A STRUCTURE/ACTIVITY STUDY OF CATION AFFINITYAND SELECTIVITY USING A SYNTHETIC PEPTIDE MODEL OF THEHELIX-LOOP-HELIX CALCIUM-BINDING MOTIFbyRIC MICHAEL PROCYSHYNB.Sc. (Pharm.), The University of Manitoba, 1983M.Sc., The University of Manitoba, 1989A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFDOCTOR OF PHILOSOPHYinTHE FACULTY OF GRADUATE STUDIESDivision of Pharmaceutical ChemistryFaculty of Pharmaceutical SciencesWe accept this thesis as conformingto the required standardTHE UNIVERSITY OF BRITISH COLUMBIAJune 20, 1994©Ric Michael Procyshyn, 1994In presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.(Signature)________________________________Department of PH i’1AC6L’7,c4I- S/&7JC&S’The University of British ColumbiaVancouver, CanadaDate - 1ø.-srDE-6 (2188)ABSTRACTThe acid pair hypothesis predicts the calcium affinity of the helix-loop-helix calcium binding motif based on the number and location ofacidic amino acid residues in chelating positions of the calcium-bindingloop region. This study investigates the effects of the number and positionof acidic residues in the loop region on calcium affinity and selectivityusing 33-residue synthetic models of single helix-loop-helix calciumbinding motifs.Increasing the number of acidic residues in the octahedrallyarranged chelating positions of the loop region from 3 to 4 by replacing anasparagine in the i-V position with an aspartic acid increases the calciumaffinity of the models between 2- and 38-fold. Differences in affinities aremore pronounced in the models containing an X-axis acid pair.The calcium affinities of peptide models containing 3 or 4 acidicresidues in chelating positions of the loop region and an X-axis acid pairare reduced when the residue in the +Z is changed from asparagine toserine. A similar reduction in calcium affinity occurs in the Z-axis acidpaired peptides when the —X chelating residue is changed from serine toasparagine.In order to increase interaction of the —X chelating residue with thecation, helix-loop-helix calcium binding motifs were synthesized containing3 and 4 acid residues in chelating position, with a glutamic acid replacingaspartic acid in the —X chelating position. The glutamate containing motifgave an unexpected 6-fold decrease in cation affinity for the 3 acid residueloop motif and a 46-fold decrease for the 4 acid residue loop motif. Toimprove calcium binding of the glutamate containing motifs, peptides wereIIsynthesized keeping glutamate in the —x position and inserting serine inthe +Z position to provide a hydrogen bonded system stabilizing theglutamate interaction with the cation. The serine residue further reducedcalcium affinity in both the 3 and 4 acid residue loop. These resultsindicate that glutamate and serine residues in the —X and +Z positionrespectively, can have a detrimental effect on calcium binding. However, innatural calcium binding proteins, glutamate in the —X chelating positioncan confer high affinity for calcium in helix-loop-helix calcium bindingmotifs, but this may be dependent on the environment created by as yetundetermined factors.Models with 3 acidic residues in chelating positions containing a Zaxis acid pair have greater calcium affinity than comparable peptidemodels with an X-axis acid pair. The presence of X- or Z-axis acid pairs incomparable peptides containing 4 acidic residues in chelating positionsdoes not greatly alter calcium affinity. Calcium selectivity residues in Xaxis acid paired peptides, whereas Z-axis acid paired peptides may exhibitboth magnesium- and calcium-induced structural changes. This ionselectivity may be explained by postulating that the Z-axis residue sidechains produce the initial, rate-limiting interactions with the cation, causinghydration shell destabilization and initiating the subsequent ligandinteractions.IllTABLE OF CONTENTSSection PageABSTRACT iiTABLE OF CONTENTS ivLIST OF TABLES viiiLIST OF FIGURES xLIST OF ABBREVIATIONS xiiiLIST OF AMINO ACID CODES xvPEPTIDE NOMENCLATURE xviACKNOWLEDGMENTS xviiiDEDICATION xixINTRODUCTION II. BACKGROUND III. GENERAL FEATURES OF THE HLH-CALCIUM BINDING MOTIF 41. Structural Overview 42. The 12 Residue Calcium Binding Loop Region 63. Calcium Coordination 10III. CONTRIBUTING FACTORS INFLUENCING CALCIUM BINDING TOTHE HLH-MOTIF 131. Calcium Ligands 132. Hydrogen Bonded Interactions Within the Calcium Binding LoopRegion 153. Hydrophobic Interactions of the x-Helices 17iv4. Dipole Moments.18IV. STRUCTURE-ACTIVITY RELATIONSHIP STUDIES OF CALCIUMAFFINITY IN HLH-MOTIFS 201.cc-Helices 202. Chelating Residues of the Loop Region 22a. +X Position 22b. +Y Position 23c. —X Position 24d. —Z Position 24e. +X, -i-Y, and +Z Positions 25f. +Y, +Z, and -X Positions 263. Nonchelating Residues of the Loop Region 27a. Position 2 27b. Position 6 28c. Positions 2, 4, and 10 28V. MAGNESIUMICALCIUM BINDING SITES 30VI. OBJECTIVES OF STUDY 33VII. MODEL 36MATERIALS AND METHODS 38I. MATERIALS 38II. EQUIPMENT 42III. METHODS 441. Solid Phase Peptide Synthesis (SPPS) 44a. Attachment of the First Amino Acid to the Resin 47b. Acetylation of the Peptide Resin 48c. Deprotection 49d. Active Ester Coupling 49e. Monitoring of SPPS (Ninhydrin Reaction) 50f. HF Cleavage 51g. Extraction of the Peptide 52V2. Purification of Peptides.52a. Dialysis 52b. HPLC Purification of Peptides 523. Determination of Peptide Concentration 544. Circular Dichroism 545. Calculations 56a. Free Calcium and Magnesium Concentrations 56b. Dissociation Constants 576. Polyacrylamide Gel Electrophoresis 58RESULTS 59I. PEPTIDE PURIFICATION 59II. CD ANALYSES 611. CaM:3(DNS) 612. CaM:3X Peptides 613. CaM:4X Peptides 704. CaM:3Z Peptides 755. CaM:4Z Peptides 816. CaM:4XZ Peptides 86Ill. DIMERIZATION 931. CD Analyses 932. Polyacrylamide Gel Electrophoresis of CaM:4Z(DDS), CaM:2/3,CaM:314, and TnC 96DISCUSSION 103I. STRUCTUREIACTIVITY RELATIONSHIP OF CALCIUM AFFINITY OFTHE HELIX-LOOP-HELIX CALCIUM BINDING MOTIF 103vi1. Evaluation of Serine and Asparagine in the +Z ChelatingPosition of X-Axis Acid Paired HLH-Calcium Binding Motifs 1042. Evaluation of Serine and Asparagine in the —X ChelatingPosition of Z-Axis Acid Paired HLH-Calcium Binding Motifs 1093. Evaluation of Glutamate and Aspartate in the —X ChelatingPosition of X-Axis Acid Paired HLH-Calcium Binding Motifs 1134. Contributions of Aspartate and Asparagine in the +Y ChelatingPosition on Calcium Affinity 1195. Differential Calcium Affinities of X vs. Z-Axis Acid Paired HLHCalcium Binding Motifs 1216. Evaluation of XZ-Axis Acid Paired HLH-Calcium Binding Motifs 1227. Evaluation of Glutamic Acid in the +Z Chelating Position 125II. MAGNESIUM BINDING TO THE HLH-CALCIUM BINDING MOTIFS 128III. DIMERIZATION 131IV. FUTURE DIRECTION 134SUMMARY AND CONCLUSIONS 136REFERENCES 142APPENDIX A 150APPENDIX B 155viiLIST OF TABLESTable # Description Page1. Semi-Permanent Protecting Groups for Boc-AminoAcids 502. Composition of PAGE Gels 583. Correspondence Between the Number of Residues in xHelical Regions and Mean Residue Ellipticity Values ofCaM:3(DNS) 634. Correspondence Between the Number of Residues in cHelical Regions and Mean Residue Ellipticity Values ofthe CaM:3X Peptides 675. Correspondence Between the Number of Residues in cHelical Regions and Mean Residue Ellipticity Values ofthe CaM:4X Peptides 726. Correspondence Between the Number of Residues in cHelical Regions and Mean Residue Ellipticity Values ofthe CaM:3Z Peptides 777. Correspondence Between the Number of Residues in a-Helical Regions and Mean Residue Ellipticity Values ofthe CaM:4Z Peptides 838. Correspondence Between the Number of Residues in a-Helical Regions and Mean Residue Ellipticity Values ofthe CaM:4XZ Peptides 899. Correspondence Between the Number of Residues in a-Helical Regions and Mean Residue Ellipticity Values ofCaM:4Z(DDS), CaM:4X(DSD) and a 1:1 Combination ofCaM:4Z(DDS) and CaM:4X(DSD) 9510. Physical Characteristics of CaM:4Z(DDS), Troponin CSite Ill, CaM:2/3, and CaM:3/4 Used in PolyacrylamideGel Electrophoresis 101vu’11. Correspondence Between the Number of Residues in a-Helical Regions and Mean Residue Ellipticity Values of AllSynthetic Analogs 10612. Amino Acid Composition of Purified CaM:3XPeptides 15513. Amino Acid Composition of Purified CaM:4XPeptides 15614. Amino Acid Composition of Purified CaM:3ZPeptides 15715. Amino Acid Composition of Purified CaM:4ZPeptides 15816. Amino Acid Composition of Purified CaM:4XZPeptides 15917. Amino Acid Composition of Purified CaM:3(DNS),CaM:2/3, CaM:3!4, and TnC Peptides 160ixLIST OF FIGURESFigure # Description Page1. The EF Hand 42. Schematic Representation of a Typical HLH-CalciumBinding Domain 53. Schematic Representation of an EF Hand CalciumBinding Site 74. Amino Acid Sequence of the 12 Residue Loop Region ofVarious Calcium Binding Proteins Utilizing the One LetterCode for Amino Acids 215. Scheme of Solid Phase Peptide Synthesis 456. Reverse Phase HPLC Purification Profile ofCaM:4Z(DES) 607. CD Spectra Showing Mean Residue Ellipticity ofCaM:3(DNS) 628. CD-Monitored Calcium Titration of CaM:3(DNS) 649. CD Spectra Showing Mean Residue Ellipticity ofCaM:3X(NSD), CaM:3X(NSE), CaM:3X(NND), andCaM:3X(NNE) 6510. CD-Monitored Calcium Titration of the CaM:3X Peptides... 6911. CD Spectra Showing Mean Residue Ellipticity ofCaM:4X(DSD), CaM:4X(DSE), CaM:4X(DND), andCaM:4X(DNE) 7112. CD-Monitored Calcium Titration of the CaM:4X Peptides... 7413. CD Spectra Showing Mean Residue Ellipticity ofCaM:3Z(NDN), CaM:3Z(NDS), and CaM:3Z(NES) 7614. CD-Monitored Calcium Titration of the CaM:3Z Peptides 7915. CD-Monitored Magnesium Titration of CaM:3Z(NDS) 80x16. CD Spectra Showing Mean Residue Ellipticity ofCaM:4Z(DDN), CaM:4Z(DDS), and CaM:4Z(DES) 8217. CD-Monitored Calcium Titration of the CaM:4Z Peptides 8518. CD-Monitored Magnesium Titration of CaM:4Z(DDS) andCaM:4Z(DDN) 8719. CD Spectra Showing Mean Residue Ellipticity ofCaM:4XZ(NDD), CaM:4XZ(NDE), CaM:4XZ(NED), andCaM:4XZ(NEE) 8820. CD-Monitored Calcium Titration of CaM:4XZ(NDD) andCaM:4XZ(NDE) 9121. CD-Monitored Calcium Titration of CaM:4XZ(NED) andCaM:4XZ(NEE) 9222. CD Spectra Showing Mean Residue Ellipticity ofCaM:4Z(DDS), CaM:4X(DSD), and a 1:1 Combination ofCaM:4Z(DDS), and CaM:4X(DSD) 9423. CD-Monitored Calcium Titration of CaM:4Z(DDS),CaM:4X(DSD), and a 1:1 Combination of CaM:4Z(DDS),and CaM:4X(DSD) 9724. Hill Plot of CaM:4Z(DDS), CaM:4X(DSD) and a 1:1Combination of CaM:4Z(DDS) and CaM:4X(DSD) 9825. Polyacrylamide Gel Electrophoresis of CaM:314 andCaM:4Z(DDS) 9926. Polyacrylamide Gel Electrophoresis of CaM:2/3,CaM:4Z(DDS), and TnC Ill 10027. Schematic Representation of the Twelve Residue CalciumBinding Loop Region of the Helix-Loop-Helix Motif 10328. Amino Acid Sequence of SCIII, LIIL, LII, CaM:4X(DSD),and the CD Site of Parvalbumin (Carp 4.25) 10829. Illustration of the Difference in Backbone Geometry WhenEither Serine or Asparagine is Found in the N-CapPosition Calcium 112xi30. Postulated Glu/Ser Side Chain Hydrogen Bonding andInteraction with Calcium 11431. Amino Acid Sequence of the X-Axis Paired Peptides:CaM:3X(NSD), CaM:3X(NSE), CaM:3X(NND),CaM:3X(NNE), CaM:4X(DSD), CaM:4X(DSE),CaM:4X(DND), and CaM:4X(DNE) 15132. Amino Acid Sequence of the Z-Axis Acid Paired Peptides:CaM:3Z(NDN), CaM:3Z(NDS), CaM:3Z(NES),CaM:4Z(DDN), CaM:4Z(DDS), and CaM:4Z(DES) 15233. Amino Acid Sequence of the XZ-Axis Acid PairedPeptides: CaM:4XZ(N DD), CaM:4XZ(NDE),CaM:4XZ(NED), and CaM:4XZ(NEE) 15334. Amino Acid Sequence of CaM:3(DNS), CaM:2/3, CaM:3/4,and Troponin C Site III 154xiiLIST OF ABBREVIATIONSAc - AcetylATP - Adenosine TriphosphateBHA - Benzhydrylamine ResinBoc - BenzyloxycarbonylBorn - BenzyloxymethylBO P - [Benzotriazol-1 -yI-oxy-tris-(dimethylamino)phosphoniumhexafluorophosphate]CaM - CalmodulinCD - Circular DichroismDCM - DichioromethaneDI EA - N, N-DiisopropylethylamineDMF - N,N-DimethylforrnamideDnp - 2,4-DinitrophenylEDTA - Ethylenediaminetetraacetic AcidEGTA - Ethyleneglycol-bis-(f3-Aminoethyl ether)N, N, N’, N’-tetraacetic AcidHF - Hydrofluoric Acidhlh - helix-loop-helixHOBt - I -HydroxybenzotriazoleHPLC - High Pressure Liquid ChromatographyKca - Dissociation Constant of CalciumKMg - Dissociation Constant of magnesiumiPrOH - Isopropyl AlcoholXIIIMOPS - 3-[N-Morpholine]propanesulfonic AcidMRW - Mean Residue WeightMW - Molecular WeightNMR - Nuclear Magnetic ResonancenOes - nuclear Overhauser effectsNTA - Nitrilotriacetic AcidSPPS - Solid Phase Peptide SynthesisSTnC - Skeletal Troponin CTCA - Trichloroacetic AcidTFA - Trifluoroacietic AcidTEE - TrifluoroethanolTnC - Troponin CTos - TosylxivLIST OF AMINO ACIDS CODESAmino Acid 3 Letter Code 1 Letter CodeAlanine Ala AArginine Arg RAsparagine Asn NAspartic Acid Asp DCysteine Cys CGlutamic Acid Glu EGlutamine Gin QGlycine Gly GHistidine His HIsoleucine lieLeucine Leu LLysine Lys KMethionine Met MPhenylalanine Phe FProline Pro PSerine Ser SThreonine Thr TTryptophan Trp WTyrosine Tyr YValine Val VxvPEPTIDE NOMENCLATURECaM:3X Peptides - Peptide analogs having three acid residues in the calciumbinding loop region and an acid pair at the vertices of the X-axis of theoctahedral arrangement of chelating atoms.CaM:4X Peptides - Peptide analogs having four acid residues in the calciumbinding loop region and an acid pair at the vertices of the X-axis of theoctahedral arrangement of chelating atoms.CaM:3Z Peptides - Peptide analogs having three acid residues in the calciumbinding loop region and an acid pair at the vertices of the Z-axis of theoctahedral arrangement of chelating atoms.CaM:4Z Peptides - Peptide analogs having four acid residues in the calciumbinding loop region and an acid pair at the vertices of the Z-axis of theoctahedral arrangement of chelating atoms.CaM:4)(Z Peptides - Peptide analogs having four acid residues in the calciumbinding loop region and acid pairs at the vertices of the X-and Z-axes of theoctahedral arrangement of chelating atoms.CaM:3(DNS) - Ac L1 CaM:(81-l 13) amideCaM:3X(NSD) - Ac N95,s7,D101,L109 CaM:(81-1 13) amideCaM:3X(NSE) - Ac N95,S97,E101,L109 CaM:(81-1 13) amideCaM:3X(NND) - Ac N95,N97,D101,L109 CaM:(81-l 13) amideCaM:3X(NNE) - Ac N95,N97,E101,L109 CaM:(81 -113) amideCaM:4X(DSD) - Acs7,D101,L109 CaM:(81-l 13) amideCaM:4X(DSE) - Ac S97,E101,L1 CaM:(81-1 13) amideCaM:4X(DND) - Ac D101,L109 CaM:(81-1 13) amideCaM:4X(DNE) - Ac E101,L1 CaM:(81-1 13) amideCaM:3Z(NDN) - Ac N95, D97,N101,L1 CaM:(81-1 13) amideCaM:3Z(NDS) - Ac N95, D97,L109 CaM:(81-1 13) amidexviCaM:3Z(NES) - Ac N95,E97, L1 CaM:(81-113) amideCaM:4Z(DDN) - Ac D97,N101,L1 CaM:(81-1 13) amideCaM:4Z(DDS) - Ac D97,L109 CaM:(81-113)amideCaM:4Z(DES) - Ac E97,L109 CaM:(81-1 13) amideCaM:4XZ(NDD) - Ac N95,D97,D101,L1 CaM:(81-1 13) amideCaM:4XZ(NDE) - Ac N95,D97,E101,L1 CaM:(81-113) amideCaM:4XZ(NED) - Ac N95,E97,D101,L1 CaM:(81-1 13) amideCaM:4XZ(NEE) - Ac N95,E97,E101,L1 CaM:(81 -113) amideCaM:2/3 - Ac L109 CaM(46-1 13) amideCaM:3/4 - Ac L1 CaM(81 -1 48) amideTnC III - Ac A98 STnC (90-123) amidexviiACKNOWLEDGMENTSResearch Supervisor: I would like to express my sincere gratitude to myresearch supervisor, Dr. Ronald Reid, for his guidance and support throughoutthis study.Supervisory Committee:Chair: Dr. G. Bellward (Pharmaceutical Sciences)Dr. S. Bandiera (Pharmaceutical Sciences)Dr. J. Diamond (Pharmaceutical Sciences)Dr. ft Molday (Biochemistry)I would also like to thank Dr. Grant Mauk (Department of Biochemistry,Faculty of Medicine, University of British Columbia) for the use of the Jasco J720CD spectrophotometer; Phil Owen (laboratory of Dr. lan Clark-Lewis, BiomedicalResearch Center) for synthesizing some of the CaM peptides; Peter Mills(laboratory of Dr. W. Palmer, Department of Animal Sciences, Faculty ofAgriculture, University of Manitoba) for performing the amino acid analyses; Dr.Tatsuya Fujimiya for providing a nonlinear curve fitting program; and mycolleagues in the laboratory of Dr. R. Reid for their technical support and usefuldiscussions.xviiiTO MY PARENTSxixINTRODUCTIONINTRODUCTIONI. BACKGROUNDCalcium is ubiquitous in living organisms and is involved either directly orindirectly in almost every cellular function. The resting free intracellular calciumconcentration in animal cells is between 100 and 300 nM, which is about 10,000-fold lower than extracellular calcium concentrations (Ashley and Campbell,1979; Blinks eta!., 1982). This large chemical gradient of calcium across theplasma membrane, along with the fact that intracellular free calcium canincrease rapidly in response to both extracellular and intracellular stimuli,creates an environment in which calcium can act as an intracellular regulator (fora review see Pietrobon et aL, 1990). Conversely, the extracellular calciumconcentration is relatively constant, thereby limiting its biological role to primarilythe stabilization of extracellular target molecules. The target molecules forextracellular calcium are enzymes, and apart from increasing thermal stability,calcium binding to these molecules also confers an increased resistance toproteolysis. Calcium binding may also serve a catalytic function in suchextracellular enzymes as phospholipase A2 and a-amylase (Dijkstra et a!., 1981;Matsuura etah, 1984).In contrast to extracellular calcium, intracellular calcium plays a regulatoryrole and is involved in several processes that include nerve impulsetransmission, muscle contraction and cell motility, metabolic regulation, celldivision and growth, secretion, and membrane permeability (for review seeRasmussen and Rasmussen, 1990). These diverse functions are, in part,regulated by a family of intracellular target molecules referred to as the helix-loop-helix (hlh) calcium binding proteins (for review see Strynadka and James,IINTRODUCTION1989). Calcium binding to this family of proteins induces the rearrangement oftheir tertiary structures such that the proteins can interact with their targetmolecules and elicit a response. Calmodulin and troponin C are classicexamples of this phenomenon. Calcium binding to these proteins results in arelative change in the disposition of the a-helices that flank the calcium bindingsite (Liu and Cheung, 1976; Klee, 1977; Leavis etah, 1978; Potter and Gergely,1975; Sin et al., 1978), thereby exposing hydrophobic patches on the surfaces ofthe molecules that enhance interactions with their effector molecules (Cheung,1980; Klee et a!., 1980; Inagraham and Swenson, 1984; Wang and Cheung,1985; Hasselgrove, 1973). In this sense, hlh-calcium binding proteins can bereferred to as calcium modulated proteins. Parvalbumin and calbindin9k,whichare also members of the hlh-calcium binding protein family, do not interact withother proteins; rather, they appear to act as calcium buffers or calciumtransporters involved in the regulation of intracellular calcium concentrations(Wnuk eta!., 1982; Gillis, 1985; Wasserman and Fullmer, 1982; Wasserman eta!., 1983; Levine and Williams, 1982; Schachter, 1980; Freund, 1982).HIh-calcium binding proteins have been extensively characterized;however, the molecular mechanism of calcium binding to these proteins is stillnot understood. Although these calcium binding sites exhibit similar structuralfeatures and show high sequence homology, great differences exist in theiraffinities for calcium. To understand the mechanism of cation binding whichdescribes these differences in calcium affinity, investigators have examined suchfactors as the high propensity of 3-turns in the N-terminal region of the calciumbinding loop (Vogt et a!., 1979), net charge on the ligands, 13-sheethydrophobicity, helical hydrophobicity (Sekharudu and Sundaralingam, 1988),secondary structure predictions (Boguta et al., 1988), cooperative association ofpaired calcium binding sites (Potter, 1977), and the number and location of2INTRODUCTIONacidic amino acid residues in chelating positions (Reid and Hodges, 1980;Marsden et a!., 1988). Although these studies have increased our knowledgesomewhat, the structural similarities and primary structure homology among thecalcium binding sites in these proteins belies the diversity of cation affinities andselectivities demonstrated by these sites.There are still many questions that need to be answered in order to havea better understanding of the molecular mechanism of cation binding to the hlhfamily of calcium binding proteins. What constitutes an ideal calcium bindingsite in an hlh-motif? Which calcium ligands provide optimal calcium binding?Do the residues of the flanking a-helices or the loop region impose the mostinfluence on calcium binding? Which hydrogen bond interactions optimizecalcium binding? What determines differential calcium affinities? Whatdetermines calcium-selective binding sites as opposed to sites that also bindmagnesium? The answers to these questions should bring us closer to apredictive ability of cation affinity and selectivity upon inspection of the primarystructure of an hlh-motif.3INTRODUCTIONII. GENERAL FEATURES OF THE HLH-CALCIUM BINDING MOTIF1. Structural OverviewKretsinger, and Nockolds, (1973) characterized the first known hlhcalcium binding motif from the crystal structure of carp parvalbumin. The tertiarystructure of a single hlh-motif has been referred to as the “EF hand”, with theindex finger corresponding to the E helix, the curled middle finger correspondingto the loop region, and the thumb corresponding to the F helix of the thirdcalcium binding site of carp parvalbumin (Figure 1). To date, more thanFigure 1. The EF Hand. The index finger and the thumb represent the E and F helices,respectively. The vertices of the octahedral coordination sphere about the calcium cationare designated by X, Y, Z, —Y, —X, and —Z.(Taken from Kretsinger and Wasserman, 1980)4INTRODUCTION153 hlh-calcium binding proteins have been identified, and based on theirprimary structures, have been placed into twelve distinct subfamilies (Moncrief,et a!., 1990). Although the primary structures of these proteins are known, thecrystal structures have only been determined for calmodulin (Babu et aL, 1985,1988; and Chattopadhyaya, et aL, 1992), turkey skeletal troponin C (Herzberg,and James, 1985,1988), chicken skeletal troponin C (Sundaralingam, et al.,1985; Satyshur et a!., 1988), carp parvalbumin (Kretsinger and Nockolds, 1973;Moews and Kretsinger, 1975; Kumar et aL, 1990), pike parvalbumin (Declercq,et a!., 1988, 1991), vitamin D-dependent calcium binding protein (Szebenyi eta!., 1981; and Szebenyi and Moffat, 1986), and oncomodulin (Ahmed et a!.,1990). The number of helix-loop-helix motifs within these calcium bindingproteins range from two to four.Figure 2 is a diagrammatic representation of a typical calcium bindingdomain made up of two hlh-calcium binding motifs related to each other by a twoADH IBFigure 2. Schematic Representation of a Typical HLH-Calcium Binding Domain. Thereare two hlh-motifs (one represented by helix A - loop I - helix B, and the other representedby helix C - loop II - helix 0). The solid dots represent the location of calcium binding.(Taken from Strynadka and James, 1989).5INTRODUCTIONfold axis of symmetry and joined by a linker region of 5-10 residues in length.The c-helices are 10-12 residues in length and flank both ends of the 12 residuecalcium binding loop region. In all of the calcium binding proteins mentionedabove, the hlh-motifs exist in pairs; however, one of the motifs may not have anyaffinity for calcium. Such is the case with cardiac troponin C and parvalbumin(van Eerd and Kawasaki, 1972; Collin et a!., 1977; van Eerd and Takahashi,1975 and 1976).In the calcium bound form, the paired hlh-calcium binding motifs areassociated with each other via their loop regions and their flanking a-helices. Inthe case of the loop regions, there is a short anti-parallel p-sheet (the only 3-sheet structure to occur within hlh-calcium binding proteins) encompassingresidues 7 to 9 of each calcium binding loop. The amphipathic a-helices withinthe calcium binding domain pack together such that their hydrophobic portionsface inward creating a hydrophobic core within the domain. Strynadka andJames (1989) referred to the calcium binding domain (two hlh-motifs) as “cup-shaped” with the inside surface of the cup being rich in hydrophobic resides (ahelices), the bottom of the cup containing the calcium binding loops, and theoutside of the cup being rich in hydrophilic residues (cz-helices).2. The 12 Residue Calcium Binding Loop RegionA common feature among the calcium binding loop regions are Asx turns.This type of secondary structure involves a hydrogen bond interaction between aside chain oxygen atom (hydrogen bond acceptor) at position n, and a mainchain NH (hydrogen bond donor) of a residue in position n+2. In order ofdecreasing frequency, the most common residues initiating the Asx turn areaspartic acid, asparagine, serine, and threonine (Richardson, 1981; Rees et aL,6INTRODUCTION1983; Baker and Hubbard, 1984). The Asx turns, along with the conservedglycine residue in position 6 (Figure 3), are critical factors governing thebackbone conformation of the calcium binding loop region.6810Figure 3. Schematic Representation of an EF Hand Calcium Binding Site. The solid linerepresents the loop region whereas the zigzag lines represent the helices. The numbersrepresent the 12 amino acid residues located within the loop region starting from the N-terminal. The letters designate the paired coordinating residues that chelate with calciumin a near octahedral arrangement.Calcium is chelated via seven ligands belonging to the residues inpositions 1 (+X), 3(+Y), 5(+Z), 7(—Y), 9(—X), and 1 2(—Z) of the loop region (Figure3). The +X position is occupied by an invariant aspartic acid in all known loopsequences. X-ray crystallography has shown that one of the carboxylate oxygenatoms of this invariant residue ligates calcium while the other is involved in ahydrogen bonded interaction with the main-chain NH of an invariant glycineresidue located in position 6 (Kretsinger and Nockolds, 1973; Szebenyi andMoffat, 1986; Babu eta!., 1988; Herzberg and James 1985; and Satyshur etaL,5 732IIINH CO7INTRODUCTION1988). This observation was also confirmed using isotope labeling andmultidimensional NMR on cardiac troponin C (sites II, Ill, and IV), skeletaltroponin C (sites Ill and IV), calmodulin (sites l-lV), calbindin D9k (site II) and ahomodimer of the fourth calcium binding site of skeletal troponin C (Krudy et a!.,1992; Tsuda eta!., 1988, 1990; lkura etal., 1987, 1990; Kordel eta!., 1989; Kayet a!., 1991). In the absence of calcium, this hydrogen bond between thecarboxylate of the invariant aspartate and glycine (position 6) has also beendemonstrated in site II of cardiac troponin C (Krudy et a!., 1992), sites I and II ofskeletal troponin C (Tsuda et a!., 1990), and sites I, Il, and Ill of calmodulin(Ikura eta!., 1987).The +Y position of the calcium binding loop region is most commonlyoccupied by either an aspartic acid or an asparagine, with aspartic acid beingmost prevalent (Marsden et a!., 1990). The +Z position also shows a preferencefor aspartic acid followed by serine and asparagine (Marsden et a!., 1990).Crystallographic studies of the calcium-saturated intestinal calcium bindingprotein have shown that the side chain oxygen of the residue occupying the +Zchelating position forms a hydrogen bond with the main chain amino group atposition 7(—Y) (Szebenyi and Moffat, 1986).Several amino acid residues have been found at position 9(—X), howeveronly the carboxyl oxygen of a glutamic acid residue at this position has beenshown to ligate calcium directly (Kretsinger and Nockolds, 1973). When thisposition is occupied by a residue other than glutamic acid, the side chain doesnot coordinate calcium directly, but rather one or two water molecules are foundbridging the amino acid with the calcium cation (Strynadka and James, 1989;Satyshur, eta!., 1988).The residue at position 1 2(—Z) is an invariant glutamic acid. The role ofthe glutamic acid is two-fold: firstly, the two carboxyl groups of the side chain8INTRODUCTIONinteract with calcium in a bidentate manner, and secondly, the acidic nature ofthe residue at this particular location contributes to the structural integrity of theh/h-motif and has been speculated to play an important role in stabilizing theexiting a-helix (Maune eta!., 1992b).Finally, the residue at the 7(—Y) coordinate is also involved in ligatingcalcium. However, in this case, the side chain of this residue is directed awayfrom the calcium binding site, thereby resulting in calcium ligation via thebackbone carbonyl oxygen.The nonchelating residues of the loop region are found in positions 2, 4,6, 8, 10, and 11 (Figure 3). The residue found in position 2 is not highlyconserved; however, the presence of the basic lysine residue is common(Marsden et al. 1990). In troponin C site IV, the side chain of the lysine residuein this position points towards the C-terminus of the incoming a-helix andcontributes to the stabilization of the helix dipole (Strynadka and James, 1989).This interaction may explain why lysine residues are more commonly found inposition 2 than other amino acid residues.The most frequently occurring residue in position 4 is glycine (Strynadkaand James, 1989; Marsden eta!., 1990). The amide hydrogen of the residue inthis position participates in two hydrogen bonds involving both side chainoxygen atoms of the invariant aspartate in position 1.Similar to position 4, position 6 contains a conserved glycine residue. Asmentioned previously, the main-chain NH forms a hydrogen bonded interactionwith the nonchelating oxygen atom of the invariant aspartate in position 1 (+X).The absence of a bulky side chain is unique to glycine and offers the loop regionadded flexibility. This lack of conformational restriction enables the loop regionto make a 900 turn, thereby allowing proper alignment of the remainingcoordinating residues with calcium.9INTRODUCTIONThe nonchelating residue found in position 8 is usually hydrophobic, withisoleucine being the most common (Marsden et aL, 1990). The residue at thisposition is also the central residue of the short p-sheet found between the twoloop regions of a pair of h/h-calcium binding motifs. Strynadka and James(1989) have implicated this region within the loop in the mechanism ofcooperativity.The nonchelating amino acid