-liquid diffusion capillaries with crushed iso-l-cytochrome c crys-tals (Salemme, 1972; Leung et al., 1989). The final crystallization conditions for the Pro7lAla,Pro7lIle and Pro7lSer mutants were: 92% saturated (NH4)2SO4, 0.07 M dithionite and 0.1 Msodium phosphate buffer adjusted to pH 6.8. The Pro7lVal crystals were grown under simi-lar conditions: 90% saturated (NH4)2SO4, 0.01 M dithiothreitol and 0.1 M sodium phosphatebuffer adjusted to pH 6.0. Crystals used in diffraction analyses were transferred into a freshlyprepared solution of 95% saturated (NH4)2SO4, 0.1 M sodium phosphate buffer of the same pHas the crystallization buffer, and 0.04 M (0.01 M for Pro7lVal) dithiothreitol, to maintain thereduced state of the heme group.Each of the data sets listed in Table 5.25 was collected from a single crystal. Diffractometerdata sets were collected of the Pro7lAla, Pro7lSer and Pro71Val cytochromes c with an EnrafNonius CAD4 instrument using continuous S2 scans of 0.5 to 0.6° at 0.55°/min at an ambienttemperature of 15 °C. Standard reflections were measured every 8 hours of X-ray exposuretime to monitor crystal decay and slippage. Two different area detector systems were used tocollect two independent data sets of the Pro7lIle mutant. The first data set was collected witha Rigaku R-AXIS II imaging plate detector system. Each of the 60 frames was exposed for 15minutes to CuIC, X-rays from a rotating anode generator operating at 80 mA and 50 kV. The94Chapter 5. Replacements of an Invariant Proline in Iso-l-Cytochrome c^95Table 5.25: Data collection statistics for Pro71 mutant cytochromes cParameter Pro7lAla Pro7lIle(1) Pro7lIle(2) Pro7lSer Pro71ValCrystal volume (mm3 ) 0.04 0.01 0.005 0.03 0.06Space group P43212 P43212 P43212 P43212 P43212Unit cell parameters (A):a, b 36.45 36.45 36.54 36.44 36.45c 137.3 137.1 137.3 137.3 137.0Instrument CAD4 FAST R-AXIS CAD4 CAD4Resolution (A) 1.8 1.7 1.7 1.8 1.9No. of reflections measured 12376 25290 40082 15374 10215No. of unique reflections 9266 8899 8546 9295 7864Max. decay correctiont 0.81 0.68 0.68Max. absorption correctiont 0.54 0.62 0.61R-scale with iso-1t (%) 11.7 8.6 9.2 14.5 15.3tThis R-scale is defined as:2 E I Fo - E (F61 ±where Foi and F0J are the observed structure factor amplitudes of wild-type iso-l-cytochrome cand the data set being compared, respectively.t These corrections are handled by inter-frame scaling for area detector data sets.crystal was oscillated through a cb angle of 1.0° for each frame. Due to the high symmetry ofspace group P43212, a nearly complete data set could be collected with an overall 60° rotationof the angle 0. The second data set was collected with an Enraf Nonius FAST detector using thesame X-ray radiation. Data collection statistics for both data sets are presented in Table 5.25.The two Pro7lIle mutant data sets were then merged together giving an R-factor of 7.7% basedon 7246 duplicate intensities, to produce a 92% complete data set to 1.7 A having 10198 uniquereflections. The absolute scale of the structure factors of each mutant diffraction data set wasderived by a linear rescale against an iso-l-cytochrome c data set having the same resolutionrange.Chapter 5. Replacements of an Invariant Proline in Iso-l-Cytochrome c^96The starting model of each mutant protein was constructed from the coordinates of yeast iso-1-cytochrome c (Louie & Brayer, 1990) by deleting all the side chain atoms except the 3-carbonat position 71 to create an alanine residue. A sulfate ion and a selection of 52 water moleculesof the iso-l-cytochrome c structure were also included in the initial model. The model wasthen refined by a restrained parameter least-squares method (Hendrickson & Konnert, 1981)using all data with F > 2a(F) and a resolution greater than 6 A. Fragment deleted, F0 — Fe,2F0 — I', and 3Fo — 2F, maps were examined at intervals during the course of refinement to fitthe complete side chain at position 71 and to adjust the conformations of other side chains in thestructure. Peaks along the protein surface in F, — I', maps were considered water molecules ifat least one hydrogen bond was formed to the existing structure and the refined thermal factorswere less than 55 A2 . The R-factor, number of solvent molecules, thermal factor statistics andstereochemistry of the final refined mutant structures are presented in Tables 5.26 and 5.27.Plots of the conventional R-factor as a function of resolution, along with theoretical curvescalculated by assuming various r.m.s. coordinate errors (Luzzati, 1952), indicate a r.m.s. errorof approximately 0.20 A for all four structures (Figure 5.32). If a complete model is assumed,the r.m.s. coordinate errors range from 0.11 A for Pro7lIle to 0.16 A for Pro7lVal using theCruickshank approach (Cruickshank, 1949; Chambers & Stroud, 1979).5.2 Results5.2.1 The Environment of Pro71 in Yeast Iso-1-Cytochrome cThe effects of amino acid replacements at position 71 in yeast iso-1-cytochrome c were analyzedthrough comparison with the wild-type protein (Louie & Brayer, 1990). In yeast cytochrome c,Pro71 is adjacent to the carboxy terminus of one a-helix and initiates another a-helix (Fig-ure 5.33). One of these is Helix III formed by residues 60 through 70 and the second helix isHelix IV consisting of residues 70 to 74 (Table 1.4). Pro71 prevents the elongation of the firsthelix because the pyrrole ring removes the amide functional group and prevents the formationof a hydrogen bond to the carbonyl group of Tyr67. Furthermore, a steric clash between theChapter 5. Replacements of an Invariant Proline in Iso-l-Cytochrome c^97Table 5.26: Final refinement statistics for Pro71 mutant cytochromes cRefinement Statistic Pro7lAla Pro7lIle Pro7lSer Pro7lValNo. of reflections 6868 9805 6197 5381R-factort(%) 18.9 18.6 18.3 18.9Overall thermal factor (A 2 ) 17.2 19.4 17.0 17.7No. of protein atoms 891 894 892 893No. of solvent molecules 73 80 67 61Avg. solvent thermal factor (A 2 ) 28.2 36.0 29.6 26.6t This R-factor is defined as:R = ^hkl liFohkri ohialTable 5.27: Final stereochemistry for Pro71 mutant cytochromes cStereochemicalrestraintr.m.s. deviation fromideal valuesRestraintweightPro7lAla Pro7lIle Pro7lSer Pro7lValDistances (A)Bond 0.019 0.019 0.019 0.019 0.020Angle 0.039 0.038 0.040 0.040 0.030Planar 1-4 0.051 0.056 0.051 0.052 0.050Plane (A) 0.015 0.016 0.016 0.015 0.020Chiral volume (A3 ) 0.209 0.199 0.200 0.230 0.180Non-bonded contact (A)tSingle torsion 0.219 0.211 0.225 0.218 0.250Multiple torsion 0.185 0.182 0.172 0.214 0.250Possible hydrogen bonds 0.217 0.190 0.220 0.260 0.250Torsional angles (deg.)Planar 2.4 2.7 2.4 2.4 2.5Staggered 21.4 20.3 23.5 22.8 20.0Orthonormal 19.0 19.3 20.2 20.6 20.0t 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.Resolution3.3 A^2.5 A^2.0 A^1.7 A,^ ,^f^i^r I1.4 A'^ 1.0-era)-0.9 L.-0.8 0- 0- 0.7 ti10.0 A^5.0 AI^I0.35-0.24 A0.30-0.20 A00.25 -130.20-00.15-' o •^I ^ /^/ /^.... I/^I /0.10^/t 1^t^t0.16 AI^'0Chapter 5. Replacements of an Invariant Proline in Iso-l-Cytochrome c^980.05^0.10^0.15 0.20 0.25^0.30 0.35SIN(0) / xFigure 5.32: A plot of the conventional R-factor as a function of resolution at the end ofrefinement. The fit to the data of the ■ Pro71Ala, ^ Pro71Ile, • Pro7lSer, 0 Pro71Val cyto-chrome c structure determinations are represented by solid lines. The theoretical dependenceof the R-factor on resolution, assuming various levels of r.m.s. error in the atomic positions ofthe model (Luzzati, 1952), are drawn with dashed lines. Above, the fraction of data used inrefinement for each structure is presented, using the same data point and curve scheme.Chapter 5. Replacements of an Invariant Proline in Iso-l-Cytochrome c^99Figure 5.33: The region near Pro71 in yeast iso-l-cytochrome c. Main chain atoms in theregion of Pro71 of yeast iso-l-cytochrome c are depicted as ribbons. The ribbon correspondingto Helices III and IV (Table 1.4) is shaded darker. Residues within 4 A of the Pro71 side chainare displayed and labeled. The heme moiety, the heme iron ligands (His18 and Met80) and theheme thioether linkages to Cys14 and Cys17 are also displayed for reference.Chapter 5. Replacements of an Invariant Proline in Iso-l-Cytochrome c^100pyrrole ring of