in position 10 is the first residue of theexisting x-heIix. The type of residue occupying this position vanes between thevarious hlh-calcium binding proteins; however, there is a high frequency ofaromatic residues at this position in the second and fourth loops of a calciumbinding domain. In this case, the aromatic residues are involved in the formationof an aromatic cluster with residues of the flanking a-helices (Babu eta!., 1988;Satyshur eta!., 1988).Finally, the nonchelating residue found in position 11 is most often anegatively charged aspartic or glutamic acid, although a positively chargedlysine residue is the third most commonly found amino acid in this position(Marsden et a!., 1990). The negatively charged carboxylate group of eitheraspartate or glutamate electrostatically stabilizes the N-terminus of the exiting xhelix.3. Calcium CoordinationStudies on calcium complexes involving small molecules have showncoordination numbers for calcium ranging from 4 to 12 (Brown, 1988). Inconsidering cation-ligand complexation, both the ionic radius and chargeassociated with the cation and ligand will have an effect on the arrangement10INTRODUCTION(geometry) and the number (coordination) of ligands that can bind to the cation.A useful indicator of the packing requirements of the ligand with the cation canbe ascertained from the radius ratio (r÷/r—) of the cation-ligand complex that isderived by dividing the ionic radius of the cation (1 A for calcium) by the ionicradius of the ligating species. This value also allows an estimation of thecoordination number of the cation (Huheey, 1972).Considering the first crystal structure of a calcium binding protein (carpparvalbumin), it was proposed that calcium was coordinated by six oxygen atomsin an octahedral arrangement (Kretsinger and Nockolds, 1973). However,increased resolution of crystal structures of various calcium binding proteins(including parvalbumin) has established that calcium is coordinated via sevenoxygen atoms (Declercq et a!., 1988; Strynadka and James, 1989; Swain et a!.,1989; and Kumar et aL, 1990). The discrepancy was a result of the lowerresolution structure of carp parvalbumin not being able to identify the bidentateinteraction with calcium involving both carboxylate oxygen atoms of the invariantglutamate in the —z position. Thus, the arrangement of ligands within a calciumbinding site can best be described as a pentagonal bipyramid with an averageoxygen-calcium distance of 2.4 A (Strynadka and James, 1989). Thecoordinating oxygen atoms in the equatorial pentagonal plane are contributed bythe residues at positions 3(+Y), 5(+Z), 7(—Y) and 12(—Z) (the —Z contributes abidentate ligand). The apical vertices are from residues located at positions1 (+X) and 9(—X).Crystal structures have shown that the calcium coordination sphere of thehlh-motif does not show ideal polygon geometry, but rather reflects a distortedpentagonal bipyramid. Deviation from an ideal pentagonal bipyramidal geometryis, in part, the result of the bidentate ligands of the invariant glutamic acid in the—Z chelating position. It is not possible for the two carboxylate oxygen atoms of11INTRODUCTIONthe invariant glutamic acid to be positioned at two vertices of an equatorial planeof a pentagonal bipyramid. The distance between two vertices on the equatorialplane of a pentagonal bipyramid is 2.82 A, whereas the distance between thetwo carboxylate oxygen atoms is only 2.2 A. As a result, one of the bidentateoxygen ligands of the invariant glutamic acid lies out of the equatorialpentagonal plane by 1.2 A (Strynadka and James, 1989). It has been suggestedthat the presence of the bidentate ligands may, in part, account for the calciumselective nature of some calcium binding sites. The calcium binding site couldconceivably be modified to accommodate magnesium binding in an octahedralfashion if the glutamate is transformed into a unidentate ligand by changing thedisposition of the side chain such that the plane of the carboxyl functional groupis perpendicular to the equatorial ligand plane (Strynadka and James, 1989).12INTRODUCTIONIII. CONTRIBUTING FACTORS INFLUENCING CALCIUM BINDING TO THEHLH-MOTIF1. Calcium LigandsA ligand is any atom having the capacity to act as a donor of electronsand form one or more coordinate bonds. As mentioned previously (Section 11.2),the calcium ligands of the hlh-motif originate from the coordinating residues inpositions 1 (+X), 3(+Y), 5(+Z), 7(—Y), 9(—X), and 1 2(—Z). To date, all of thecalcium ligands in hlh-motifs are oxygen atoms from either amino acid sidechains, backbone carbonyl groups, or water molecules (McPhalen et aL, 1991).The most common amino acid to coordinate with calcium in an hlh-motif isaspartic acid that is followed by glutamic acid, asparagine, serine, threonine,and glutamine (Marsden et aI., 1990). Apart from forming a coordinate bond withcalcium, the functional group of the acidic residues serves two other roles: 1) itacts as a counterbalance to the charge associated with calcium, and 2) itparticipates in hydrogen bonded interactions which aid in the stabilization of thecalcium complex. Similar to the acidic residues, the amide functional groups ofasparagine and glutamine are also involved in hydrogen bonded interactions.However, unlike the acidic residues the amide group will have a lesser effect oncounterbalancing the charge associated with the cation. The lower incidence ofserine and threonine residues coordinating with calcium may be explained inpart, by the fact that they do not have any other functional groups on their sidechains that could contribute to the stabilization of the calcium binding complex.All h/h-calcium binding motifs within the calmodulin family have at leastone main-chain carbonyl oxygen atom coordinating with calcium. In the case ofthe first calcium binding site of calbindin, there are four such backbone carbonyl13INTRODUCTIONoxygen ligands (Szebenyi and Moffat, 1986). This aberrant feature is due to twoextra amino acids in the calcium binding site that alters the conformation in theN-terminal region of the loop. Completing the list of calcium ligands in hIhcalcium binding sites is water. Except for the CD calcium binding site ofparvalbumin, water serves as a bridge between the calcium cation and the sidechain of the residue in position 9(—X). The absence of a water molecule in theCD calcium binding site of parvalbumin is probably due to the longer side chainof the glutamic acid in position 9(—X) which ligates the cation directly (Kretsingerand Nockolds, 1973; Kumar eta!., 1990; and Declercq eta!., 1991).In a review of calcium-ligand interactions, Einspahr and Bugg (1984)found little variation in calcium-ligand distances among various ligand types.These ranged from 2.42 A for water, 2.36 A for carbonyl oxygens, 2.38 A forunidentate carboxylate oxygen atoms, and 2.53 A for a bideritate carboxylateoxygen atom. In another survey of 40 calcium binding proteins, the overall meancalcium-ligand distance was found to be 2.4 A with no significant differencebetween the various ligands (McPhalen eta!., 1991).Thus, the interpretation of varying calcium affinities among the h!hcalcium binding proteins must take into account the contributing effects of thecalcium ligands. Differences in ionic radii and charge between the variousligands will dictate the geometry of the cation coordination sphere thatsubsequently effect cation binding and affinity. Along with the influences ofdirect interaction with the cation, the hydrogen bonded interactions involving thecoordinating side chains will also have an effect on calcium affinity.14INTRODUCTION2. Hydrogen Bonded Interactions Within the Calcium Binding LoopRegionOne of the major intramolecular interactions contributing to proteinstabilization is thought to come from hydrogen bonds. Hydrogen bonds betweenuncharged donors and acceptors can contribute from 0.5 - 1.8 kcal/mole to theassociation energy, and hydrogen bonds between charged groups cancontribute up to 6 kcal/mole (Page, 1984; Fersht et a!., 1985; Fersht, 1987).Apart from acting as a uni- or bidentate ligands(s) for calcium, the side chaincarboxylate oxygens of aspartate and glutamate also participate in at least one,but usually two, three or more hydrogen bonded interactions with main-chainamide nitrogen atoms, hydroxyl, amide, and basic side chain moieties, andordered solvent molecules (McPhalen etaL, 1991). An interaction with a mainchain NH group is by far the most common. Quite commonly, this hydrogenbond is of the n to n+2 Asx type, however, n to n-i-4, and n to n+5 interactionsare also found as in the case of the invariant aspartic acid in the +X chelatingposition of parvalbumin, calbindin, and chicken troponin C (Szebenyi and Moffat,1986; Satyshur et a!., 1988). The hydrogen bonded interactions associated withAsx turns aid in stabilizing the cluster of negatively charged oxygen atoms in theloop region, by directing the ligating side chains towards the vertices of thecoordinating sphere thereby reducing electrostatic repulsion.Water is the second most frequent hydrogen-bond donor for the calciumliganding carboxylates of aspartate and glutamate. In many cases, the watermolecule serves as a calcium ligand. The carboxylate ligands have also beenfound to hydrogen bond directly to the side chains of asparagine, glutamine,lysine, and arginine. Taken together, the prevalence of hydrogen bondedinteractions involving aspartate and glutamate calcium ligands serve to: 1) aid in15INTRODUCTIONthe proper orientation of carboxylate side chain oxygen atoms to coordinate withcalcium, 2) stabilize neighboring amide or water molecules which are calciumligands themselves, 3) stabilize the main chain conformation of the calciumbinding loop region, thereby allowing proper alignment of other calciumcoordinating ligands, and 4) counteract the repulsive electrostatic forcesassociated with neighboring negatively charged amino acid residues locatedwithin the loop region.The amide functional groups of asparagine and glutamine that participatein calcium coordination can also accept or donate hydrogen bonds. Inspectionof hlh-calcium binding sites reveals that the majority of amide oxygen atomsinvolved in calcium coordination also receive a hydrogen bond from a mainchain nitrogen atom in the form of an n to n+2 hydrogen bonding pattern of anAsx turn (McPhalen et a!., 1991). On the other hand, the side chain NH group isprimarily involved in hydrogen bonded interactions with the side chains ofneighboring aspartate and glutamate residues. There is no evidence of the sidechain NH group interacting with either a coordinating water molecule or a main-chain backbone carbonyl group. As a result of the limitations placed upon theamide functional group, asparagine and glutamine play a smaller role in proteinstabilization compared to aspartate or glutamate.Serine and threonine are still further limited in their abilities to stabilize astructure, since they have fewer atoms in their side chains capable of donatingor accepting a hydrogen bond. Some degree of stability, however, is offered bythe polar side chains of serine and threonine interacting with charged residuesthat ligate directly with calcium. Thus, it appears that the major contributingfactors determining the hydrogen bonded networks within an hlh-calcium bindingsite are the side chain functional groups that coordinate with calcium. Since hlhcalcium binding proteins have different amino acid residues in coordinating16INTRODUCTIONpositions, it is not surprising that there are differences in the hydrogen bondednetworks among the various calcium binding sites. It is conceivable that thesedifferences in hydrogen bonded networks account, in part, for the differences incalcium affinities seen among the hlh-calcium binding proteins. Thus, anyexplanation of calcium affinity would be incomplete if it did not take into accountthe influences of hydrogen bonding within the calcium binding loop region.3. Hydrophobic Interactions of the a-HelicesHydrophobic bonding is one of the most structure-stabilizing forces foundin proteins. Free energy of stabilization from buried hydrophobic surface areashave been estimated at 22-25 cal/A2 (Richards, 1977). Upon calcium binding,hlh-proteins undergo a conformational change such that the hih-motifs of acalcium binding domain pack against each other producing a hydrophobic core.Hydrophobic interactions between the first and fourth helices within a calciumbinding domain are primarily between aromatic residues in an edge to facemanner approximately perpendicular to each other. This aromatic cluster isfound in all calcium binding domains of bovine brain calmodulin, skeletaltroponin C, and bovine intestinal calcium binding protein and is made up of threephenylalanine residues (Babu eta!., 1988; Satyshur etal., 1988; Szebenyi andMoffat, 1986). One of the phenlyalanines is located four residues before thetermination of the first a-helix. The second and third phenylalanine residues arelocated in position 10 and immediately following the —Z chelating residue of thesecond loop region. An aromatic cluster is also found in carp parvalbumin,however the residue in position 10 of the second loop region is valine (Kumar eta!., 1990).17INTRODUCTIONUnlike the hydrophobic interaction of the first and fourth a-helices, thehydrophobic interactions of the second and third a-helices involve aliphaticresidues. In addition, the hydrophobic interactions between the second andthird a-helices are not as highly conserved as the hydrophobic interactions ofthe first and fourth a-helices. In any event, the presence of these structurestabilizing hydrophobic interactions between the a-helices of adjacent hlh-motifsalmost certainly plays an important role in influencing the stability of the calciumbound complex. This being the case, one can only speculate on what effect, ifany, the removal of one or all of these hydrophobic interactions would have oncalcium affinity.4. Dipole MomentsIn general, the c-helices of hlh-motifs are amphipathic and arranged in ann to n+4 hydrogen bonding pattern typical of a-helices. The different polaritiesbetween amide and carbonyl groups within an a-helix result in a large net dipolemoment that is aligned along the helical axis with a partial positive charge at theamino end and a partial negative charge at the carboxyl end (Wada, 1976; Holet aL, 1978). The strongest interaction to occur between an ion and a helixdipole is when the ion lies on the axis of the helix near the oppositely chargedend of the a-helix. In the case of the hlh-motif, the negatively charged ends ofthe helix dipole are not involved in direct or strong interactions with the calciumcation as a result of both flanking a-helices pointing away from the calcium suchthat the cation lies 6 and 8 A away from the helical axis (McPhalen et a!., 1991).Thus, the helix dipole of the h/h-motif does not appear to contribute to balancingthe charge on the cation or stabilization of the coordination structure.18INTRODUCTIONApart from interacting with an ion, a helix dipole can also serve tost?bilize adjacent anti-parallel x-helices via electrostatic stabilization to thetertiary structures. This type of helix-helix interaction can increase the stabilityof a structure by 5-7 kcal/mole (Sheridan et a!., 1982). As an example, the anti-parallel disposition of the flanking cz-helices of a calcium-free hlh-motif is moreenergetically stable than the a-helices of the calcium-bound hlh-motif in whichthe cL-helices have crossing angles ranging from 84° to 109° at their ends(Strynadka and James, 1989). Helix dipole interactions are also limited betweenhelices of neighboring hlh-motifs that have an average crossing angle of 118°.On first impression, it would appear that the helix dipole does not have anyeffect on calcium affinity; however, the electrostatic forces resulting from thehelix dipoles in the calcium free state must be overcome upon calcium binding.Thus, the relative anti-parallel alignment of the flanking a-helices in the calciumfree site may determine in part, the degree of calcium binding.19INTRODUCTIONIV. STRUCTURE-ACTIVITY RELA TIONSHIP STUDIES OF CALCIUMAFFINITY IN HLH-MOTIFSAs a result of amino acid substitutions (via-site directed mutagenesis orsolid phase peptide synthesis), a wealth of information has been obtained withrespect to the nature of calcium binding to hlh-calcium binding motifs. Calciumaffinity will be examined in terms of the influences of the chelating andnonchelating residues of the loop region as well as residues located within thex-helices flanking the loop region. Figure 4 provides the amino acid sequencesof the loop regions of the calcium binding proteins discussed below.1. a-HelicesSite specific replacement of the hydrophobic residues at positions 45-49of helix B in skeletal troponin C (V45T, M46Q, M48A, and L49T) with lesshydrophobic residues all showed an increase in the calcium sensitivity of thefluorescence emission spectra using a reporter mutant (F29W) (Pearlstone, etaL, 1992). These results are in agreement with the structural transition from a2Ca-- state to a 4Ca2-- state model proposed by Herzberg et a!. (1986).According to this model, the binding of calcium to the C-terminal calcium bindingsites results in the exposure of a number of hydrophobic side chains to thesolvent. The exposure of these hydrophobic residues is energeticallyunfavorable and is speculated to be compensated for by calcium binding at theN-terminal domain. In view of this, reducing the number of solvent-exposedhydrophobic residues should decrease the energy needed to maintain thecalcium-occupied structure or, in other words, the affinity for calcium at thisdomain should increase as was seen above (Pearlstone eta!., 1992).20INTRODUCTIONPosition# 1 2 3 4 5 6 7 8 9 10 11 12-Y -zBovine BrainCalmodulinSitel D k D g N g T i T t k ESitell D a D g N g T i D f p ESitelli D k D g N g Y I S a a ESiteIV Di D g D g QvN ye ERabbitSkeletalTroponin CSitel D a D g G g D I S v k ESitell D e D g S g T I D f e ESiteill D k N a D g Y I D i e ESiteIV D k N n D g R i D f d ERatParvalbuminSitell D k D k S g F I E e d ESitelil D k D g D g K I G v e ECalbindin a pSitel A””k E g D’”n 0 I S k e ESitell D k N g D g E v S f e EOnco modul inSitell D n D q S g Y I D g d ESiteill D n D g D g K I G a d ESCIII D k N a D g Y i D I e ELIIL D k D aS g Ti Die ELII D e D g S g Ti D feEFigure 4. Amino Acid Sequence of the 12 Residue Loop Region of Various CalciumBinding Proteins Utilizing the One Letter Code For Amino Acids.21INTRODUCTIONda Silva et a!. (1993) also constructed two troponin C mutants similar tothose mentioned above (V45T, and M48A). In this case, however, calciumaffinity was determined directly by flow dialysis. The M48A mutant had a 2.6-fold increase in calcium affinity, and the V45T had a 5.1-fold increase in calciumaffinity. As in the previous example, reducing the hydrophobic residues exposedto the solvent resulted in an increase in calcium affinity. Similarly, Sekharuduand Sundaralingam (1988) noted an inverse correlation between calcium affinityand the number of hydrophobic triplets in the flanking x-helices. Whereas thehigh affinity C-terminal domain of troponin C has eight hydrophobic triplets, thelower affinity N-terminal domain contains 13.Amino acid substitutions in the linker region between two hlh-motifs havealso been shown to effect calcium affinity. Golosinska et a!. (1991) substituted athreonine residue by isoleucine at position 130 of troponin C. Refined crystalstructures of troponin C from both chicken and turkey muscle have shown thatthis threonine is extensively involved in a hydrogen bonded network thatincludes ordered water molecules and residues in the N-terminal region of theG-helix (Satyshur et al., 1988; Herzberg and James, 1988). Disrupting thesehydrogen bonds by replacing the threonine with isoleucine resulted in a 2.6-folddecrease in calcium affinity. Thus, destabilization of a hydrogen bond networkdistant to the calcium binding site (>20 A) can have an influence on calciumaffinity.2. Chelating Residues of The Loop Regiona. ÷X PositionThree mutants of cardiac troponin C were engineered by replacing theinvariant aspartic acid residue located in the ÷X chelating position of calcium22INTRODUCTIONbinding sites II, Ill, and IV with an alanine residue (Putkey, et a!., 1989; Negele,et aL, 1992). The changes resulted in loss of calcium binding at these sites.This suggests that perhaps the residue in the +X cheating position must beacidic, however, other changes would be necessary to confirm this. Not onlydoes the aspartic acid in the ÷X position provide one of the ligands for calciumbinding, but it is also involved in a strong hydrogen bond with the amidebackbone proton of the glycine residue of position 6 (Strynadka and James,1989; Krudy eta!., 1992).Babu et a!. (1992), utilizing protein engineering, replaced the invariantaspartic acid of the +X position in rabbit skeletal troponin C site II with eitherglutamate or asparagine. Both substitutions afforded functionally inactiveproteins. In both cases, the calcium binding capacities were reduced by 1 moleof calcium per mole of troponin C. Thus, it appears that aspartic acid is requiredat the +X position for effective calcium binding. The requirement probablyreflects its extensive participation in hydrogen bonding of its carboxylate oxygenatoms with several residues within the loop as well as the calcium cation(Strynadka and James, 1989).b. +Y PositionTwo analogs of rabbit skeletal troponin C were prepared in which theaspartic acid residue found in the +Y position of site II was replaced by eitherglutamic acid or alanine (Babu eta!., 1992). In both cases, the mutated proteinbound only three, rather than four moles of calcium per mole of protein. Asmentioned previously (Introduction 11.2.), the +Y chelating position is almostalways occupied by either aspartic acid or asparagine. In fact, of 165 calciumbinding proteins examined by Marsden et a!. (1990), only four proteins havebeen found with a glutamate residue and none with an alanine residue in this23INTRODUCTIONposition. It appears that this position has requirements which only aspartic acidand asparagine can accommodate.c. —x PositionIn an attempt to make the CD calcium binding site of rat oncomodulinmore like the equivalent high affinity calcium-magnesium binding site of carpparvalbumin, the aspartate residue in the —X chelating position was replacedwith a glutamate residue using site-directed mutagenesis (Williams et a!., 1987;Golden et a!., 1989; Hapak, et aL, 1989, MacManus et aL, 1989). Thissubstitution resulted in a modest change in the calcium dissociation constant(0.78 M for the native protein and 0.55 jiM in the mutated protein) (Hapak, etaL, 1989). However, using Lu3 as a calcium probe, there was a five-foldincrease in affinity (Golden et a!., 1989). The fact that the calcium binding sitehad a much greater affinity for the lanthanide suggests that lanthanides may notbe acceptable as calcium probes and information obtained using lanthanidesshould be interpreted with caution.d. —z PositionTwo mutant proteins of rabbit skeletal troponin C were engineered suchthat the invariant glutamic acid found in the —Z chelating position of the firstcalcium binding site was replaced with either an aspartic acid or a glutamineresidue (Babu et a!., 1992). Radiolabelled studies showed that both analogsbound approximately three moles of calcium per mole of protein compared tofour in the wild type. The fact that the glutamine residue provides only onecalcium ligand compared to the bidentate ligands of glutamate (Strynadka andJames, 1989), perhaps explains the inability of glutamine to chelate calcium. On24INTRODUCTIONthe other hand, the shorter side chain of aspartic acid may be responsible for itsinability to chelate with calcium.In another experiment using Drosophila melanogaster calmodulin,consecutive replacement of the invariant glutamic acid residue (—Z position) ineach of the four calcium binding sites with either a glutamine or lysine eliminatedcalcium binding to the mutated site. As well, there was loss of cooperativecalcium binding between paired sites and a decrease in x-helical structure(Beckingham, 1991; Maune eta!., 1 992a, 1 992b).e. +X,+Y, and +Z PositionsEight 13-residue synthetic analogs of rabbit skeletal troponin C site Illwere synthesized such that all possible combinations of aspartic acid in the +X,+Y, and +Z chelating positions were represented either individually or incombination (Marsden eta!., 1988). Using La3 as a calcium probe, the resultsshowed a general trend of increased La3-’- affinity (1.1 mM - 4 M) as the numberof aspartic acid residues increased and that the aspartate residues in thesepositions contributed equally to La3-’- affinity. However, aspartate residues inneighboring chelating positions (i.e. +X and -i-Y, ÷Y and +Z, or +X, +Y and +Zpositions) were found to lower La3-’- affinity, and it was rationalized that dentate-dentate repulsion occurred between these aspartate residues. This result isalso supported by the “acid pair hypothesis” (Reid et a!., 1980; Reid, 1987),which states that cation affinity of a helix-loop-helix will increase with the numberof acidic residues in chelating positions to a maximum of four. The analog inquestion in the above cited example had a total of five acidic residues inchelating positions (+Z, +Y, +Z, —X, and —Z) of the loop region.Interestingly, synthetic analogs having asparagine at the +X positioninstead of the invariant aspartic acid, also displayed significant affinity for La325INTRODUCTION(19 tM — 1.1 mM). These finding are inconsistent with the results of Babu et al.,(1992) in which the replacement of the invariant aspartic acid in the ÷X positionwith asparagine resulted in loss of calcium binding. This recurrent inconsistencybetween lanthanide and calcium binding characteristics emphasizes that thelanthanides may not faithfully mimic calcium binding properties.f. +Y, +Z, and —x PositionsUsing synthetic analogs of the third calcium binding site of calmodulin,Reid (1990) was able to make comparisons between: 1) peptides possessingdifferent numbers of acid residues in chelating positions and 2) peptides withacid residues located on the vertices of the X-axis, Z-axis, or both X and Z-axes.Results showed higher calcium affinity in a peptide having acidic residues on thevertices of the Z-axis compared to a peptide having the acidic residues on thevertices of the Xxis (KCa = 58.8 pM versus 524 pM respectively). It wasreasoned that the lower calcium affinity of the X-axis acid paired peptide wasdue to the indirect chelation of the aspartic acid in the —x position with thecation. The aspartic acid in the —X position chelates calcium indirectly via awater molecule (Strynadka and James, 1989), resulting in incompletedehydration of the cation that could conceivably decrease stabilization of thecalcium complex. The results also showed that increasing the number of acidresidues in the loop region to four and pairing them on the vertices of the X andZ-axes produced the peptide with significantly increased affinity for calcium (KCa= 19.1 jiM).26INTRODUCTION3. Nonchelating Residues of the Loop Regiona. Position 2Utilizing protein engineering, Babu et a!., (1992) replaced the glutamicacid residue found between the +X and +Y chelating position in site II of rabbitskeletal troponin C with an alanine residue. Determination of calcium bindingcapacity showed no significant difference compared to the wild type proteinpossibly indicating the flexible nature of the type of residue in this position.Using 34-residue analogs of rabbit skeletal troponin C site Ill (SCIII),Shaw et a!. (1991) replaced the chelating residues with those from site II to give(LIIL) (Figure 4). In a second analog (LII), the entire loop region of site Ill wasreplaced with that of site II. The result indicated that there was little change incalcium affinity between SCIII (native site) and LIIL (Ka = 3 tM versus 8 jiMrespectively), suggesting that the chelating residues had very little effect oncalcium affinity. LII, on the other hand, showed a 1000-fold decrease in calciumaffinity compared to SClll and LIIL. The authors attributed this large decrease incalcium affinity to the substitution of lysine by a glutamic acid residue betweenthe +X and +Y chelating positions of LII. Upon analysis of the primary structureof all known troponin C proteins (sites III and IV), this position is alwaysoccupied by a basic residue (Marsden eta!., 1990). It has been suggested thatthis charged side chain may stabilize the helix through a helix-dipole mechanism(Strynadka and James, 1989). Support for this proposal comes from crystalstudies of troponin C, which have shown the side chain of lysine in position 2 tobe directed back towards the N-terminal a-helix (Herzberg and James, 1988;Satyshur eta!., 1988).27INTRODUCTIONb. Position 6Using site-directed mutagenesis, Matsuura, et a!. (1991) replaced thehistidine residue at position 6 in the second calcium binding site of yeastcalmodulin (Saccharomyces cerevisiae) with a glycine residue in order toreplicate a typical EF hand containing an invariant glycine residue at thisposition. The investigators found that calcium bound to both sites with highaffinity. However, when an alanine residue replaced the invariant glycine inposition 6 of a 23 residue peptide analog of skeletal troponin C site U, there wasa four-fold decrease in calcium affinity compared to the unsubstituted peptide(Malik etaL, 1987).c. Positions 2, 4, and 10The CD calcium binding site of oncomodulin was mutated such that thecoordinating residues were identical to the high affinity CD calcium-magnesiumbinding site of parvalbumin (Hapak et a!., 1989). The resulting oncomodulinanalog failed to show calcium affinity comparable to that of parvalbumin (1 jiMversus 0.37 - 8 nM respectively) thereby suggesting that residues other thanthose found chelating the cation play a pivotal role in cation affinity. On thispremise, single site substitutions at nonchelating sites 2, 4, and 10 of the CDcalcium binding site of oncomodulin were replaced with residues occupying