Pro71 and the side chain of Asn70 would exist if these residues were in an a-helical conformation. Conversely, the same pyrrole side chain locks the main chain torsionalangle q5 near —60° initiating helical segment IV. In this way, the carbonyl group of residue 71interacts with 11e75 N forming the first hydrogen bond in Helix IV. It has been proposed thatone function of Pro71 is to direct the proper folding of the polypeptide chain in this region(Takano & Dickerson, 1981a; Louie et al., 1988a).The overall polypeptide chain fold of cytochrome c completely occludes the main chainand side chain atoms of Pro71 from bulk solvent. Thus, the side chain of Pro71 is in thehydrophobic core of this protein and packs against Tyr67, the Met80 side chain, Phe82 andGly83 (Figure 5.33). Of these residues, the heme iron ligand Met80 would be expected to bethe most functionally sensitive to amino acid replacements at position 71. Furthermore, thepolypeptide chain in the region of both Pro71 and Gly83 is observed to be more flexible whencytochrome c is in the oxidized state suggesting that these portions of the polypeptide chainmay be of functional importance in oxidation state dependent conformational changes (Berghuis& Brayer, 1992).5.2.2 Structural Differences of Pro71 Mutant Cytochromes cMatrices showing main chain distance differences and thermal factor differences of each position71 mutant structure from the values of wild-type iso-l-cytochrome c are presented in Figure 5.34.These analyses indicate that observed main chain structural differences are localized to tworegions. These are in the immediate vicinity of the mutation site and about residue 83. Inaddition, the four Pro71 mutant structures have been superimposed onto the wild-type structureby a least-squares fit of the main chain and heme atoms. The average deviations for all themain chain atoms of Pro7lAla, Pro71Ile. Pro7lSer and Pro71Val cytochromes c are presentedin Table 5.28. Like the difference matrices, these results indicate that the main chain fold isperturbed in the vicinity of Pro71 and Gly83 but is otherwise comparable to the wild-typeprotein. Analysis of mutant protein heme solvent accessibility and heme ligand geometry showsC4) 080 90 10010010^20^30^40^50^60^70^80 10^20^30^40^50 60^70Residue NumberResidue Number60^70^80^90^100 40^50^60^70^80^90^100Pro71 SerCa00 00.>0^I10090800_a 60EZ 50a)40a)CC 302010Chapter 5. Replacements of an Invariant Proline in Iso-l-Cytochrome c^101Figure 5.34: Positional and thermal factor difference matrices for the comparison of Pro7lAla,Pro7lIle, Pro7lSer and Pro71Val cytochromes c with the wild-type protein. The upper lefttriangular matrix contains the a-carbon distance differences and the lower right triangle containsthe average main chain thermal factor differences. Each triangular matrix is contoured at onestandard deviation intervals above the mean difference. The first five residues (-5 to —1) arenot included in the matrices due to positional disorder in this region. See Section 2.2.2 for acomplete description of difference matrix construction.Chapter 5. Replacements of an Invariant Proline in Iso-l-Cytochrome c^102Table 5.28: Positional deviations of groups in the vicinity of residue 71GroupPositional deviation (A)Pro7lAla Pro7lSer Pro7lIle Pro71ValA. Main chainAsn70 0.27 0.11 0.38 0.5771 0.10 0.24 0.34 0.45Phe82 0.31 0.32 0.28 0.17G1y83 0.75 0.84 0.44 0.68G1y84 0.40 0.45 0.19 0.47B. Side chainTyr67 0.19 0.10 0.16 0.22Met80 0.12 0.15 0.19 0.20Phe82 0.27 0.24 0.66 0.33C. All main chain atoms 0.15 0.16 0.15 0.19these to be within experimental error the same as those of wild-type iso-l-cytochrome c.The average main chain deviations of residues near the mutation site in all four mutantstructures are presented in Table 5.28. Only average deviations greater than two times theoverall average deviation are considered significant in the present discussion. The four mutantproteins can be divided into two groups based on the size and shape of the replacement sidechain. In the first group, Pro7lAla and Pro7lSer cytochromes c, the replacement side chainsare smaller in volume than the original proline side chain. In Figure 5.35, these two mutantstructures have been superimposed onto the structure of iso-l-cytochrome c and the regionabout Pro71 drawn. The main chain atom positions of residue 71 are unaffected by the alanineand serine replacements (Table 5.28), however, the space vacated by these smaller side chainsis partially filled by G1y84 0 as a result of the reorganization of the main chain atoms aboutG1y83. Despite these main chain shifts the packing of the hydrophobic core in the region ofresidue 71 is particularly open in the Pro7lAla structure. The CB atom in this structure ispositioned similar to that of the wild-type protein and does not shift to compensate for the lossChapter 5. Replacements of an Invariant Proline in Iso-l-Cytochrome c^103Figure 5.35: Stereo-diagram of the structures of Pro7lAla, Pro7lSer (thin lines) and wild-typeiso-1 (thick lines) cytochromes c. Substitution of smaller side chains at position 71 results inshifts about the main chain atoms of residues 82 to 84.of the other two side chain atoms present in the wild-type structure.The side chain of Ser71 adopts an unusual conformation to fit into the hydrophobic core.As shown in Figure 5.35, the Ser71 OG atom in Pro7lSer cytochrome c is in a position similarto that of the Pro71 CG atom in the iso-1 structure. The xi angle of residue 71 in Pro7lSercytochrome c is —17° as compared to —33° for the same residue in the wild-type protein.This xi angle in the Pro7lSer structure is distant from frequently observed angles for serineresidues (Ponder & Richards, 1987; James & Sielecki, 1983). A potential explanation for thisconformation is the formation of a new bifurcated hydrogen bond between the hydroxyl groupof Ser71 and the main chain carbonyl group of Tyr67.The second group of mutants, Pro71Ile and Pro71Val cytochromes c, possess side chainsthat are either larger or differently shaped than the normally resident Pro71. The valine sidechain, although constructed from the same number of atoms as a proline residue, has a ,Q-branch that results in a differently shaped side chain than the ring structure of a proline. Theisoleucine side chain is both /3-branched and larger, containing an additional methyl group.The structures of Pro7lIle and Pro71Val cytochromes c in the vicinity of the mutation site areplotted in Figure 5.36 along with the corresponding wild-type iso-l-cytochrome c structure.Chapter 5. Replacements of an Invariant Proline in Iso-l-Cytochrome c^104Figure 5.36: Stereo-diagram of the structures of Pro71Ile, Pro71Val (thin lines) and wild-typeiso-1 (thick lines) cytochromes c in the same orientation as Figure 5.35. Larger side chains atposition 71 displace the main chain atoms of residues 69 to 71 and 81 to 83.In the Pro71Val mutant, one of the 7-carbons is positioned near the corresponding atomof the proline residue in the wild-type structure. The second 7-carbon is directed into thehydrophobic core of the molecule. The xi torsional angle of Va171 is —40° and is within 20°of a frequently observed torsional angle for that residue (Ponder & Richards, 1987; James SiSielecki, 1983). The 11e71 side chain is in a similar conformation to that of Va171, however, theadditional methyl group, 11e71 CD1, is directed into the hydrophobic core. (Figure 5.36). Thexi and X2 torsional angles are 177°and —52°, respectively. These angles are also commonlyobserved for this residue.Unlike the smaller alanine and serine replacements at Pro71, both the Pro71Val and Pro7lIlemutant structures display positional perturbations in the region of the mutation site, in additionto the region about G1y83 (Figure 5.34). As detailed in Table 5.28, the average main chaindeviations of residues 70 and 71 range from 0.34 to 0.57 A to accommodate the bulkier sidechains in these mutants. In particular, these displacements appear to be necessary to allowfor a second 7-carbon atom to fit within the hydrophobic core without the development ofunacceptable steric clashes with residues Tyr67 and Met80 (Figure 5.33). Note that the sidechains of Met80 