thesame positions in the CD calcium-magnesium binding site of parvalbumin(Palmisano et a!., 1990). Whereas the substitution of asparagine with lysine inposition 2 resulted in a 1.6 decrease in calcium affinity (Ka = 0.78 jiM for wild-type versus 1.27 jtM for the mutant), the substitution of glycine with glutamate atposition 10 resulted in a 1.5-fold increase in calcium affinity (KCa = 0.53 jiM).Substituting glutamine with lysine at position 4 did not result in any significantchange in calcium affinity (KCa = 0.72 jiM). The greatest change in calcium28INTRODUCTIONaffinity was observed in an analog in which aspartate and glycine of positions 9and 10 respectively have been replaced by two glutamate residues (Palmisano,et a!., 1990). The affinity for both calcium and magnesium of this mutatedanalog was greater than the wild type (Ka 0.41 iiM and KMg 0.74 mMversus KCa = 0.78 iiM and KMg = 3.5 mM respectively). This modest change incalcium affinity of the mutated oncomodulin is still however two orders ofmagnitude lower than the CD site of parvalbumin suggesting that residues otherthan those occupying the loop region play a role in dictating ion affinity.29INTRODUCTIONV. MAGNESIUM-CALCIUM BINDING SITESEven though the molecular mechanism of cation binding to a hlh-motifremains obscure, discrimination between calcium and magnesium binding islikely governed in part by the size of the binding cavity and the coordinationgeometry (Williams, 1977; Martin, 1983). Two classes of calcium binding siteshave been designated based on their affinities for magnesium (Kretsinger, 1980;Seamon, and Kretsinger, 1983; Haiech et a!., 1979; Haiech et a!., 1981). Thenonspecific calcium-magnesium sites have Ka’S <1O M and a KMg’S 1O M,whereas the specific calcium specific sites have KCa’S between 10—6 - 1 0 Mand KMg’S >1 O M. Nonspecific cation binding sites include the two C-terminalcalcium binding sites of troponin C, the two functional calcium binding sites ofpan/albumin, and the EF site of oncomodulin.Attempts at rationalizing a structure/function relationship between theamino acid sequence of the twelve residue loop region of the hlh-calciumbinding motif and the selectivity of the low affinity sites for calcium or the lack ofselectivity by the high affinity calcium sites suggested that the +Y chelatingresidues and the following residue in the loop sequence should be Asp-Gly for alow affinity calcium specific site and Asn-X (where X is not Gly) for a high affinitycalcium-magnesium site (Collins, 1976; Potter et a!., 1977). A more stringentstructure/function relationship was suggested by Kobayashi et al. (1989) whodistinguished between calcium specific and nonspecific sites by the followingthree criteria:1. Basic residues occur between the +X and +Y chelating positions innonspecific binding sites.2. Calcium specific sites have Asp at the +Y position, while nonspecificsites have Asn in the same position.30iNTRODUCTION3. Calcium specific sites have the sequence Gly-Gly/Ser-Gly following the+Y position, while nonspecific sites have the sequence X-Asp-Gly (where X isnot Gly) following the +Y position.It has also been speculated that the —X chelating position determines, inpart, ion-binding specificity (Williams et a!., 1987). The calcium-magnesiumbinding site from various parvalbumins contain a glutamic acid residue in the —Xposition compared to the calcium specific sites of calmodulin, calbindin, troponinC (N—terminal domain) and oncomodulin (CD site) which have either anaspartate or serine residue. Occupation of aspartate and serine in this positionbinds the cation indirectly via a water molecule, whereas the longer side chain ofglutamic acid chelates directly with the cation. This direct chelation involvingglutamate should increase the state of dehydration of the bound ion and therebyincrease the stabilization of the metal-complex. Thus, magnesium sensitivityresulting from the presence of glutamate in the —X chelating position may be theresult of 1) increased stabilization of the metal-complex or 2) the longer sidechain of the glutamate residue creating in a smaller cavity size conducive tomagnesium binding. Support for this glutamate-dependent magnesiumselectivity is found in studies in which the aspartate in the —X position ofoncomodulin was replaced by glutamate (Golden eta!., 1989; MacManus eta!.,1989). The glutamate containing analogs resulted in the elimination of thecalcium selective nature of the site and increased its sensitivity for Lu3 andMg2-’- (Golden et a!., 1989; MacManus et a!., 1989). As further support, spectralstudies employing UV absorbency, fluorescence, NMR techniques, and Eu3luminescence (MacManus et a!., 1989; Hapak et a!., 1989) have shown theglutamate substituted site of oncomodulin to behave in a similar fashion to theCD binding site of parvalbumin. However, it should be noted that there was onlya minor increase in magnesium affinity for the glutamate substituted site (3 mM31INTRODUCTIONfor the native protein and 1 mM for the glutamate substituted protein) (Hapak eta!., 1989). It should also be noted that the high affinity calcium-magnesiumbinding sites of troponin C have an aspartic acid residue rather than a glutamicacid residue in the —X chelating position.32INTRODUCTIONVI. OBJECTIVES OF STUDYThe objectives of this study have their roots going back to 1980 whenReid and Hodges suggested that the primary sequence of the loop region of anhlh-calcium binding motif encodes for 1) cation specificity and 2) the relativeaffinity for a specific cation. From these assumptions the acid pair hypothesiswas developed which states:“Affinity of the hlh-calcium binding loop unit for calcium willincrease with the number of anionic oxygen ligands found in chelatingpositions in the loop region to a maximum number of four and nonbondedinteractions between the anion ligands will be minimized while interactionof the chelating atoms with the cation will be maximized if the chelatingatoms are paired at the vertices of an octahedral coordination sphere”(Reid and Hodges, 1980,).Using the acid pair hypothesis as our working hypothesis, the objectivesof this study are three-fold:1) To explain calcium affinity in the hlh-calcium binding proteins basedon the nature of the amino acid residues in the six coordinating residueslocated within the 12 residue loop region.This is not to imply that the remaining nonchelating residues of the loopregion or the residues constituting the flanking a-helices do not contribute tocalcium affinity. It is however, an attempt to simplify the prediction of calciumaffinity in these calcium binding proteins by examining the contributions of the33INTRODUCTIONchelating residues in positions 1(+X), 3(-i-Y), 5(+Z), 7(—Y), 9(—X), and 12(—Z).Once the relationship between the chelating residues of the loop region andcalcium affinity has been established, the synthetic model can be used to furtherprobe the contributions of the nonchelating residues of the loop region as well asthe residues of the flanking cL-helices to calcium affinity.2) To explain cation selectivity in the hlh-calcium binding proteinsbased on the nature of the amino acid residues in the six coordinatingresidues located within the 12 residue loop region.Along with binding calcium, some hlh-motifs also bind magnesium. Suchis the case for the two C-terminal h/h-motifs of skeletal troponin C, the two sitesin parvalbumin, and the EF site of oncomodulin (Potter and Gergely, 1975;Lehky, et aL, 1977; MacManus et al., 1984; and Williams, et al., 1987). Thelower affinity constants seen for magnesium at these sites may be attributed inpart, to the different chemical properties of the cations. For example,magnesium has a small ionic radius compared to calcium (0.78 A vs. 1.06 A;Lehn, 1973), resulting in a higher charge density and a lower hydration number.Magnesium also has a free energy of hydration of —454 kcal/mol at 25 °Ccompared to —379 kcal/mol for calcium (Lehn, 1973), which translates into agreater amount of energy required to dehydrate the magnesium cation. Alongwith the different chemical natures of these cations, the lower observed affinitiesof hlh-motifs for magnesium may be due to dentate-dentate repulsion and stericinterference of the ligands approaching each other as a result of the smallerradius associated with magnesium (Reid and Hodges, 1980). As with explainingcalcium affinity, this study attempts to explain cation selectivity of hlh-calcium34INTRODUCTIONbinding proteins based on the contributions made by the chelating residues ofthe loop region.3) To resolve whether or not synthetic analogs of the third calciumbinding site of calmodulin exist as monomers or as homodimers.In calculating cation affinity constants of the hlh-motifs, our laboratory haspreviously assumed that the synthetic h!h-peptides have a single independentcation-binding site. However, using proteolytic and synthetic fragments ofskeletal troponin C, dimers of single helix-loop-helix motifs related by a 2-foldaxis of symmetry have been observed (Kay et aL, 1991; Shaw, 1990). Theoverall structure of the dimers were described as being similar to the C-terminaldomain of the crystal structure of skeletal troponin C. The significance of thesefindings makes it imperative to ascertain whether or not our model peptidesbehave similarly. If so, the assumption of a single independent binding site isnot valid, and, as a result, the cation binding data must be treated to reflect atwo-site model.35INTRODUCTIONVII. MODELAlthough natural fragments have been used extensively to examineprotein function, the method suffers from not being able to make modifications tothe primary structure. Site-directed mutagenesis overcomes this problem;however, the technique is limited to mutations involving only naturally occurringamino acids. Chemical synthesis of peptide fragments has also been employedpreviously to study protein function. Initially, this method was disappointing,producing synthetic fragments with very little activity (Maximov et al., 1978; Reid,et a!., 1980; Kanellis et a!., 1983; Pavone, et a!., 1984; Bonn, et a!., 1985;Buchta, et aL, 1986; and Foffani et al., 1991). The lack of activity resulted fromdifficulty in synthesizing fragments of sufficient size to be of biochemicalimportance. This limitation has been largely overcome by the introduction ofautomated synthesizers and more efficient coupling agents.One of the first fragments of a calcium binding site to be chemicallysynthesized was a 21 residue analog comprising the loop region and the C-terminal x-helix of troponin C site Ill (Reid et a!., 1980). Unfortunately, lack ofresponse to UV or CD spectroscopy in aqueous medium made it a poor modelfor structure-affinity analysis (Reid et a!., 1980). However, upon acetylation ofthis analog, calcium affinity could be measured (Ka = 4 mM) suggesting arepulsive interaction between the free amino group of the previousnonacetylated peptide with the cation. Considering this assumption, two moreanalogs (26 residues, and 34 residues) were prepared that yielded in calciumdissociation constants of 26 j.tM and 4 i.tM respectively (Reid et a!., 1981). Thefact that the KCa’S increased by orders of magnitude compared to the acetylated21-residue analog indicated that factors besides simple electrostatic repulsionbetween the free amino terminus and the cation were at work (perhaps36INTRODUCTIONthermodynamic stabilization of the calcium-peptide complex). Taken together,these results suggest that a suitable synthetic model should be of sufficientlength to include both of the flanking a-helices.The model used in this study is the third hlh-calcium binding motif ofcalmodulin that comprises residues 81-113. This site’s poor affinity for calcium(KCa = 735 LtM), provides us a starting point for which to design subsequentanalogs in an attempt to increase calcium affinity. This fragment, along with allof the analogs presented in this study (19 in total), were prepared by solid phasepeptide synthesis, which conveniently allows one to change any desired residuewithin the sequence. Due to the acid catalyzed oxidation of methionine residuesduring HF cleavage and deprotection, a leucine residue is substituted for themethionine residue at position 109.37MATERIALS AND METHODSMATERIALS AND METHODSMATERIALSThe chemicals and solvents used are reagent grade unless otherwisespecified.Aldrich Chemical Company (Milwaukee, Wisconsin, U.S.A.)d-Camphorsulfonic AcidGuanidine HydrochlorideMu rexideSodium Hydride (60% dispersion in mineral oil)Tetraethylammonium Chloride HydrateBache m (Torrance, California, U.S.A.)Benzhydrylamine Resin (BHA) (1% crosslink)N-Boc-L-AlanineN-Boc-NY-p-Tosyl-L-Argini neN-Boc-L-Aspartic Acid-3-CyclohexyI esterN-Boc-L-Glutamic Acid-7-Benzyl esterN-c-Boc-N-it-Benzyloxymethyl-L-HistidineN-Boc-L-lsoleucine hemihydrateN-Boc-L-LeucineN-Boc-N-e-(2-chloro-CBZ)-L-LysineN-Boc-L-PhenylalanineN-Boc-L-ProlineN-Boc-O-Benzyl-L-Threonine38MATERIALS AND METHODSN-Boc-O-(2,6-dichlorobenzyl)-L-Tyrosine,N-Boc-L-Valine,BDH Chemicals Canada Limited (Vancouver, British Columbia, Canada)Acetic Acid (glacial)Acetic AnhydrideCalcium ChlorideDichloromethane (DCM) (distilled over CaCO3prior to use)Diethyl ether (anhydrous)N,N-Dimethylformamide (DMF)Hydrochloric AcidMagnesium ChlorideNinhydrinPotassium ChloridePotassium HydroxidePropan-2-olPyridine (distilled over NaH and ninhydrin prior to use),Sulphuric AcidToluene (Sodium dried)Bi o- Rad Laboratories (Hercules, California, Ii. S.A.)AcrylamideAnalytical Grade Mixed Bed Resin AG-501-X8(D) 20-50MeshAmmonium PersulfateChelex (sodium form, 200-400 Mesh)N, N’-Methylene-bis-acrylamide39MATERIALS AND METHODSFisher Scientific Company (Vancouver, British Columbia, Canada)Acetonitrile, HPLC gradeAmmonium BicarbonateMethanol, HPLC gradeTriethylamine, HPLC gradeHalocarbon Product Corporation (River Edge, New Jersey, U.S.A.)Trifluoroacetic Acid (TFA)McArthur Chemical Company Ltd. (Montreal, Quebec, Canada)Calcium ChloridePeninsula Laboratories (Belmont, California, U.S.A.)N- Boc-G lycineRichelieu Biotechnologies inc. (Montreal, Quebec, Canada)[Benzotriazol-l-yl-oxy-tris-(di methylamino) phosphoniumhexafluorophosphate] (BOP)Sigma Chemical Company (St. Louis, Missouri, U.S.A.)AnisoleBrilliant Blue GN,N-Diiso-propylethylamine (DIEA) (distilled over NaH and thenover ninhydrin prior to use)Ethylenediamine-tetraacetic Acid (EDTA)Ethyleneglycol-bis-(3-Aminoethyl ether) N,N,N’,N’-tetraacetic Acid (EGTA)Na-t-Boc-N-a-Xanthyl-L-Glutamine40MATERIALS AND METHODSGlycerol1 -Hydroxybenzotriazole (HOBt)Indole,3-[N-Morpholino]propanesulfonic acid (MOPS)Nitrilotriacetic AcidN-t-Boc-O-Benzyl-L-SerineN, N, N’, N’-Tetramethylethylenediamine (Temed)3’,3”,5’,5”-Tetrabromophenolsulfonephthalei n(Bromophenol Blue)N-tris[hydroxymethyl] Methyl glyci ne (Tricine)Trifluoroethanol[Tris (hydroxymethyl) aminomethane Hydrochloride](Trizma-Hydrochloride)Vega (Tucson, Arizona, U.S.A.)N-Boc-His (Tos)41MATERIALS AND METHODSII. EQUIPMENTCircular Dichroism Spectrophotometer (Jasco)- Model 500A and J720Concentrator (Savant)- Automatic Concentrator model AS290Gel Electrophoresis (Bio-Rad)- Mini-Protean II Electrophoresis CellHF Apparatus (Protein Research Foundation)- Type 1 HF ApparatusHPLC (Waters)- Model 510 HPLC Pump- Model 730 Data Module- Model 721 Programmable System Controller- Lamba-max Model 481 LC SpectrophotometerHPLC Columns(Chromatographic Specialties)- Zorbax C8 Column (9.4 mm X 25 cm)- Zorbax C18 Column (9.4 mm X 25 cm)(Beckman)-Ultrasphere C8 Column (10 mm X 25cm)-Ultrasphere Cl 8 column (10 mm X 25 cm)42MATERIALS AND METHODSPeptide Synthesizer (Vega Biotechnologies)-Vega coupler 1000 peptide synthesizerpH Meter (Fisher)-Accumet pH meter model 825 MPAmino Acid Analyzer-LKB Model 4151 Alpha Plus43MATERIALS AND METHODSIII. METHODS1. Solid Phase Peptide Synthesis (SPPS)The basic principles and applications of SPPS have been welldocumented in several reviews (Birr, 1978; Barany and Merrifield, 1980;Bodanszky, 1984; Stewart and Young, 1984; Atherton and Sheppard, 1989).The fundamental concept of SPPS involves the synthesis of peptides throughthe sequential addition of cr-amino protected amino acid residues, starting at thecarboxyl terminus (Figure 5). The first amino acid is attached to an insolublesupport medium (resin) which provides a scaffold from which the peptide issynthesized through repeated deprotection, neutralization, and coupling steps.Upon completion of the synthesis, the peptide is cleaved and extracted from theresin.The resin used in the syntheses of this study is a crosslinked polystyrene(1% divinylbenzene), 200-400 mesh, functionalized with benzhydrylamine(BHA). The resin swells extensively in standard SPPS solvents(dichioromethane (DCM), and N,N-dimethylformamide (DMF)), allowingpenetration of solvents and reagents into the resin to increase reactionefficiency. The advantage of the BHA resin is its resistance to repeatedtreatment with trifluoroacetic acid (TEA) during the deprotection step. Thedisadvantage of using the BHA resin is that HF cleavage of the peptide/resinproduces a peptide with a C-terminal amide group rather than a carboxyl group.In SPPS it is necessary to protect reactive chemical groups other thanthose participating in the formation of the peptide bond. Protecting groups mustbe readily removed without subsequent damage to the growing peptide chain.44MATERIALS AND METHODSBoc—NH——CH——C——OH + POLYMER 1Attach first amino4, acid to resinBoc—NH--—-CH——C—O----j_POLYMERI Deprotect with 50% TFA4, Neutralized with 5% DIEANH2—CH—C—O-—-(_PLYMERR” 0I II Coupling of anBoc— NH—CH—C —x activated amino acidr________Boc—NH-——CH—C— NH—CH—C—O----[POLYMER II Deprotection andI coupling repeatedr1 0T II____Boc—PEPTIDE——C—O----j POLYMERHF CleavagePEPTIDE ÷ POLYMERFigure 5. Scheme of Solid Phase Peptide Synthesis.(Adapted from Stewart and Young, 1984)45MATERIALS AND METHODSTwo types of protecting groups are required in SPPS: 1) a temporary protectinggroup which protects the a-amino group and is removed during the deprotectionstep, and 2) semi-permanent protecting groups which protect the side chainsand are removed by hydrofluoric acid (HF) once the synthesis is complete. Inthis study, the a-amino group is protected by the t-butyloxycarbonyl (Boc) group.Removal of the Boc protecting group is required in order to couple the nextresidue. Deprotection of the Boc-amino acid is accomplished by acidolysisusing 50% TFNDCM (v/v)(with 0.1% indole). Since reagent grade TFA maycontain impurities such as aldehydes which could be deleterious to peptidesynthesis, 0.1% indole is added to act as a scavenger. Deprotection using 50%TEA is specific for the N-terminal Boc protecting group and will not affect thesemi-permanent protecting groups on the side chains. After removal of the Bocgroup, the TFA salt of the a-amino group is converted to the free base using 5%N, N-diisopropylethylamine/dichloromethane (Dl EA/DCM) (v/v). The stericallyhindered nature of DIEA curtails the formation of quaternary compounds and isfavored over other tertiary amines. DCM is the solvent of choice forneutralization, since DMF promotes nucleophilic reactions such astrifuloroacetylatio n, diketopiperazine formation, and N-terminal glutaminecyclization. Since the latter two mentioned reactions are catalyzed by weakacids, it is imperative that the peptide-resin be thoroughly washed afterdeprotection and that neutralization be carried out to completion before initiationof coupling.Formation of the peptide bond requires activation of the carboxyl group ofthe incoming a-amino protected amino acid (Boc-amino acid). The procedureused is a modification of the BOP/HOBt method described by Hudson, 1988. Inthis method, the [benzotriazol-l-yl-oxy-tris-(dimethylamino) phosphoniumhexafluorophosphate] (BOP) acts to produce an acyloxyphosphonium salt with46MATERIALS AND METHODSthe activated amino acid. This salt, in the presence of 1-hydroxybenzotriazole(HOBt), forms a benzotriazolyl active ester which favors coupling to the freeamino group of the resin linked peptide chain. In this procedure, DIEA is used toinitiate the activation reaction by producing the carboxylate anion of the Bocamino acid.After synthesis is complete, the final N-terminal residue is deprotected,and the exposed amino group is acetylated, resulting in a peptide with an N-terminal aminoacyl functional group. This is done to better simulate a fragmentexcised from the native protein which would have an N-terminal amide grouprather than a terminally charged amino group. Removal of the N-terminal Bocprotecting group reduces possible t-butylation of susceptible side chain residuesduring HF cleavage.Advantages of SPPS include: 1) All of the reactions are carried out in thesame reaction vessel eliminating loss of material due to transfer; 2) Productpurification after every addition of a protected amino acid residue is eliminatedand substituted with simple washings and filtrations of soluble contaminants; 3)Excess reactants can be used to force the reaction to completion; and 4)Procedures can be automated. A disadvantage of SPPS is that it requiresisolation and purification of the peptide. Incomplete coupling steps during thesynthesis will produce peptide chains which may differ from the desired peptideby only a few amino acid residues. However, HPLC has alleviated many of theproblems encountered in the purification process. The sequence of the peptidessynthesized for this study are listed in appendix A.a. Attachment of the First Amino Acid to the ResinBHA resin (approximately 3.0 g) is placed into a 50 mL reaction vessel,saturated with DCM, and allowed to swell overnight. After the resin has swelled,47MATERIALS AND METHODSthe DCM is drained from the vessel and replaced with 5% DIEAIDCM (v/v) toneutralize the resin. The resin is treated with two aliquots (approximately 40 mLeach) of the 5% DIEA/DCM (v/v) for 10 minutes. After the base treatment, theresin is washed six times with DCM and selectively coupled with Boc-Gly. BocGly (157.95 mg; 0.3 mmols/g resin) is placed into a beaker with HOBt (137.7 mg;0.3 mmols/g resin) and dissolved in DMF (15 mL). In a separate beaker, BOP(398.1 mg; 0.3 mmols/g resin) is dissolved in DMF (15 mL). DIEA (282 pi; 0.54mmols/g resin) is added to the beaker containing the BOP reagent. The reactionvessel is drained, and the contents of the two beakers are added manually to thereaction vessel. The coupling reaction proceeds for 30 minutes, after whichtimes the resin is washed with DCM (six times), with 80% isopropanol/DCM (v/v)(three times), and finally with DCM (six times). The peptide-resin is then treatedwith 5% DIEA/DCM followed by a final six washes with DCM.b. Acetylation of the Peptide ResinAfter the selective coupling of the Boc-Gly to the BHA resin, the remainingfree amino groups on the BHA are acetylated to prevent coupling of subsequentamino acid residues to the BHA resin. The acetylation solution consists oftoluene (sodium dried), pyridine (distilled from ninhydrin), and acetic anhydride(3:3:1), and is manually added to the reaction vessel in two aliquots. The firstaliquot is mixed with the Boc-Gly-resin for 5 minutes (prewash) and thendrained, removing with it any residual DCM from the previous coupling step.The second aliquot of the acetylation solution is then added to the reactionvessel and allowed to mix with the Boc-Gly-resin for 60 minutes. Afteracetylation is complete, the resin is washed six times with DCM. Solublecontaminants are removed by washing the resin three times with 80%isopropanol/DCM (v/v). Once the isopropanol washes are complete, the Boc48MATERIALS AND METHODSGly-resin is washed six times with DCM, once with 5% DIEA/DCM, and a final sixtimes with DCM. The acetylation procedure is also employed 1) to terminate agrowing peptide chain in the event of an inefficient coupling and 2) to modify theN-terminal amino acid residue after deprotection.c. DeprotectionThe Boc group is removed prior to coupling of the subsequent amino acidresidue of the growing peptide chain. The deprotection solution is a mixture of50% TFA/DCM (vlv) containing 0.1% of indole. The peptide-resin is treatedthree times with the TEA solution for 1, 5, and 20 minutes. After deprotection,the peptide-resin is washed six times with DCM to remove any residual TEA.The peptide-resin is then neutralized by treating the peptide-resin twice with 5%DIEAIDCM (v/v). Lastly, the peptide-resin is washed six times with DCM.d. Active Ester CouplingAfter selective coupling of the Boc-Gly to the BHA resin, the remainingcoupling reactions use a three-fold excess of Boc-amino acid. In a clean drybeaker, the appropriate Boc-amino acid (0.9 mmols/g resin) is added to HOBt(413.1 mg; 0.9 mmols/g resin). The Boc-amino acid and HOBt are dissolved inDMF (15 mL). In a second beaker, BOP (1.194 g;0.9 mmols/g resin) is dissolvedin DME (15 mL). DIEA (830 j.tL; 1.59 mmol/g resin) is added to the beakercontaining the BOP reagent, and the contents of the two beakers areimmediately added to the reaction vessel. Erom this point forward, the method isidentical to the coupling procedure for the first amino acid. To prevent sidereactions involving the side chains of the amino acids, semi-permanentprotecting groups are necessary (Table 1).49MATERIALS AND METHODSTable 1Semi-Permanent Protecting Groups for Boc-Amino AcidsBoc-Amino Acid Protecting GroupAlanine NoneArginine it-Toluenesulfonyl (Tosyl)Asparagine XanthylAspartate-Cyclohexyl esterGlutamate y-Benzyl esterGlycine NoneHistidine Benzyloxymethyl (Born) or2,4-dinitrophenyl (Dnp)Isoleucine NoneLeucine NoneLysine 2-Chloro-BenzyloxycarbonylPhenylalanine NoneSerine O-BenzylThreonine O-BenzylTyrosine 2,6-Dichloro-BenzylValine Nonee. Monitoring of SPPS (Ninhydrin Reaction)It is important in SPPS to have a rapid means of assessing the efficiencyof the coupling procedure. This is achieved by using a modified quantitativeninhydrin reaction which measures free amino groups (Sarin et aL, 1981). Thereaction produces a chromophore (Ruhemann’s purple) in solution which ismonitored at 570 nm.50MATERIALS AND METHODSSolutions needed for the ninhydrin test are as follows:Solution 1: 20 g Phenol5 mL Ethanol (100%)1 mL1OmMKCN50 mL PyridineSolution 2: 5% Ninhydrin/Ethanol (w/v),Solution 3: 0.5 M Tetraethylammonium Chloride/DCM(w/v),Between 8 and 13 mg of dried peptide-resin is placed into a small testtube. Eight drops of solution 1 and four drops of solution 2 are added to the testtube containing the dried resin. The test tubes are heated at 100 °C for 10minutes. After the reaction is complete, the contents are filtered through glasswool in a pasteur pipette. Approximately 100 .tL of solution 3 is passed throughthe pipette to displace any color bound to the resin through ionic interactions.The filtrate volume is adjusted to 2.0 mL with 60% (vlv) ethanol, and the A570 isread against a blank (reagents in the absence of resin). Free amine groups arecalculated using the following equation:Free NH2 = (Abs570)(Vol)/(Wt. Resin )(E’570where e’570 = 1.5 X 1 O M1 cm—1. This method is also used to determine thequantity of free amine groups after deprotection, giving an indication of the yieldof the growing peptide chain. In this case, the final volume used is 25 mL, andthe calculations are modified accordingly.f. HF CleavagePrior to cleaving the peptide from the resin, the peptide-resin isthoroughly washed with DCM to remove any residual DMF. One gram of dried51MATERIALS AND METHODSpeptide-resin, along with 1.0 mL of anisole and a stir bar, is placed into a teflonreaction vessel of a type 1 HF apparatus. The anisole is used as a scavenger toprevent alkylation of the peptide by t-butyl and benzyl cations. The reactionvessel is cooled in a liquid nitrogen bath for at least five minutes before distillingover approximately 10 mL of HF. The contents are stirred for 45-60 minuteswhile cooling in an ice bath (—5 to 0 °C), at which time the HF is removed undervacuum.g. Extraction of the peptideThe dried residue is washed with diethyl ether (3 X 10 mL) to removeorganic soluble impurities, and the peptide is extracted with 50% acetic acid (5 X10 mL). The resulting aqueous solution is reduced to dryness on a SavantAS290 SpeedVac.2. Purification of Peptidesa. DialysisThe peptide is dissolved in 50 mM ammonium bicarbonate andexhaustively dialyzed (SpectralPor Membrane MWCO:1 000, 38 mm flat width)against the same buffer. The dialyzed peptide solution is filtered using a 0.45i.im disposable filter, placed into a plastic container, and evaporated to drynessin the SpeedVac.b. HPLC Purification of PeptidesThe buffers used in the purification procedure are as follows:52MATERIALS AND METHODSA: 0.2% Triethylamine Phosphate, 2 mM EGTA, pH 6.5Triethylamine 11.2 mLo-Phosphoric Acid 8.0 mLEGTA 3.04 gNano Pure water made up to 4000 mLAdjust pH to 6.5 with TriethylamineB: 0.2% Triethylamine Phosphate, 2 mM EGTAIAcetonitrile (40:60), pH 6.5.Triethylamine 11.2 mL0-Phosphoric Acid 8.0 mLEGTA 3.04gAcetonitrile 2400 mLNano Pure water made up to 4000 mLAdjust pH to 6.5 with TriethylamineIn preparing buffers A and B, the EGTA is added to approximately 1000 mL ofwater and triethylamine is added to dissolve the EGTA. The remainingcomponents of the mixture are then added once the EGTA is completelydissolved. The pH is adjusted to 6.5 with 3 M KOH.Prior to purification, the peptide is dissolved in a small volume (10- 15mL) of buffer A. In the event of poor solubility, a few drops of triethylamine areadded to the peptide solution. To prevent racemization, the final pH of thepeptide solution is not to exceed 8.0.HPLC purification is performed on a C8 or C18 column using a 30 minutelinear gradient going from a 70:30 mixture of buffer A:B to a 20:80 mixture. Themajor peak on the elution profile is collected, and evaporated to dryness in theSavant AS1 60 SpeedVac concentrator, redissolved in buffer A, and purified for asecond time as described above.After the second purification step, the main peak is once again pooledand evaporated to dryness using the Savant AS16O SpeedVac concentrator.The peptide is redissolved in 50 mM NH4CO3and exhaustively dialyzed against53MATERIALS AND METHODSthe same buffer to remove the EGTA and any impurities. After dialysis, thepeptide