and Tyr67 are not displaced by any of the mutations studied at position 71Chapter 5. Replacements of an Invariant Proline in Isa-l-Cytochrome c^105(Table 5.28). The displacement of G1y83 is less pronounced than in the Pro7lAla and Pro7lSerproteins, however the phenyl ring of Phe82 is displaced towards the molecular surface by 0.7and 0.3 A in the Pro7lIle and Pro7lVal structures, respectively. The displacement of the Phe82side chain is proportional to the size of the side chain at position 71. Movements occurringin the Pro71Val mutant protein lead to the formation of a hydrogen bond between Va171 Nand G1y83 0 (3.4 A). A similar but longer range interaction (4.0 A) is seen in the Pro7lIlestructure.Inspection of the difference matrices in Figure 5.34 shows that the positional perturbationscentered at G1y83 are accompanied by a change in the main chain thermal factors in thisregion. These values are presented in Table 5.29. The average increase in main chain thermalfactor for G1y83 in all four Pro71 mutant structures is about 8.4 A2 . The largest increase isobserved in the Pro7lSer structure (A = +11 A 2 ) when compared to wild-type protein whichhas an average thermal factor of 19.0 A2 . In contrast, the main chain thermal factors at themutation site, residue 71, do not differ greatly from the wild-type protein values (Table 5.29).The largest deviation is observed in the Pro7lVal structure in which the average increase inmain chain thermal factor is found to be 5.5 A2 . The thermal factors of the Phe82 side chain inthe Pro7lSer structure have decreased in contrast to the other mutants which display a modestincrease in these thermal factors.5.3 DiscussionThe three-dimensional structures of Pro7lAla, Pro7lIle, Pro7lSer. Pro71Val and wild-typeiso-l-cytochromes c provide insight into the role of Pro71 in the folding and function of cyto-chrome c. Mutations at Pro71 are found to lead to perturbations in the atomic positions andthermal factors of residues 82 through 84 (Figure 5.34). If the replacement amino acid containsa larger side chain, the main chain atoms at the mutation site are also displaced. However, thenearby buried side chains of Tyr67 and Met80, which pack against the Pro71 side chain, arenot affected by such substitutions and appear to be more rigidly held in place than residuesChapter 5. Replacements of an Invariant Proline in Iso-l-Cytochrorne c^106Table 5.29: Average thermal factors of groups in the region of residue 71GroupAverage thermal factor (A 2 )Wild-type Pro7lAla Pro7lSer Pro7lIle Pro71ValA. Main chainAsn70 15.1 16.7 17.4 18.6 20.771 12.7 12.8 17.0 16.9 18.2Phe82 15.2 24.2 21.8 21.6 21.5G1y83 19.0 29.3 30.0 24.6 25.7G1y84 17.4 23.2 22.6 19.5 26.7B. Side chainTyr67 13.2 14.7 13.9 11.2 17.471 14.2 8.0 22.9 24.0 18.5Met80 5.2 9.4 7.7 9.6 9.4Phe82 16.9 19.2 14.7 20.3 20.4The thermal factors of the protein atoms of the Pro71 mutants were normalized to those ofiso-1-cytochrome c for these comparisons.82 through 84 which are located at the protein surface. There is a further correlation betweenthe size of the amino acid replacement and the displacement of the invariant Phe82 side chain(Table 5.28). Larger replacements of Pro71 result in an increased displacement of the phenylring of Phe82 towards the protein surface.The four Pro71 mutant structures studied all displayed positional perturbations at residues82 and 83. Phe82 and Gly83 are both located at the protein surface near to the solvent exposedheme edge (Figure 5.33). These residues are part of the proposed interface in modeled electrontransfer complexes between cytochrome c and electron transfer partners such as cytochrome cperoxidase (Poulos & Kraut, 1980; Lum et al., 1987) and cytochrome b5 (Salemme, 1976;Mauk et al., 1986). The invariance and location of Phe82 suggests that this residue may beimportant in the mechanism of electron transfer (Dickerson & Timkovich, 1975). Site-directedmutagenesis has shown that an aromatic group at position 82 is required for efficient electrontransfer (Pielak et al., 1985: Liang et al., 1988). Structural studies of the Phe82Gly mutantChapter 5. Replacements of an Invariant Proline in Iso-l-Cytochrome c^107of yeast iso-l-cytochrome c revealed that the polypeptide chain in this region had refolded(Louie & Brayer, 1989). This refolding, which altered the surface contour of cytochrome c,was suggested to account for the reduced steady state electron transfer kinetics of this mutantwith cytochrome c peroxidase (Louie & Brayer, 1989). Similarly, structural investigation ofPhe82Ser cytochrome c suggested that the phenyl ring of Phe82 must be present for optimalelectron transfer function (Louie et al., 1988b). Mutations at Phe82 have also been shown todecrease the alkaline transition pK, indicating that this residue is important for maintainingthe heme crevice structure (Pearce et al., 1989). Therefore these mutations at position 71that substantially alter the conformational positioning of Phe82 could be expected to affect theoverall structure of cytochrome c as well as potentially disrupting the putative redox partnercomplex interface.The functional properties of Pro7lIle, Pro7lLeu, Pro7lSer, Pro7lThr and Pro71Val cyto-chromes c have been determined in vivo by measuring growth rates of yeast cells on lactatemedia (Ernst et al., 1985). The thermodynamic stability of Pro7lIle, Pro7lThr and Pro7lValcytochromes c have been also measured by guanidine hydrochloride induced equilibrium unfold-ing studies (Ramdas et al., 1986). The functional and stability measurements are summarizedin Table 5.30. The variance in function and stability cannot be readily correlated to one struc-tural parameter. The functional capacity of an amino acid replacement at Pro71 appears todepend on the size, shape and chemical character of the replacement side chain.The integration of residue 71 into the cytochrome c fold is presented in Figure 5.37 forwild-type and four Pro71 replacement structures. The Pro71 side chain in the wild-type iso-1cytochrome c structure is surrounded by other sections of polypeptide chain. Inspection ofthe Pro7lAla mutant reveals the creation of additional space surrounding the A1a71 side chainresulting in a looser packing of the hydrophobic core in this region. The increased thermalmobility and reorganization of the main chain atoms about Gly83 do not entirely fill the spacevacated by the Pro7lAla substitution. This inability to optimize packing of the hydrophobiccore could be expected to reduce the stability of the protein.Chapter 5. Replacements of an Invariant Proline in Iso-l-Cytochrome c^108Table 5.30: Functional and stability studies of Pro71 mutant cytochromes cCytochrome c Function Stability(%) (kcal/mol)Wild-type 100 3.6Pro7lIle 20 1.9Pro7lLeu 0 NDPro7lSer 30 NDPro7lThr 60 1.9Pro71Val 90 2.6Function was estimated by cell growth on lactate at 22 °C (Ernst et al., 1985). Thermodynamicstability of these cytochromes c was determined by denaturation with guanidine hydrochlorideat pH 6.0 and 20 °C (Ramdas et al., 1986). In these latter experiments, Cys102 was modifiedwith methyl methanethiosulfonate.Examination of the Pro7lSer packing interactions reveals that the additional Ser71 OG atomfills more of the vacated space than CB atom of the Pro7lAla mutant protein (Figure 5.37).However, the function of this replacement is 30% of that of wild-type cytochrome c (Table 5.30).The poor functional capacity of this mutant may be a consequence of the increased polarity andthe hydrogen bonding potential of the Ser71 side chain. In the reduced structure, the highlymobile Ser71 OG atom (thermal factor of 27 A 2 ) is hydrogen bonded to Tyr67 0. By rotatingXi, the Ser71 side chain could hydrogen bond to Tyr67 OH which is a component of a hydrogenbonding network implicated in mediating electron transfer (Berghuis & Brayer, 1992). It ispossible that this alternate hydrogen bond involving Ser71 OG could be formed in the moreflexible oxidized state and thereby diminish electron transfer capacity.It is of interest to note that Pro7lThr cytochrome c functional capacity is twice that ofthe Pro7lSer protein (Table 5.30). If the Pro7lThr structure is modelled by comparison to thePro71Val structure, the 0G1 atom is positioned similar to the Ser71 OG atom. However, theadditional methyl group in the Pro7lThr protein relative to the Pro7lSer structure could beexpected to lock the orientation of the hydroxyl group of the Thr71 side chain such that onlythe hydrogen bond to the Tyr67 main chain is formed in both the reduced and oxidized states.Chapter 5. Replacements of an Invariant Proline in Iso-I-Cytochrome c^109Figure 5.37: Space-filled representations of wild-type and Pro71 mutant cytochromes c. Theresidue 71 side chain is shaded black, the heme group is shaded gray and all other proteingroups are drawn in white. Atoms in front of the plane of view were removed to expose residue71. The modelled placement of the Leu71 side chain results in a steric clash with the side chainsof Met80 and Phe82.Chapter 5. Replacements of an Invariant Proline in Iso-l-Cytochrome c^110The stability of the Pro7lThr mutant has also been measured and found to be 2.7 kcal/mol lessthan that of the wild-type protein (Ramdas et al., 1986). In addition, the apparent pl( c, of thealkaline transition of the iso-2 Pro7lThr mutant has also been measured by examination of the695 nm absorption band and found to decrease from 8.45 to 6.63 (White et al., 1987). Theseresults indicate that despite its higher functional capacity the exchange of proline for threonineat position 71 leads to a considerable loss of protein stability.Of the four replacements of known structure, the Pro7lVal mutant protein has the highestfunctional capacity (Table 5.30). In addition, Pro7lVal cytochrome c is the most stable replace-ment of the mutants studied, undoubtedly because the valine side chain is best able to mimicthe size and chemical character of the Pro71 side chain in iso-l-cytochrome c (Figure 5.37).Also, the formation of a hydrogen bond between Va171 N and Gly83 0 may help stabilize thePro7lVal mutant protein.The least functional and stable replacement for which a three-dimensional structure is avail-able is Pro7lIle cytochrome c (Table 5.30). As shown in Figures 5.36 and 5.37, the 11e71 sidechain does not fit into the cytochrome c fold without substantial displacements in adjacentgroups. Furthermore, the 695 nm absorbance band has been found to be significantly reducedin the Pro7lIle mutant protein (Ramdas et al., 1986). This absorbance band is an indicator ofthe presence of the Met80 heme iron ligand (Schechter & Saludjian, 1967) and the integrity ofthe heme pocket. These results suggest that the 11e71 side chain disrupts the heme environmentleading to substantial disruption of the electron transfer mechanism of cytochrome c.The folding and functional requirements of residue 71 of yeast iso-l-cytochrome c have alsobeen characterized by the examination of functional revertants of yeast strains containing aleucine at this position (Ernst et al., 1985). Of the 29 revertants examined, only four differentpartially functional proteins were observed, Pro7lIle, Pro7lSer, Pro7lThr and Pro71Val cyto-chromes c. Six other amino acid replacements, Pro7lGlu, Pro7lGln, Pro7lLys, Pro7lArg,Pro7lTyr and Pro7lPhe, could have been generated by single base pair substitutions but werenot observed and are presumed to be non-functional. All of these residues are larger thanChapter 5. Replacements of an Invariant Proline in Iso-1-Cytochrome c^111the wild-type proline and most contain polar functional groups which would be expected tosignificantly alter the character of the hydrophobic heme pocket.This suggests that most mutations in this region cannot be tolerated. A model constructedby replacing the side chain of Pro71 with that of a leucine is shown in Figure 5.37. The Leu71side chain could not be positioned such that a steric clash with protein groups surroundingresidue 71 does not occur. Although leucine and isoleucine side chains contain the same numberof atoms, the difference in the arrangement of these atoms greatly effects the ability of these sidechains to fit into the hydrophobic core at position 71. As shown in Figure 5.37, a minimalistmodelling of the Leu71 side chain results in steric clashes with both the Met80 and Phe82side chains. The difference in packing of the Leu71 and Ile71 side chains appear to definethe functional limitation of the cytochrome c fold to accommodate replacement side chains atposition 71.A further factor in the instability of Pro71 mutant proteins may be related to amino acidpreferences for specific helix positions. The occurrence of an amino acid residue at a givenposition in an a-helix was examined in 215 a-helices in 45 globular proteins (Richardson &Richardson, 1988). The amino acid preference ratio is derived by normalizing the number ofoccurrences by the expected number based on the overall percentage of each amino acid foundin a sample of 135 proteins of known structure. In particular, there is a strong preference for thefirst and second residues of an a-helix to be asparagine (3.5:1) and proline (2.6:1), respectively.Helix IV of yeast cytochrome c is initiated by the residues Asn70 and Pro71. Of the Pro71replacements discussed, two have preference ratios greater than one: alanine (1.2:1) and valine(1.1:1). Leucine and isoleucine side chains are equally undesirable (0.9:1), whereas a serine isthe least preferred residue (0.7:1) at the second helix position. All of the partially functionalamino acid replacements have near neutral preference for the second helix position as comparedto the strong preference for proline.The structural results obtained for the four Pro71 mutants clearly show the limitations im-posed on substitutions at position 71. The most acceptable by functional and thermal stabilityChapter 5. Replacements of an Invariant Proline in Iso-l-Cytochrome c^112criteria appears to be the Pro71Val substitution (Ernst et al., 1985: Ramdas et al., 1986). Thevaline side chain is best able to mimic the size and chemical character of the Pro71 side chainin iso-l-cytochrome c (Figure 5.36). It seems likely that the structural perturbations observedat residues 82 through 84 as a result of this substitution are responsible for the partial loss offunction observed and may be related to changes in the surface contour of the protein at theredox partner interface. The other mutants investigated are less functional due to additionalstructural perturbations. For example, the serine substitution introduces a polar group adja-cent to the Met80 ligand which was formerly in a highly hydrophobic environment. Smallerreplacements, such as Pro7lAla, result in poor packing in the hydrophobic core which couldbe expected to lead to reduced thermal stability. The larger replacement, Pro7lIle, is onlytolerated by rearrangements of the main chain at the mutation site and a displacement of thePhe82 side chain. The combination of both these displacements seems to lead to a low levelof functional activity. As evident from our studies (Figure 5.37), even larger side chains atthis position such as Leu71 would require large structural reorganizations which would abolishcytochrome c function probably as a result of disrupting the Met80 ligand bond. It is also en-tirely possible that the substitution of even larger side chains would prevent polypeptide chainfolding altogether.Chapter 6SummaryIn this work, a total of eight cytochrome c structures were elucidated to atomic resolution. Oneof these, iso-2-cytochrome c, is a wild-type yeast isozyme differing from iso-l-cytochrome c at17 amino acid positions. The composite mutant protein, B-2036 cytochrome c, possesses 10of these 17 amino acid replacements relative to the iso-1 protein. The a-Loop A replacementprotein, RepA2 cytochrome c, contained five amino acid substitutions, one of which was backsubstituted to that of the wild-type iso-1 protein to produce RepA2(Va120) cytochrome c.Finally, four single site mutants were structurally characterized at position 71. In all, theseproteins have 25 different substitutions at 20 sites (Table 6.31).Table 6.31. Amino acid additions and replacements in the eight cytochrome c structures studiedin this workCytochrome c No. Residue additions and