solution is evaporated to dryness and stored at —20 °C.3. Determination of Peptide ConcentrationA 50 jiL sample of peptide is hydrolyzed using 6 N HCI in an evacuatedchamber at 110 °C for 24 hours. After hydrolysis, the peptide samples areanalyzed on an LKB Model 4151 alpha-plus amino acid analyzer using post-column derivatization with ninhydrin. The concentration of peptide is calculatedusing the mean of the molar ratios of the amino acids in the acid hydrolysate.Alternatively, the peptide concentration can be estimated by UV absorbance,employing a molar extinction coefficient (e275 = 1.63 X io M-1 cml) which waspreviously determined using quantitative amino acid analysis.4. Circular DichroismIn order to prepare a peptide sample for CD spectroscopy, theconcentration is estimated by UV absorbance as mentioned above on a HewlettPackard 8452A Diode Array spectrophotometer at room temperature over awavelength range from 250 to 350 nm. Spectra are recorded using quartz cellshaving a path length of 1 cm. The peptide is dissolved in the CD buffercomposed of 100 mM MOPS, 150 mM KCI, and 1 mM NTA, pH 7.2 to afford afinal concentration between 0.5 and 1.0 mg/mL.Prior to CD spectroscopy, the peptide solution is placed into aSpectra/Por Membrane (MWCO:1000, 12 mm flat width) and dialyzed overnightat 4 °C against 500 mL of the CD buffer mentioned above. Following dialysis,54MATERIALS AND METHODSthe peptide is filtered through a 0.45 .tm disposable filter and scanned a secondtime between 250 and 350 nm.All solutions for CD spectroscopy are prepared from deionized water, andNalgene® laboratory equipment is used in place of glassware to avoidcontamination of solutions with calcium leaching from glass. The followingbuffer is used to dissolve the peptide sample prior to CD spectroscopy.100 mM MOPS, 150 mM KCI, and 1 mM NTA at pH 7.2-MOPS 41.86 g-KCI 22.37 g-NTA 568.8 mg-Chelexed-treated water made up to 2000 mL-Adjust pH to 7.2 with 3 M KOHThe calcium and magnesium solutions are prepared from deionizedwater. The buffer used in preparing the two solutions is identical to the CD bufferexcept that the NTA is omitted.The concentration of the calcium solutions used in the titrations aredetermined by titration with EGTA as the primary standard, using murexide asthe end-point indicator; and the magnesium solutions are titrated with EDTA asthe primary standard using methyl thymol blue as indicator (West, 1969).Circular dichroism spectroscopy is performed on either a Jasco J500A ora J720 CD spectrophotometer. Circular dichroism is a technique capable ofexamining the conformation of macromolecules in solution. Circular dichroismmeasures the differential absorption of right- and left-circularly polarized light byan optically active chromophore (Freifelder, 1976). Use of circular dichroism inproteins can aid in determining secondary structures such as cc-helices, Isheets, and random coil. The model peptides of this study are monitored at 222nm to ascertain any change in a-helical content.55MA TERIALS AND METHODSThe peptide (900 1iL) is titrated with calcium, and the change in ellipticityat 222 nm is recorded in millidegrees. When there is no further change in thepeptide’s response to calcium, a second scan between 200 and 250 nm isrecorded. The titration is performed in triplicate. A CD scan of the peptidesaturated with TFE is also recorded between 200 and 250 nm.The mean residue ellipticity ([]2), at 222 nm is calculated from thefollowing equation:[]222 = (moeg) 222(MRW)/1 0 (I)(c)where MRW is the mean residue weight of the peptide (i.e. molecularweight/number of amino acid residues), mDeg is the observed ellipticity value at222 nm, I represents the path length (0.1 cm), and c is the peptide concentration(mglmL). The units of [9] are degreecm2mole1.These calculated values areplotted against wavelength (nm) and are indicated as dots on the CD spectra.5. Calculationsa. Free Calcium and Magnesium ConcentrationsFree calcium and magnesium concentrations are determined using theEQCAL computer program from Biosoft, with log10 of the association constant(Sillen, and Martell, 1971) for the complexing species of nitrilotriacetic acid(NTA) set at: H to NTA3,9.73; H to HNTA2,2.49; H toH2NTA1,1.89; Ca2to NTA3-, 6.5; and Mg2 to NTA3-, 5.5. The H to peptide log10 associationconstant is set at 4.00 to correspond roughly to the pKa of the acid side chainsinvolved in calcium chelation. Free cation concentrations are calculated, and56MATERIALS AND METHODSthese values are used to determine an association constant for calcium andmagnesium as described below.b. Dissociation ConstantsOur data is reported as dissociation constants (ie. the inverse of theassociation constant) of the peptides for calcium and magnesium and iscalculated using a modified Perrin and Sayce (1967), nonlinear regressioncomputer program which fits the CD cation titration data to an equation of theform:nKcat[Cat2iN1 + Kcat[Cat2])where f is the fraction of peptide molecules in the calcium-bound state and isdetermined as the ratio of the change in ellipticity at 222 nm to the maximumchange in ellipticity which can be elicited by calcium at 222 nm. n is the numberof binding sites, and Kc5t is the apparent association constant of the peptide forcalcium or magnesium. [Cat2] is the concentration of free calcium ormagnesium calculated as described above.The procedure to estimate the helical content of the synthetic peptidebased on []222 values is described elsewhere (Chen et a!., 1974). The fractionof ca-helix (rh) present is calculated as follows:= [e]2221[e],.(1-kin)where [0]H and k are calculated constants (Chen et a!., 1974) which are 39,500and 2.57, respectively, and n is the average helical length taken as nine in thesecases (Reid eta!., 1981, and Nagy etal., 1978).57MATERIALS AND METHODS6. Polyacrylamide Gel ElectrophoresisSeparation of peptides was achieved using a modified Schagger and VonJagow (1987) discontinuous polyacrylamide gel electrophoresis system. Theuse of tricine, as the trailing ion, allows an increase in the resolution of peptidesas compared to glycine. The compositions of the gels are listed in Table 2.Table 2Composition of PAGE GelsReagent Stacking Gel “Spacer “ Gel Separating Gel4%T,3%C 1O%T,3%C 16.5%T,3%C49.5 % T, 3% C 1.0 mL 6.1 mL 10 mLSolutionGel Buffer 3.1 mL 10 mL 10 mLGlycerol 3.2 mLDEGASWater 8.33 mL 13.6 mL 6.25 mL10%Ammonium 100iL lOOiiL 100PersuiphateTemed lOitL lOitL 10pLAfter electrophoresis is complete, the gels are stained using coomassieblue (100 mg/mL) for 1-2 hours. Destaining the gels is accomplished using 10%acetic acid for 2 hours, exchanging the solution every 30 minutes.58RESULTSRESULTSI. PEPTIDE PURIFICATIONA typical HPLC profile using CaM:4Z(DES) is shown in Figure 6. Insert Ashows the HPLC profile of the crude peptide. The major peak has a retentiontime of 23.71 minutes and is followed by several minor peaks which mayrepresent peptide byproducts resulting either from amino acid deletions duringthe synthesis or peptide fragments generated by exposure of the peptide tohydrogen fluoride.Insert B (Figure 6), represents the “clean-up” stage in the purificationprocess. At this point, most of the peaks resulting from the byproducts havebeen eliminated. A minor peak at approximately 9 minutes and a slight shoulderon the downward slope of the major peak remain to be separated. As a result ofdecreasing the flow rate from 2.5 (insert A) to 2.0 mL/min, there is an increase inretention time to 25.16 minutes. Insert C shows the profile of the final purifiedpeptide.59RESULTSI I I I I0 5 10 15 20 25 30 35 40TIME (mins)Figure 6. Reverse Phase HPLC Purification Profile of CaM:4Z(DES) Using a C18Column. Insert A (Crude Peptide): Volume injected: 20 ilL, wavelength: 210 nm, AUFsetting: 2.0, flow rate: 2.5 mLlmin, retention time: 23.71 mm. Insert B (Clean-Up Step):Volume injected: 1.5 mL, wavelength: 280 nm, AUF setting: 2.0, flow rate: 2.0 mL!min,retention time: 25.16 mm. Insert C (Isolated Peptide): Volume injected: 50 giL,wavelength: 280 nm, AUF setting: 0.5, flow rate: 2.0 mL/min, retention time: 25.29 mins.Buffers and gradient as per Material and Methods, section Ill.2.b.ApooledpooledBC60RESULTSII. CD ANAL YSESThe results of CD analyses are discussed according to groups of peptidesrelated to each other by the number and position of acidic residues in chelatingpositions. However, for easy reference during the discussion, the CD data iscompiled in Table 11 (page 106).1. CaM:3(DNS)Figure 7 illustrates the combined CD spectra of CaM:3X(DNS) in theabsence of calcium, presence of calcium, and presence of trifluoroethanol(TFE). In the absence of calcium, CaM:3(DNS) displays a [6] of 5,470 ± 145,as compared to a value of 13,755 ± 60 in the presence of calcium (Table 3).These values correspond to 6 and 16 residues, respectively, making up the chelical portions of the peptide. In the presence of the conformation-inducingsolvent TFE, the mean-residue ellipticity is 19,760 which corresponds to an -helical content of 23 residues (Table 3).Figure 8 is a plot of the ratio of calcium-induced change in ellipticity at222 nm versus the negative log of the calcium concentration (p[Ca]). The resultis a curve which is theoretically sigmoidal in nature. Using nonlinear regressionanalysis, the calcium dissociation constant is determined. This correlates with a50% change in ellipticity (f = 0.5). The calculated p[Ca]f05 of CaM:3(DNS) is3.13 which translates to a Kca of 735 tM (Figure 8, and Table 3).2. CaM:3X PeptidesFigure 9 depicts the CD spectra of CaM:3X(NSD), CaM:3X(NSE),CaM:3X(NND), and CaM:3X(NNE) in the absence of cation, in the presence ofeither calcium or magnesium, and in the presence of added TFE. In the apo61RESULTSCaM:3(DNS)420—2—4—6—8—10—12—14—16—18—20—22—24—26—28—30210 220 230 240 250WAVLENGTH (nm)G)0EC.”EC)0)a)-oxFigure 7. CD Spectra Showing Mean Residue Ellipticity of CaM:3(DNS).1. Apopeptide in aqueous buffer.2. Calcium-saturated peptide in the aqueous buffer.3. Apopeptide in hydrophobic buffer.Aqueous buffer is 100 mM MOPS, 150 mM KCI, 1 mM NTA, pH 7.2.Hydrophobic buffer is a 1:3 (v/v) mixture of aqueous buffer in TFE.Cell path length is 0_i cm.62Table3CorrespondenceBetweentheNumber of ResiduesinCL-HelicalRegionsandMeanResidueEllipticityValuesof CaM:3(DNS)—[01222(Degcm2/dmol)(a)CalculatedResidues(C)—Ca2+Ca2+TFE(b)—Ca2+Ca2+TFEKa(.tM)CaM:3(DNS)5,470±14513,755±6019,76061623735±61aMeanresidueellipticityat222nmisexpressedinmillidegrees±S.D.(n=3).bThedataindicatedaretheresultofasingleexperiment.cProceduretoestimatethehelicalcontentdescribedbyChenet aL,1974.0RESULTSf1.21.11 .00.90.80.70.60.50.40.30.20.10.00.5 1.0 1.5 2.0 2.5p[Caj4.0 4.5Figure 8. CD-Monitored Calcium Titration of CaM:3(DNS).OCaM:3(DNS) (peptide concentration = 0.40 mg/mL)f is the ratio of the calcium-induced change in ellipticity at 222 nm to the maximumchange in ellipticity that can be elicited by calcium at 222 nm.Cell path length is 0.1 cm.I II I I - I i I I3.0 3.564C’LtiCaM:3X(NSD)CaM:3X(NSE)CaM:3X(NND)CaM:3X(NNE)4 2 0—2—4-—6--6 °—12-142—16—18-X-22-24—26—28—30210220230240250210220230240250210220230240250210220230240250WAVLENGTH(nm)Figure9.CDSpectraShowingMeanResidueEllipticityofCaM:3X(NSD),CaM:3X(NSE),CaM:3X(NND),andCaM:3X(NNE).1.ApopeptideInaqueousbuffer.2.Calcium-saturatedpeptideIntheaqueousbuffer.3.Apopeptideinhydrophobicbuffer.4.MagnesIum-saturatedpeptideIntheaqueousbuffer.AqueousbufferIs100mMMOPS,150mMKCI,1mMNTA,pH7.2.HydrophobicbufferIseithera1:1or1:3(v/v)mixtureof aqueousbufferInTFE.CellpathlengthIs0.1cm.RESULTSstate, both CaM:3X(NS1 E) and CaM:3X(NNE) possess a similar mean residueellipticity value ([O]222 = 3,636 ± 259 and 3,646 ± 158 respectively)corresponding to 4 residues in the x-helical regions of both peptides (Table 4).CaM:3X(NSD) and CaM:3X(NND) show slightly more structure in the absence ofcation, with mean residue ellipticity values of 4,136 ± 17 and 5,715 ± 247,respectively, corresponding to an cc-helical content of 5 and 7 residues,respectively.The addition of magnesium to the CaM:3X peptides does not result in anysignificant change in mean residue ellipticities when compared to the apopeptides (Figure 9, and Table 4). Addition of magnesium does not result in anincreased number of residues making up the cc-helical region of the peptides(Table 4).The addition of calcium results in the individual CaM:3X peptides showingvarying degrees of induced structure (Figure 9, Table 4). CaM:3X(NNE) andCaM:3X(NSE), which both have the lowest mean residue ellipticity values in theapo-state, also have the lowest mean residue ellipticity values in the calciumbound state (7,261 ± 80 and 9,188 ± 47 respectively, Table 4), whichcorresponds to 8 and 11 residues in the cc-helical regions, respebtively. On theother hand, CaM:3X(NSD) and CaM:3X(NND) show significantly higher meanresidue ellipticities of 13,084 ± 209 and 15,251 ± 197, correlating to an cc-helicalcontent of 15 and 18 residues, respectively.Maximal induction of helicity results upon the addition of TFE to theCaM:3X series of peptides. CaM:3X(NND) shows the greatest response to TFE,with a []222 of 27,599 corresponding to 32 residues in the cc-helical regions(Table 4). As previously mentioned, the CaM:3X(NND) peptide also showed thelargest calcium-induced change in ellipticity. This positive correlation betweencalcium and TFE-induced conformation does not always hold true. For example,66Table4CorrespondenceBetweentheNumber ofResiduesina-HelicalRegionsandMeanResidueEllipticityValuesoftheCaM:3XPeptides—[01222(Degcm2/dmol)(a)CalculatedResidues(C)3X-Peptides+Ca2+Mg2+(b)+TFE(b)—Ca2+Ca2+Mg2+TFEKKMQ(.tM)(KM)CaM:3X(NSD)4,136±1713,084±2094,28619,05551552215,400±600CaM:3X(NSE)3,636±2599,188±473,59018,96741142219,600±2400CaM:3X(NND)5,715±24715,251±1975,56627,599718732524±_16CaM:3X(NNE)3,646±1587,261±803,29216,657484193,140±_540aMeanresidueellipticityat222nmisexpressedinmillidegrees±S.D.