replacementsIso-2B-2036RepA2RepA2(Va120)Pro71AlaPro7lIlePro71SerPro71Val21^Ala(-9), Lys(-8), Glu(-7), Ser(-6), Glu(-3)Gly, Ala(-1)Pro,Leul5Gln, Va120Ile, Lys22Glu, His26Asn, Ala43Val, Glu44Lys,Lys54Asn, Leu58Lys, Asn62Asp, Asn63Ser, Gly83Ala, Leu98Met,Lys99Thr, Cys102Ala, Glu103Lys10 Leul5G1n, Va120Ile, Lys22Glu, His26Asn, Ala43Val, Glu44Lys,Lys54Asn, Leu58Lys, Asn62Asp, Asn63Ser5^Va120Phe, Glu2lAsp, Lys22G1n, Pro25Ala, His26Asn4^Glu21Asp, Lys22Gln, Pro25Ala, His26Asn1^Pro71Ala1^Pro7lIle1^Pro71Ser1^Pro7lVal113Chapter 6. Summary^ 114An overview of the positional deviations in the various cytochromes c examined in this workfrom wild-type iso-l-cytochrome c is presented in Figure 6.38. The overall average main chaindeviation of the eight structures is 0.25 A. Despite the large number of amino acid substitutionspresent in these structures, the main chain fold is highly conserved. The largest deviation occursat G1y37 and is due to an alternative conformation in main chain atom positions in the iso-2 andB-2036 structures (Figure 3.14). The overall average positional deviations of side chain atoms(0.45 A) is almost twice that observed for main chain atoms. The majority of large deviationsinvolve residues with disordered solvent accessible side chains (Figure 6.38). Additionally, thetwo buried side chains with the largest deviations, Leu9 and G1u61, are disordered in the iso-1-cytochrome c structure (Louie & Brayer, 1990).The iso-2, B-2036 and RepA2 structures share some common features with respect to residue20 and the subsequent alterations in hydrophobic core packing when compared to iso-l-cyto-chrome c. In iso-2-cytochrome c, the replacement of isoleucine for valine at position 20 iscompensated for by the substitution Cys102Ala. In the B-2036 protein where the Cys102 isretained, disruptions of the hydrophobic core (residues Ile20, 11e35, Leu98 and Cys102) occurto accommodate the addition of the buried Ile20 CD1 group (Figures 3.17 and 6.38). Thesealterations in hydrophobic core packing also lead to an increase in the volume of a buriedcavity located adjacent to the heme group (Figure 3.18). In RepA2 cytochrome c, there is aphenylalanine at position 20. The larger side chain of this amino acid is redirected towardsthe surface of the protein (Figure 4.26). This Va120Phe substitution creates a void in thehydrophobic core and Leu98 moves to fill the vacated space. In addition, Tyr97 and Ala101are displaced to accommodate the phenylalanine side chain at the protein surface. Theseresults indicate that there is a limit as to the size of the amino acid replacement that may beaccommodated into the hydrophobic core by adjustment of neighboring side chain positions. Atposition 20 in yeast iso-l-cytochrome c, whereas an extra methyl group was accommodated, aphenyl ring cannot be accommodated and is redirected away from the hydrophobic core. Evenwith the disruptions observed to the hydrophobic core in the case of Va120Phe it is remarkable45^55^65^75 85 11 1951^1^11^111i1^1^1 1^1^11^1^1^1^1^I^1^1^1^1^I^1^1^1^1^1^1^1^1^1 ^11111111Chapter 6. Summary^ 11500a)Da)0)0La)2.52.0L'1.5 -1.0 -0.5-0.0 0.00.5 7 -1.5 -2.0-2.53.0 -3.54.0II^1^1 ^ 11111I^II^11^1111^11111^1^1^1 1^1^1 1^1 1^1^i^1111^1^1^1I^1^1^1^1^1^1 11Residue NumberFigure 6.38: A plot of the overall average deviation of main (above) and side (below) chainatoms of the structures examined in this work (Table 6.31). The vertical bars in the upper graphrepresent the range of individual pairwise average deviations. The dashed and solid verticalbars in the lower graph represent solvent accessible and inaccessible side chains, respectively.A side chain was considered solvent inaccessible if less then 20% of its surface area was solventexposed in any one of the structures examined. The overall average deviations for the mainand side chain atoms are 0.25 and 0.45 A, respectively.Chapter 6. Summary^ 116how little the overall conformation of the main chain was perturbed.The reduction potential of the RepA2 protein is observed to be 19 mV less than that ofeither RepA2(Va120) or wild-type iso-l-cytochrome c (Table 4.23). The largely solvent exposedPhe20 side chain in the RepA2 structure may modify the reduction potential by altering solventordering upon reduction. The presence of a phenylalanine and the associated structural pertur-bations also diminish the in vivo function of cytochrome c. This loss in function may be due tothe replacement of amino acids at the protein surface combined with the increase in mobilityof Q-Loop A in the RepA2 structure thereby interfering with the binding of cytochrome c withphysiologically important electron transfer partners.All of iso-2, B-2036, RepA2 and RepA2(Va120) cytochromes c have an asparagine at position26 as compared to a histidine in the iso-1 protein. His26 forms two hydrogen bonds, one to eachof the main chain atoms of residues 31 and 44 (Table 3.14). This forms a crosslink between themain chain backbones of the polypeptide segments involved. The substituted asparagine sidechain of the iso-2 and mutant cytochromes c possesses equivalent functional groups to formthese hydrogen bonds, but because of its smaller size the hydrogen bonded groups cannot beplaced in the same position as observed in wild-type iso-l-cytochrome c. As a consequence, themain chain about Glu44 shifts (Figure 6.38) in the a-Loop A mutants to allow the formationof a hydrogen bond between Asn26 ND2 and G1u44 0 (Figure 4.25). In the iso-2 and B-2036proteins this hydrogen bond is not formed. This appears to result from a second substitution,Ala43Val, which fills the space resulting from the His26Asn substitution (Figure 3.15). In theend this double substitution provides stronger hydrophobic interactions in this region. Thepositions of the main chain atoms surrounding G1u44 are adjusted in the iso-2 and B-2036proteins to optimize the fit of Va143 (Figures 3.13 and 6.38).The invariant residue Pro71 is located near the Met80 ligand (Figure 5.33). It is a com-pletely solvent inaccessible residue that bridges Helices III and IV (Table 1.4). The uniqueproperties of proline residues are thought to be important for the stabilization of this struc-turally conserved region of cytochrome c. The structures of four substitutions at this positionChapter 6. Summary^ 117revealed that changes in side chain size, shape and chemical character resulted in displacementsin the positions of residues 70, 71 and 82 through 84 (Figure 5.34). Replacement of Pro7lwith the smaller side chains, Pro7lAla and Pro7lSer, required only adjustment of residues 82through 84 (Figure 5.35), however, larger side chain replacements, for example Pro7lIle andPro71Val (Figure 5.36), were accommodated only at the expense of altering the structure of themain chain at the mutation site. It appears that unlike residues 20 and 26 as discussed above,the internal positioning and rigid polypeptide chain structure at the mutation site prevents theprotein from adapting to the presence of a wider range of residues at position 71. The structuralperturbations observed at residues 82 through 84 as a result of substitutions at position 71 aresuggested to be responsible for the observed partial loss of function in these mutant proteins.Of the 25 amino acid substitutions listed in Table 6.31, only a fraction give rise to the sub-stantial structural perturbations. The majority do not appear to result in observable structuraldifferences. Those amino acid substitutions that cannot be incorporated into the cytochrome cfold without structural perturbations occur in conformationally restricted regions. In addition,the displacement of groups by these replacements indicates the limits present in specific regionsin terms of conformational flexibility. For example, the substitution Va120Ile results in smallalterations in the packing of the hydrophobic core, however, the Va120Phe substitution resultsin the redirection of the side chain of this residue apparently because the hydrophobic coreis incapable of adjusting to accommodate the larger phenyl ring. The interaction of the twosegments of polypeptide chain bridged by hydrogen bonds to His26 represents a conformationalrestriction in the polypeptide chain at this point. The observed shifts of residues 42 to 44 inthe iso-2, B-2036. RepA2 and RepA2(Va120) mutant proteins as a result of the His26Asn re-placement suggests that this segment of polypeptide chain is more flexible than the polypeptidechain segment that contains residue 26. The small range of acceptable substitutions at position71 reflects the low tolerance for side chain replacements in this conformationally restricted andrigid region of the protein.BibliographyArmstrong, F.A., Hill, A.O. & Walton, N.J. (1988). Direct electrochemistry of redox proteins.Acc. Chem. Res., 21, 407-413.Arndt, U.W. & Wonacott, A.J., eds. (1977). The Rotation Method in Crystallography. North-Holland Publishing Company, Amsterdam.Barker, P.D., Hill, A.O. & Walton, N.J. (1989). Fast second order electron transfer reactionscoupled to redox protein electrochemistry. J. Electroanal. Chem., 260, 303-326.Berghuis, A.M. (1993). The Nature and Role of Oxidation State Dependent ConformationalDifferences in Cytochrome c. PhD thesis, University of British Columbia In preparation.Berghuis, A.M. & Brayer, G.D. (1992). Oxidation state-dependent conformational changes incytochrome c. J. Mol. Biol., 223, 959-976.Bernstein, F.C., Koetzle, T.F., Williams, G.J.B., Meyer, E.F. Jr., Brice, M.D., Rodgers, J.R.,Kennard, 0., Shimanouchi, T. & Tasmui, M. (1977). The Protein Data Bank: a computer-based archival file for macromolecular structures. J. Mol. Biol., 112, 535-542.Bhat, T.N., Sasisekharan, V. & Vijayan, M. (1978). An analysis of side chain conformations inproteins. Int. J. Pept. Protein Res., 13, 170-184.Brautigan, D.L., Feinburg, B.A., Hoffman, B.M., Margoliash, E., Peisach, J. & Blumberg, W.B.(1977). Multiple low spin forms of cytochrome c ferrihemochrome. J. Biol. Chem., 252,574-582.Brayer, G.D. Si Murphy, M.E.P. (1993). Structural studies of eukaryotic cytochrome c. InThe Cytochrome c Handbook (Mauk, A.G. & Scott, R.A., eds.). Plenum Press, New York.Submitted.Buerger, M.J. (1964). The Precession Method. John Wiley, New York.Bushnell, G.W., Louie, G.V. & Brayer, G.D. (1990). High-resolution three-dimensional struc-ture of horse heart cytochrome c. J. Mol. Biol., 214, 585-595.Chambers, J.L. Si Stroud, R.M. (1979). The accuracy of refined protein structures: comparisonof two independently refined models of bovine trypsin. Acta Crystallogr. sect. B, 35, 1861-1874.Chothia, C. & Lesk, A.M. (1986). The relation between the divergence of sequence and structurein proteins. EMBO J., 5, 823-826.118BIBLIOGRAPHY^ 119Churg, A.K. Weiss, R.M. & Warshel, A. (1983). On the action of cytochrome c: correlatinggeometry changes upon oxidation with activation energies of electron transfer. J. Phys.Chem., 87, 1683-1694.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. Amer. Chem. Soc., 107, 1118-1124.Corradin, G. & Harbury, H.A. (1974). Reconstitution of horse heart cytochrome c. Reformationof the peptide bond linking residues 65 & 68. Biochem. Biophys. Res. Comm., 61, 1400-1406.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: correc-tion. Acta Crystallogr., 7, 519.Davis, L.A., Schejter, A. & Hess, G.P. (1974). Alkaline isomerization of oxidized cytochrome c.J. Biol. Chem., 249, 2624-2632.Dayhoff, M.O. & Barker, W.C. (1976). Cytochrome c. In Atlas of Protein Sequence and Struc-ture (Dayhoff, M.O., ed.), vol. 5, pp. 25-49. National Biomedical Research Foundation,Silver Spring, Md.Denning, P.J., Dennis, J.B. & Quality, J.E. (1978). Machines, Languages, and Computation.Prentice-Hall, Englewood Cliffs, New Jersey 07632.Dethmers, J.K., Ferguson-Miller, S. & Margoliash, E. (1979). Comparison of yeast and beefcytochrome c oxidases. J. Biol. Chem., 254, 11973-11981.Dickerson, R.E., Takano, T., Eisenberg, D., Kallai, 0.B., Samson, L., Cooper, A. & Margoliash,E. (1971). Ferricytochrome c, general features of the horse and bonito proteins at 2.8 Aresolution. J. Biol. Chem., 246, 1511-1532.Dickerson, R.E. & Timkovich, R. (1975). Cytochromes c. In The enzymes (Boyer, P.D.,vol. 11, part A, pp. 397-547. Academic Press, New York, 3rd edition.Dumont, M.E., Ernst, J.F. & Sherman, F. (1987). Identification and sequence of the geneencoding cytochrome c heme lyase in the yeast Saccharomyces cerevisiae. EMBO J., 6,235-241.Dumont, M.E., Ernst, J.F. & Sherman, F. (1988). Coupling of heme attachment to import ofcytochrome c into yeast mitochondria: Studies with heme lyase deficient mitochondria andaltered apocytochromes c. J. Biol. Chem., 263, 15928-15937.BIBLIOGRAPHY^ 120Dumont, M.E., Mathews, A.J., Nall, B.T., Bann, S.B., Eustice, D.C. & Sherman, F. (1990).Differential stability of two apo-isocytochromes c in the yeast Saccharomyces cerevisiae. J.Biol. Chem., 265, 2733-2739.Dus, K., Sletten, K. & Kamen, M.D. (1968). Cytochrome c2 of Rhodospirillum rubrum, completeamino acid sequence and phylogenetic relationships. J. Biol. Chem., 243, 5507-5518.Eden, D., Matthew, J.B., Rosa, J.J. & Richards, F.M. (1982). Increase in apparent compress-ibility of cytochrome c upon oxidation. Proc. Nat. Acad. Sci., U.S.A., 79, 815-819.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.Ernst, J.F., Hampsey, D.M., Stewart, J.W. Rackovsky, S., Goldstein, D. & Sherman, F. (1985).Substitutions of proline 76 in yeast iso-l-cytochrome c, analysis of residues compatible andincompatible with folding requirements. J. Biol. Chem., 260, 13225-13236.Ernst, J.F., Stewart, J.W. & Sherman, F. (1981). The cyc1-11 mutation in yeast reverts byrecombination with a nonallelic gene: composite genes determining the iso-cytochromes c.Proc. Nat. Acad. Sci., U.S.A., 78, 6334-6338.Ernst, J.F., Stewart, J.W. & Sherman, F. (1982). Formation of composite iso-cytochromes cby recombination between non-allelic genes of yeast. J. Mol. Biol., 161, 373-394.Ferguson-Miller, S., Brautigan, D.L. & Margoliash, E. (1979). The electron transfer function ofcytochrome c. In The Porphyrins (Dolphin, D., ed.), vol. 7, pp. 149-240. Academic Press,New York.Fetrow, J.S., Cardillo, T.S. & Sherman, F. (1989). Deletions and replacements of omega loopsin yeast iso-l-cytochrome c. Proteins: Struct. Plate . Genet., 6, 372-381.Greef, R., Peat, R., Peter, L.M., Pletcher, D. & Robinson, J. (1985). Instrumental Methods inElectrochemistry / Southampton Electrochemistry Group (Kemp, T.J., ed.). Ellis Horwood,Chichester, England.Hampsey, D.M., Das, G. & Sherman, F. (1986). Amino acid replacements in yeast iso-l-cytochrome c, comparison with the phylogenetic series and the tertiary structures of relatedcytochromes c. J. Biol. Chem., 261, 3259-3271.Hampsey, D.M., Das, G. & Sherman, F. (1988). Yeast iso-l-cytochrome c: genetic analysis ofstructural requirements. FEBS Letters, 231, 275-283.Hendrickson, W.A. & Konnert, J. (1981). Stereochemically restrained crystallographic least-squares refinement of macromolecular structures. In Biomolecular Structure, Function,Conformation and Evolution (Sunivasan, R., ed.). vol. 1, pp. 43-57. Pergamon Press,Oxford.BIBLIOGRAPHY^ 121Hickey, 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-cytochrome c with amino acid replacements at position 52 and 102. J. Biol.Chem., 266, 11686-11694.Hickey, D.R., McLendon, G. & Sherman, F. (1988). Thermodynamic stabilities of yeast iso-1-cytochromes c having amino acid substitutions for lysine 32. J. Biol. Chem., 263, 18298-18305.Hong, X. & Dixon, D.W. (1989). NMR study of the alkaline isomerization of ferricytochromec. FEBS Letters, 246, 105-108.Ikai, A., Fish, W.W. & Tanford, C. (1973). Kinetics of unfolding and refolding of proteins: II.Results for cytochrome c. J. Mol. Biol., 73, 165-184.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.James, M.N.G. & Sielecki, A.R. (1983). Structure and refinement of penicillopepsin at 1.8 Aresolution. J. Mol. Biol., 163, 299-361.Janin, J., Wodak, S., Levitt, M. & Maigret, B. (1978). Conformation of amino-acid side chainsin proteins. J. Mol. Biol., 125, 357-386.Jones, T.A. (1978). A graphics model building and refinement system for macromolecules. J.Appl. Cryst., 11, 268-272.Kabsch, W. & Sander, C. (1983). Dictionary of protein secondary structure: pattern recognitionof hydrogen bonded and geometrical features. Biopolymers, 22, 2577-2637.Karpusas, M.K., Baase, W.A., Matsumura, M. & Matthews, B.W. (1989). Hydrophobic packingin T4 lysozyme probed by cavity-filling mutants. Proc. Nat. Acad. Sci., U.S.A., 86, 8237-8241.Kassner, R.J. (1972). Effects of nonpolar environments on the redox potentials of heme com-plexes. Proc. Nat. Acad. Sci., U.S.A., 69, 2263-2267.Kassner, R.J. (1973). A theoretical model for the effects of local nonpolar heme environmentson the redox potentials of cytochromes. J. Amer. Chem. Soc., 95, 2674-2677.Knapp, J.A. & Pace, N.P. (1974). Guanidine hydrochloride and acid denaturation of horse,cow and Candida krusei cytochromes c. Biochemistry, 13, 1289-1294.Kreil, G. (1965). Die c-terminale aminosauresequenz des tunfischcytochroms c. Z. Physiol.Che•ie., 340, 86-87.BIBLIOGRAPHY^ 122Kreishman, G.P., Anderson. C.W., Su, C.-H., Halsall, H.B. & Heineman, W.R. (1978). Theeffects of bulk solvent structure on the temperature dependence of the reduction potentialof cytochrome c. Bioelect. and Bioenerg., 5, 196-203.Kunkel, T.A., Roberts, J.D. & Zakour, R.A. (1987). Rapid and efficient site-specific mutagenesiswithout phenotypic selection. Methods of Enzymol., 154, 367-382.Langen, R., Brayer, G.D., Berghuis, A.M., McLendon, G., Sherman, F. & Warshel, A. (1992).Effect of the Asn52—*Ile mutation on the redox potential of yeast cytochrome c; theoryand experiment. J. Mol. Biol., 224, 589-600.Latimer, W.M. (1952). Oxidation Potentials. Prentice Hall, New York, 2nd edition.Latimer, W.M., Ptizer, K.S. & Smith, W.V. (1938). The entropies of aqueous ions. J. Amer.Chem. Soc., 60, 1829-1831.Laz, T.M., Pietras, D.F. & Sherman, F. (1984). Differential regulation of the duplicated isocy-tochrome c genes in yeast. Proc. Nat. Acad. Sci., U.S.A., 81, 4475-4479.Lee, B. & Richards, F.M. (1971). Estimation of protein structures: static accessibility. J. Mol.Biol., 55, 379-400.Leszczynski, J.F. & Rose, G.D. (1986). Loops in globular proteins: a novel category of secondarystructure. Science, 234, 849-855.Leung, C.J., Nall, B.T. & Brayer, G.D. (1989). Crystallization of yeast iso-2-cytochrome cusing a novel hair seeding technique. J. Mol. Biol., 206, 783-785.Liang, N., Mauk, A.G., Pielak, G.J., Johnson, J.A., Smith, M. & Hoffman, B. (1988). Regu-lation of interprotein electron transfer by residue 82 in yeast cytochrome c. Science, 240,311-313.Lo, T.P., Murphy, M.E.P., Guillemette, J.G., Smith, M. & Brayer, G.D. (1993). Structuralconsequences of mutating leucine residues in the hydrophobic core of yeast iso-l-cytochromec. In preparation.Louie, G.V. (1990). Structural Studies of Wild-type and Variant Yeast Iso-1-Cytochromes c.PhD thesis, University of British Columbia.Louie, G.V. & Brayer, G.D. (1989). A polypeptide chain-refolding event occurs in the G1y82variant of yeast iso-l-cytochrome c. J. Mol. Biol., 210, 313-322.Louie, G.V. & Brayer, G.D. (1990). High-resolution refinement of yeast iso-l-cytochrome c andcomparisons 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.BIBLIOGRAPHY^ 123Louie, G.V., Pielak, G.J., Smith, M. & Brayer, G.D. (1988b). The role of Phe-82 in yeastiso-l-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 Struct.,Folding Design, 2, 143-150.Lumry, R. & Rajender, S. (1970). Enthalpy-entropy compensation phenomena in water solu-tions of proteins and small molecules: a ubiquitous property of water. Biopolymers, 9,1125-1227.Luzzati, V. (1952). Traitement statistique des erreurs dans la determination des structurescristallines Acta Crystallogr., 5, 802-810.Margalit, R. & Schejter, A. (1973a). Cytochrome c: a thermodynamic study of the relationshipsamong oxidation state, ion binding and structural parameters. 1. The effects of tempera-ture, pH and electrostatic media on the standard redox potential of cytochrome c. Eur. J.Biochem, 32, 492-499.Margalit, R. & Schejter, A. (1973b). Cytochrome c: a thermodynamic study of the relationshipsamong oxidation state, ion binding and structural parameters. 2. Ion-binding linked tooxidation state. Eur. J. Biochem, 32, 492-499.Margoliash, E., Smith, E.L., Kreil, G. & Tuppy, H. (1961). Amino acid sequence of horse heartcytochrome c. Nature (London), 192, 1121-1127.Mathews, F.S. (1985). The structure, function and evolution of cytochromes. Progr. Biophys.Mol. Biol., 33, 1-56.Matsuura, Y., Hata, Y., Yamaguchi, T., Tanaka, N. & Kakudo, M. (1979). Structure of bonitoheart ferricytochrome c and remarks on molecular interaction in its crystalline state. J.Biochem. (Tokyo), 85, 729-737.Matthews, B.W. (1968). Solvent content of protein crystals. J. Mol. Biol., 33, 491-497.Mattoon, J.R. & Sherman, F. (1966). Reconstitution of phosphorylating electron transport inmitochondria from a cytochrome c-deficient yeast mutant. J. Biol. Chem., 241, 4330-4338.Mauk, M.R., Mauk, A.G., Weber, P.C. & Matthew, J.B. (1986). Electrostatic analysis of theinteraction of cytochrome c with native and dimethyl ester heme substituted cytochromeb5. Biochemistry, 25, 7085-7091.McLendon, G. & Smith, M. (1978). Equilibrium and kinetic studies of unfolding of homologouscytochromes c. J. Biol. Chem., 253, 4004-4008.McPherson, A. (1982). Preparation and Analysis of Protein Crystals. John Wiley & Sons, NewYork.BIBLIOGRAPHY^ 124Messerschmidt, A. & Pflugrath, J.W. (1987). Crystal orientation and X-ray pattern predictionroutines for area detector diffractometer systems in macromolecular crystallography. J.Appl. Cryst., 20, 306-315.Montgomery, D.L., Hall, B.D., Gillam, S. & Smith, M. (1978). Identification and isolation ofthe yeast cytochrome c gene. Cell, 14, 673-680.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. Nat. Acad.Sci., U.S.A., 77, 541-545.Moore, G.R. (1983). Control of redox properties of cytochrome c by special electrostatic inter-actions. FEBS Letters, 161, 171-175.Moore, G.R. & Pettigrew, G.W. (1990). Cytochrome c, Evolution, Structural and Physiochem-ical Aspects. Springer-Verlag, Berlin.Mori, E. & Morita, Y. (1980). Amino acid sequence of cytochrome c from rice. Tokyo J.Biochem., 87, 249-266.Moult, J. & James, M.N.G. (1986). An algorithm for determining the conformation of polypep-tide segments in proteins by systematic search. Proteins: Struct. Func. Genet., 1, 146-163.Nall, B.T. (1983). Structural intermediates in the folding of yeast iso-2 cytochrome c. Bio-chemistry, 22, 1423-1429.Nall, B.T. (1986). Native or nativelike species are transient intermediates in folding of alkalineiso-2 cytochrome c. Biochemistry, 25, 2974-2978.Nall, B.T. & Landers, T.A. (1981). Guanidine hydrochloride induced unfolding of yeast iso-2cytochrome c. Biochemistry, 20, 5403-5411.Nall, B.T., Zuniga, E.H., White, T.B., Wood, L.C. & Ramdas, L. (1989). Replacement of a con-served proline and the alkaline conformational change in iso-2 cytochrome c. Biochemistry,28, 9834-9839.North, A.C.T., Phillips, D.C. & Mathews, F.S. (1968). A semi-empirical method of absorptioncorrection. Acta Crystallogr., A24, 351-359.Ochi, H., Hata, Y., Tanaka, N., Kakudo, M., Sakuri, T., Aihara, S. & Morita, Y. (1983).Structure of rice ferricytochrome c at 2.0 A resolution. J. Mol. Biol., 166, 407-418.Offord, R.E. (1987). Protein engineering by chemical means? Protein Eng., 1, 151-157.Osterhout, J.J., Muthukrishnan, K. & Nall, B.T. (1985). pH-induced conformational changesand equilibrium unfolding in yeast iso-2 cytochrome c. Biochemistry, 24, 6680-6684.Osterhout, J.J. Si Nall, B.T. (1985). Slow refolding kinetics in yeast iso-2 cytochrome c. Bio-chemistry, 24, 7999-8005.BIBLIOGRAPHY^ 125Pearce, L.L., Gãrtner, A.L., Smith, M. & Mauk, A.G. (1989). Mutation—induced perturbationof the cytochrome c alkaline transition. Biochemistry, 28, 3152-3156.Pettigrew, G.W. & Moore, G.R. (1987). Cytochrome c, Biological Aspects. Springer-Verlag,Berlin.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, 152-153.Ponder, J.W. & Richards, F.M. (1987). Tertiary templates for proteins. Use of packing criteriain the enumeration of allowed sequences for different structural classes. J. Mol. Biol., 193,775-791.Poulos, T.L.^Finzel, B.C. (1984). Heme enzyme structure and function. Pept. Protein Rev.,4, 115-171.Poulos, T.L. & Kraut, J. (1980). A hypothetical model of the cytochrome c peroxidase-cytochrome c electron transfer complex. J. Biol. Chem., 255, 10322-10330.Rafferty, S.P. (1992). Functional Properties of Active Site Variants of Yeast Cytochrome c. PhDthesis, University of British Columbia.Rafferty, S.P., Pearce, L.L., Barker, P.D., Guillemette, J.G., Kay, C.M., Smith, M. & Mauk,A.G. (1990). Electrochemical, kinetic, and circular dichroic consequences of mutations atposition 82 of yeast iso-l-cytochrome c. Biochemistry, 29, 9365-9369.Ramakrishnan, C. & Ramachandran, G.N. (1965). Stereochemical criteria for polypeptide andprotein chain conformations. Biophys. J., 5, 909-933.Ramdas, L., Sherman, F. & Nall, B.T. (1986). Guanidine hydrochloride induced equilibriumunfolding of mutant forms of iso-l-cytochrome c with replacements at proline-71. Biochem-istry, 25, 6952-6958.Richardson, J.S. (1981). The anatomy and taxonomy of protein structure. Ann. Rev. of Prot.Chem., 34, 167-339.Richardson, J.S. & Richardson, D.C. (1988). Amino acid preferences for specific locations atthe ends of a helices. Science, 240, 1648-1652.Roder. H., ElEwe, G.A. & Englander, W. (1988). Structural characterization of folding interme-diates in cytochrome c by H-exchange labelling and proton NMR. Nature (London), 335,700-704.Rossmann, M.G., Leslie, A.G.W., Abdel-Meguid, S.S. & Tsukihara, T. (1979). Processing andpost-refinement of oscillation camera data. J. Appl. Cryst., 12, 570-581.Saigo, S. (1981). A transient spin-state change during alkaline isomerization of ferricytochromec. J. Biochem, 89, 1977-1980.BIBLIOGRAPHY^ 126Salemme, 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). A hypothetical structure for an intermolecular electron transfer complexof cytochrome c and b5. J. Mol. Biol., 102, 563-568.Salemme, F.R. (1977). Structure and function of cytochromes c. Ann. Rev. Biochem., 46,299-329.Salemme, F.R., Freer, S.T., Xuong, N.H., Alden, R.A. & Kraut, J. (1973). The structure ofoxidized cytochrome C2 of Rhodospirillum rubrum. J. Biol. Chem., 248, 3910-3921.Schechter, E. & Saludjian, P. (1967). Conformation of ferricytochrome c IV: relationship be-tween optical absorbance and protein conformation. Biopolymers, 5, 788-790.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, Amsterdam.Schellman, J.A. (1978). Solvent denaturation. Biopolymers, 17, 1305-1322.Schweingruber, M.E., Sherman, F. & Stewart, J.W. (1977). Altered absorption spectra of iso-1cytochromes c from mutants of yeast. J. Biol. Chem., 252, 6577-6580.Schweingruber, M.E., Stewart, J.W. & Sherman, F. (1979). Primary site and second site rever-tants of missense mutants of the evolutionary invariant tryptophan 64 in iso-l-cytochromec from yeast. J. Biol. Chem., 254, 4132-4143.Sels, A.A., Fukuhara, G., Pere, G. & Slonimski, P.P. (1965). Cintique de la biosynthese induitede l'iso-l-cytochrome c et de l'iso-2-cytochrome c au course de l'adaptation a l'oxygene.Biochim. Biophys. Acta, 95, 486-502.Sherman, F. (1990). Studies of yeast cytochrome c: how and why they started and why theycontinued. Genetics, 125, 9-12.Sherman, F., Fink, G.R. & Hicks, J. (1983). Yeast Genetics. Cold Spring Harbor Laboratory,Cold Spring Harbor, New York.Sherman, F., Jackson, M., Liebman, S.W., Schweingruber, A.M. & Stewart, J.W. (1975).A deletion map of cycl mutants and its correspondence to mutationally altered iso-l-cytochromes c of yeast. Genetics, 81, 51-73.Sherman, F. & Slonimski, P.P. (1964). Respiration-deficient mutants of yeast II. Biochim.Biophys. Acta, 90, 1-15.Sherman, F., Stewart, J.W., Jackson, M., Gilmore, R.A. & Parker, J.H. (1974). Mutantsdefective in iso-l-cytochrome c. Genetics, 77, 255-284.BIBLIOGRAPHY^ 127Sherman, F., Stewart, J.W., Parker, J.H., Inhaber, E., Shipman, N.A., Putterman, G.J.,Gardisky, R.L. & Margoliash, E. (1968). The mutational alteration of the primary structureof yeast iso-l-cytochrome c. J. Biol. Chem., 243, 5446-5456.Sherman, F., Taber, H. & Campbell, W. (1965). Genetic determination of iso-l-cytochrome cin yeast. J. Mol. Biol., 13, 21-39.Sherwood, C. & Brayer, G.D. (1985). Crystallization and preliminary diffraction data for iso-1-cytochrome c from yeast. J. Mol. Biol., 185, 209-210.Smith, M., Leung, D.W., Gillam, S., Astell, C.R., Montgomery, D.L. & Hall, B.D. (1979).Sequence of the gene for iso-l-cytochrome c in Saccharomyces cerevisiae. Cell, 16, 753-761.Sokolik, C.W. & Cohen, R.E. (1991). Structure of ubiquitin conjugates of yeast iso-2-cytochromec. J. Biol. Chem., 266, 9100-9107.Sokolik, C.W. & Cohen, R.E. (1992). Ubiquitin conjugation to cytochrome c; structure ofthe yeast iso-1 conjugate and possible recognition determinants. J. Biol. Chem., 267,1067-1071.Stellwagen, E. (1978). Haem exposure as the determinate of oxidation-reduction potential ofhaem proteins. Nature (London), 330, 86-88.Swanson, R., Trus, B.L., Mandel, N., Mandel, G., Kallai, 0.B., & Dickerson, R.E. (1977). Tunacytochrome c at 2.01 resolution. I: Ferricytochrome structure analysis. J. Biol. Chem.,252, 759-775.Takano, T. & Dickerson, R.E. (1981a). Conformation change of cytochrome c: ferrocytochromec structure refined at 1.5 A resolution. J. Mol. Biol., 153, 79-94.Takano, T. & Dickerson, R.E. (1981b). Conformational change in cytochrome c: ferricytochromec refinement at 1.8 A and comparison with the ferrocytochrome structure. J. Mol. Biol.,153, 95-115.Tanaka, N., Yamane, T., Tsukihara, T., Asida, T. & Kakudo, M. (1975). The crystal structureof bonito (kasuo) ferrocytochrome c at 2.3 A II: structure and function. J. Biochem (Tokyo),77, 147-162.Taniguchi, V.T., Sailasuta-Scott, N.. Anson, F.C. & Gray, H.B. (1980). Thermodynamics ofmetalloprotein electron transfer reactions. Pure & Applied Chem., 52, 2275-2281.Theorell, H. & Akesson, A. (1941). Studies on cytochrome c II. The optical properties of purecytochrome c and some of its derivatives. J. Amer. Chem. Soc., 63, 1812-1818.Timkovich, R. (1979). Cytochrome c: the architecture of a protein-porphyrin complex. In ThePorphyrins (Dolphin, D., ed.), vol. 7, pp. 241-289. Academic Press, New York.BIBLIOGRAPHY^ 128Tong, H., Guillemette, J.G., Mauk, A.G., Smith, M., Sherman, F. Si Brayer, G.D. (1993). Thefunctional roles of amino acids around heme propionate A in cytochrome c: a structuraland functional study of four mutants. In preparation.Trewhella, J., Carlson, V.A.P., Curtis, E.H.^Heidorn, D.B. (1988). Differences in the solutionstructures of oxidized and reduced cytochrome c measured by small-angle X-ray scattering.Biochemistry, 27, 1121-1125.Tsong, T.Y. (1976). Ferricytochrome c chain folding by energy transfer of tryptophan 59 to theheme group. Biochemistry, 15, 5467-5473.Wallace, C.J.A. (1984). Modulation of the alkaline transition in cytochrome c and cytochromec—T by full or specific partial acetimidylation. Biochem. J., 217, 601-604.Wallace, C.J.A.^Clark-Lewis, I. (1992). Functional role of heme ligation in cytochrome c:effects of replacement of methionine 80 with natural and non-natural residues by semisyn-thesis. J. Biol. Chem., 267, 3852-3861.Wallace, C.J.A., Guillemette, J.G., Hibaya, Y. Si Smith, M. (1991). Enhancing protein engi-neering capabilities by combining inutagenesis with semisynthesis. J. Biol. Chem., 266,21355-21357.White, T.B., Berget, P.B. Si Nall, B.T. (1987). Changes in conformation and slow refoldingkinetics in mutant iso-2-cytochrome c with replacement of a conserved proline residue.Biochemistry, 26, 4358-4366.Wilson, A.J.C. (1942). Determination of absolute from relative X-ray intensity data. Nature(London), 150, 151-152.Wood, L.C., Muthukrishnan, K., White, T.B., Ramdas, L. & Nall, B.T. (1988). Replace-ment of a conserved proline eliminates the absorbance-detected slow folding phase of iso-2-cytochrome c. Biochemistry, 27, 8562-8567.Wu, T.T. Kabat, E.A. (1970). An analysis of the sequences of the variable regions of Bence-Jones proteins and myeloma light chains and their implications for antibody complemen-tarity. J. Exp. Med., 132, 211-250.Wyckoff, H.W. (1985). Diffractometry. Methods in Enzymology, 114, 330-386.Zoller, M.J. Si Smith, M. (1983). Oligonucleotide-directed mutagenesis of DNA FragmentsCloned into M13 Vectors. Methods of Enzymology, 100, 468-500.Zuniga, E.H.^Nall. B.T. (1983). Folding of yeast iso-l-AM cytochrome c. Biochemistry, 22,1423-1429.


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



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"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items