(n=3).bThedataindicatedaretheresultofasingleexperiment.CProceduretoestimatethehelicalcontentdescribedbyChenetaL,1974.RESULTSCaM:3X(NSE), which has a significantly lower [O]2 than CaM:3X(NSD) in thepresence of calcium and magnesium, shows essentially the same mean residueellipticity value as CaM:3X(NSD) in the presence of TFE (18,967 versus 19,055respectively, Table 4). Both CaM:3X(NSE) and CaM:3X(NSD) have an ce-helicalcontent of 22 residues. CaM:3X(NNE) results in the lowest TFE-induced meanresidue ellipticity value of 16,657 correlating to 19 residues within the cL-helicalregions.The dissociation curves of figure 10 show that CaM:3X(NSD) andCaM:3X(NSE), with a serine residue in the +Z chelating position, havep[Ca2+]f05 values of 1.81 and 1.71, respectively, which correlate to dissociationconstants of 15.4 and 19.6 mM, respectively (Figure 10, Table 4). On the otherhand, CaM:3X(NND) and CaM:3X(NNE), with an asparagine residue in the +Zchelating position, yields in significantly higher p[Ca24]f05 values of 3.28 and2.5, respectively, equating to dissociation constants of 524 and 3,140 jiM,respectively. It is also evident that an aspartic acid residue in the —X chelatingposition results in a peptide having higher affinity for calcium. For example,CaM:3X(NSD) with an aspartic acid residue in the —X position, shows a 1.3-foldgreater affinity for calcium than CaM:3X(NSE) which has a glutamic acid residuein the —X position (KCa = 15.4 mM versus 19.6 mM respectively, Table 4).Similarly, CaM:3X(NND) has a 6-fold greater affinity for calcium thanCaM:3X(NNE).No correlation is found in-between the calcium affinities of the CaM:3Xpeptides and the degree of calcium-induced conformation. For example,CaM:3X(NNE), with only 8 residues in the cL-helical regions, has a five-foldgreater affinity for calcium than CaM:3X(NSD) which has 15 residues in the ahelical region (Table 4).68RESULTS1 .21.11.00.90.80.70.6f0.50.40.30.20.10.0—0.1—0.2p[CajFigure 10. CD-Monitored Calcium Titration of the CaM:3X PeptidesOCaM:3X(NSD) (peptide concentration = 0.61 mglmL)•CaM:3X(NSE) (peptide concentration = 0.51 mglmL)VCaM:3X(NND) (peptide concentration = 1.52 mg/mL)vCaM:3X(NNE) (peptide concentration = 0.59 mg/mL)f is the ratio of the calcium-induced change in ellipticity at 222 nm to the maximumchange in ellipticity that can be elicited by calcium at 222 nm.Cell path length is 0.1 cm.0 1 2 3 4 5 669RESULTS3. CaM:4X PeptidesFigure 11 shows the CD spectra of the CaM:4X peptides in the absenceof added cations, in the presence of either calcium or magnesium, and in thepresence of TEE. In the absence of added cations, CaM:4X(DNE) displays theleast amount of mean residue ellipticity (3,717 ± 184, Table 5) followed byCaM:4X(DSD) and CaM:4X(DSE) ([O]2 = 4,670 ± 101 and 4,212 ±136respectively). In terms of the number of residues involved in the a-helices of thepeptides, apo CaM:4X(DNE) has 4, whereas both CaM:4X(DSD) andCaM:4X(DSE) have 5. CaM:4X(DND) shows the highest ct-helical content in theapo-state ([]2 = 5,862 ± 76) accounting for 7 residues contributing to the ahelices.Along with having the greatest degree of conformation in the absence ofcation, CaM:4X(DND) also displays the greatest amount of ct-helical structure inthe presence of calcium ([e]2 = 22,263 ± 111, or 26 residues in the ct-helicalregions, Table 5). CaM:4X(DSD) has the second highest fraction of a-helicalstructure with a []2 of 16,621 ± 62 (19 residues forming the ct-helices).CaM:4X(DSE) and CaM:(DNE) show the lowest degree of calcium-inducedstructure, with 14 calculated residues in the a-helical region ([e]222 = 12,342 ±12 and 11,581 ± 67 respectively, Table 5). As with the CaM:3X peptides, thereis no significant change in mean residue ellipticity when the CaM:4X peptidesare titrated with magnesium.CaM:4X(DND), which shows the most ct-helical structure in both the apoand calcium saturated state, also shows the greatest amount of structure in theTEE saturated state ([8]2 = 23,555, or 28 residues making up the ct-helicalregions, Table 5). Similarly, CaM:4X(DNE), which displays the least amount ofstructure in the apo and calcium-saturated state, also displays amongst thelowest degree of structure in the TEE-saturated state ([O]2 = 17,724, or 21704 2 0—2—4—6—8—10—12—14—16—18—20—22—24—26—28—30210220230240250210220230240250210220230240250210220230240250WAVLENGTH(nm)Figure11.CDSpectraShowingMeanResidueEllipticityofCaM:4X(DSD), CaM:4X(DSE),CaM:4X(DND),andCaM:4X(DNE).1.Apopeptideinaqueousbuffer.2.Calcium-saturatedpeptldeintheaqueousbuffer.3.Apopeptldeinhydrophobicbuffer.4.Magnesium-saturatedpeptideintheaqueousbuffer.Aqueousbufferis100mMMOPS,150mMKCI,1mMNTA,pH7.2.Hydrophobicbufferiseithera1:1or1:3(vlv)mixtureofaqueousbufferinTFE.CellpathlengthIs0.1cm.CaM:4X(DSD)CaM:4X(DSE)CaM:4X(DND)CaM:4X(DNE)G) 0 E N C.) 0) xTable5CorrespondenceBetweentheNumber of Residuesina-HelicalRegionsandMeanResidueEllipticityValuesof theCaM:4XPeptides—t]222(Degcm2/dmol)(a)CalculatedResidues(C)4X-Peptldes—Ca2+Ca2+Mg2(b)+TFE(b)—Ca2+Ca2+Mg2+TFEKaKMG(KM)(riM)CaM:4X(DSD)4,670±10116,621±625,12217,978519621407±27CaM:4X(DSE)4,212±13612,342±2003,58620,2375143242,806±_185CaM:4X(DND)5,862±7622,263±1115,79623,55572672842.1±1.2CaM:4X(DNE)3,717±18411,581±673,16417,7244144211,950±_160aMeanresidueellipticityat222nmisexpressedinmillidegrees±S.D. (n=3).bThedataindicatedaretheresultofasingleexperiment.CProceduretoestimatethehelicalcontentdescribedbyChenet aL,1974.RESULTSresidues in the tx-helical regions). On the other hand, CaM:4X(DSE), which hasa significantly lower mean ellipticity value than CaM:4X(DSD) in the presence ofcalcium, has a considerably higher amount of structure in the presence of TEE([O]222 = 20,237 vs. 17,987 respectively).Unlike the CaM:3X peptides, the dissociation constants of the CaM:4Xpeptides obtained from the sigmoidal curves in figure 12 correlate with thedegree of calcium-induced structure. For example, CaM:4X(DND) which has thegreatest degree of structure in the presence of calcium, also shows the greatestaffinity for calcium (KCa = 42.1 j.tM, Table 5, Figure 12). Similarly, CaM:4X(DSE)and CaM:4X(DNE), which possess the least amount of calcium-inducedstructure, have the poorest affinities for calcium (Ka = 2,806 jiM and 1,950 j.tMrespectively). Lastly, CaM:4X(DSD), which shows intermediate calcium-inducedstructure, shows intermediate affinity for calcium (Ka 407 jiM).As with the CaM:3X peptides discussed previously, a serine residue in the+Z chelating position of the CaM:4X peptides results in a reduction in calciumaffinity. There is a 10-fold decrease in affinity when the asparagine residue inthe +Z position of CaM:4X(DND) is replaced by a serine residue to giveCaM:4X(DSD) (KCa = 42.1 vs. 407 jiM, respectively, Table 5). The samesubstitution in going from CaM:4X(DNE) to CaM:4X(DSE) also results in areduction in calcium affinity, albeit only 1.4-fold (KCa = 1,950 VS. 2,806 j.LM,respectively). Also in agreement with the CaM:3X peptides is the finding that anaspartic acid residue, rather than a glutamic acid residue, in the —X chelatingposition results in peptides with higher affinities for calcium. For example,CaM:4X(DSD) has a 7-fold greater affinity for calcium than CaM:4X(DSE), andCaM:4X(DND) had a 46-fold greater affinity than CaM:4X(DNE) (Table 5, Figure12).73RESULTSp[Ca]I I I I II IIIIIIIIIIIII1 .21.11 .00.90.80.70.60.50.40.30.20.10.0—0.1—0.20 6Figure 12. CD-Monitored Calcium Titration of the CaM:4X PeptidesOCaM:4X(DSD) (peptide concentration = 0.74 mglmL)•CaM:4X(DSE) (peptide concentration = 0.85 mg/mL)VCaM:4X(DND) (peptide concentration = 0.48 mg/mL)vCaM:4X(DNE) (peptide concentration = 0.94 mg/mL)f is the ratio of the calcium-induced change in ellipticity at 222 nm to the maximumchange in ellipticity that can be elicited by calcium at 222 nm.Cell path length is 0.1 cm.1 2 3 4 574RESULTS4. CaM:3Z PeptidesThe CD spectra of the CaM:3Z peptides (Figure 13) show the apoCaM:3Z(NES) as having a significantly greater mean residue ellipticity than apoCaM:3Z(NDN), (4,765 ± 104 versus 3,628 ± 267 respectively, Table 6).However, in the presence of calcium, only a minimal change in structure isobserved for CaM:3Z(NES), whereas CaM:3Z(NDN) shows a large structuralchange ([]2 = 5,610 ± 48 versus 14,434 ± 115 respectively). In terms ofcalculated residues in the a-helical regions, CaM:3Z(NES) goes from 6 residuesin the absence of calcium to 7 residues in the presence of calcium, whereasCaM:3Z(NDN) goes from 4 to 17 residues in the absence and presence ofcalcium, respectively. CaM:3Z(NDS) has greater structure in the absence ofcalcium ([O]2 = 7,095 ± 291, or 8 residues in the a-helical regions) than bothCaM:3Z(NES) and CaM:3Z(NDN). Likewise, CaM:3Z(NDS) exhibits a greaterdegree of structure in the calcium saturated state ([O] = 19,540 ±423, or 23residues in the a-helical regions) as compared to CaM:3Z(NES) andCaM:3Z(NDN).Unlike any of the peptides discussed to this point, CaM:3Z(NDS) displaysa magnesium-induced structural change. The mean residue ellipticity was 9,452in the magnesium saturated state compared to 7,095 ± 291 in the apo state(Table 6). The addition of magnesium results in an increase of 3 extra residuesin the a-helical portions of the peptide as compared to the apo peptide.CaM:3Z(NDN) and CaM:3Z(NES) both failed to show any significant change instructure resulting from the introduction of magnesium into the system.The structure inducing solvent TEE, produced the greatest response inCaM:3Z(NDS) ([8]2 = 23,450, Table 6). CaM:3Z(NDN) and CaM:3Z(NES)754 2 0—2—4—6—8—10—12—14—16—18—20—22—24—26—28—3D210220230240250210220230240250210220230240250WAVLENGTH(nm)CaM:3Z(NES)G) 0 2 0 2 C))0 IFigure13.CDSpectraShowingMeanResidueEllipticityofCaM:3Z(NDN), CaM:3Z(NDS),andCaM:3Z(NES).1.ApopeptldeInaqueousbuffer.2.Calcium-saturatedpeptideintheaqueousbuffer.3.Apopeptideinhydrophobicbuffer.4.MagnesIum-saturatedpeptideIntheaqueousbuffer.Aqueousbufferis100mMMOPS,150mMKCI,1mMNTA,pH7.2.Hydrophobicbuffer iseithera1:1or1:3(v/v)mixtureofaqueousbuffer inTFE.CellpathlengthIs0.1cm.C’Table6CorrespondenceBetweentheNumberofResiduesina-HelicalRegionsandMeanResidueEllipticityValuesoftheCaM:3ZPeptides—[01222(Degcm2/dmol)(a)CalculatedResldues(C)3Z-Peptides—Ca2+Ca2+Mg2+(b)+TFE(b)—Ca2+Ca2÷Mg2+TFEKaKMG(p.M)(mM)CaM:3Z(NDN)3,628±26714,434±1153,97218,5744175221000±140CaM:3Z(NDS)7,095±29119,540±4239,45223,450823112758.881±0.1±15CaM:3Z(NES)4,765±1045,610±484,79317,7916762183,750±30,000aMeanresidueellipticityat222nmisexpressedinmillidegrees±S.D.(n=3).bThedataindicatedaretheresultofasingleexperiment.CProceduretoestimatethehelicalcontentdescribedbyChenetaL,1974.RESULTSresponded to a lesser degree in the presence of TFE ([O]2 = 18,574 and17,791 respectively).The CD calcium titration curves of figure 14 show the great differencesobserved in the amount of calcium required to elicit a 50% change in ellipticity at222. The poor affinity of CaM:3Z(NES) for calcium (p[Ca]05 = 1.08, KCa83.75 mM, Figure 14, Table 6) is reflected in its dissociation curve lying close tothe Y-axis. Upon inspection of the titration curve for CaM:3Z(NES), it is clearthat the curve does not fit the typical sigmoidal shape and that the linear portionof the plot is not parallel to the other plots. CaM:3Z(NDS), which has thegreatest amount of ct-helical structure, compared to CaM:3Z(NDN) andCaM:3Z(NES), in any of the above mentioned conditions, also has the highestaffinity for calcium (p[Ca]05 = 4.23, KCa = 58.8 jiM, Figure 14, Table 6). TheCD curve of CaM:3Z(NDN) lies in-between CaM:3Z(NES) and CaM:3Z(NDS).Even though CaM:3Z(NDN) has less structure in the absence of calcium thanCaM:3Z(NES), it has the capacity for binding calcium (p[Ca]f05 = 3.0, Kca =1,000 jiM, Figure 14, Table 6).As mentioned above, CaM:3Z(NDS) displays a magnesium-inducedstructural change. Figure 15 represents the CD-monitored magnesium titrationcurve of CaM:3Z(NDS). Even though there is a definite response to magnesium,the curve is not sigmoidal. The calculated KMg for CaM:3Z(NDS) is 81 mM(Table 6), which corresponds to a p[Mg]f5 of 1.09 (Figure 15).Similar to the CaM:4X peptides, calcium affinity correlates with the degreeof calcium-induced conformation of the CaM:3Z peptides. For example,CaM:3Z(NDS), which has the greatest affinity for calcium, also has the greateststructural change in the presence of calcium (15 residue increase in its ct-helicalregions, Table 6). As well, CaM:3Z(NES), which showed very poor affinity forcalcium, has an increase of only one residue in the a-helical region of the78RESULTS1.21.11.0 -0.9 -0.8 -0.7 -0.6 -f 0.5 -I I0.4 - •1I I0.3-23 4 5 6p[Ca]Figure 14. CD-Monitored Calcium Titration of the CaM:3Z PeptidesOCaM:3Z(NDN) (peptide concentration = 0.57 mg/mL)•CaM:3Z(NDS) (peptide concentration = 0.54 mg/mL)VCaM:3Z(NES) (peptide concentration = 0.66 mg!mL)f is the ratio of the calcium-induced change in ellipticity at 222 nm to the maximumchange in ellipticity that can be elicited by calcium at 222 nm.Cell path length is 0.1 cm.79RESULTS1.21.11.00.90.80.70.6f 0.50.40.30.20.10.0—0.1—0.20.0Figure 15. CD-Monitored Magnesium Titration of CäM:3Z(NDS).OCaM:3Z(NDS) (peptide concentration = 0.39 mg/mL)f is the ratio of the calcium-induced change in ellipticity at 222 nm to the maximumchange in ellipticity that can be elicited by calcium at 222 nm.Cell path length is 0.1 cm.I I0.5 1.0 1.5 2.0 2.5 3.0p[Ca]80RESULTSpeptide in the presence of calcium. The 17-fold greater affinity for calcium ofCaM:3Z(NDS) compared to CaM:3Z(NDN) suggests that the presence of anasparagine residue in the —X chelating position may be detrimental to calciumaffinity. Similarly, the loss of calcium affinity in going from CaM:3Z(NDS) toCaM:3Z(NES) (KCa = 1,000 vs. 83.75 mM) suggests that helix-loop-helixcalcium binding motifs cannot accommodate a glutamic acid residue in the +Zchelating position.5. CaM:4Z PeptidesIn the apo-state, CaM:4Z(DDN) shows the least amount of structure whencompared to CaM:4Z(DDS) and CaM:4Z(DES) ([O] = 3,077 ± 88 compared to5,090 ± 229 and 5,507 ± 192 respectively; Figure 16, and Table 7). The lowdegree of structure observed for apo-CaM:4Z(DDN) (4 residues in the x-helicalregions) is also observed in the apo-CaM:3Z(NDN) (discussed in section 11.4).Both peptides have an asparagine residue in the —X chelating position.Upon saturation of the peptides with calcium, the situation changes suchthat CaM:4Z(DDN) has a greater degree of structure compared to CaM:4Z(DES)([]2 = 11,310 ± 111 vs. 8,138 ± 211, Table 7). CaM:4Z(DDS) displays themost calcium-induced structure with a []2 of 21,707 ± 140. Upon examinationof the CaM:4Z and CaM:3Z peptides, one recognizes a correlation betweencalcium affinity and the residue type occupying the +Z and -X chelatingpositions. That is, greatest calcium affinity is observed in the CaM:4Z andCaM:3Z peptides having an aspartate and serine residue in the +Z and —Xchelating positions, respectively, followed by aspartate and asparagine in the +Zand —X chelating positions, respectively, and lastly by glutamate and serine inthe +Z and —X chelating positions, respectively.81CaM:4Z(DDN)CaM:4Z(DDS)CaM:4Z(DES)4 2 0—2-—4—6-c’.J-o—12-1413)—16‘7—18C-X—22-24—26—28—30210220230240250210220230240250210220230240250WAVLENGTH(nm)FIgure16.CDSpectraShowingMeanResidueEllipticityofCaM:4Z(DDN),CaM:4Z(DDS),andCaM:4Z(DES).1.Apopeptideinaqueousbuffer.2.Calcium-saturatedpeptideintheaqueousbuffer.3.Apopeptideinhydrophobicbuffer.4.Magnesium-saturatedpeptideIntheaqueousbuffer.Aqueousbufferis100mMMOPS,150mMKCI,1mMNTA,pH7.2.HydrophobicbufferIseithera1:1or1:3(v/v)mixtureofaqueousbufferinTEE.Cellpathlengthis0.1cm.Table7CorrespondenceBetweentheNumber of Residuesino-HeIicalRegionsandMeanResidueEllipticityValuesof theCaM:4ZPeptides—[O]222(Degcm2/dmol)(a)CalculatedResidues(C)4Z-Peptides—Ca2+Ca2+Mg2(b)+TFE(b)—Ca2+Ca2+Mg2+TFEKcaKMG+(pM)(mM)CaM:4Z(DDN)3,077±8811,310±1114,86716,42141361954254±65±4CaM:4Z(DDS)5,090±22921,707±14016,38423,103625192729.217±1.0±3CaM:4Z(DES)5,507±1928,138±2115,46021,20461062531,940±5,900aMeanresidueellipticityat222nmisexpressedinmillidegrees±S.D. (n=3).bThedataindicatedaretheresultofasingleexperiment.CProceduretoestimatethehelicalcontentdescribedbyChenetaL,1974.RESULTSAs with CaM:3Z(NDS) (Figure 15), CaM:4Z(DDS) and CaM:4Z(DDN) alsodisplay magnesium-induced structural changes (Figure 16, and Table 7). Asidefrom having Z-axis acid pairs, the three magnesium sensitive peptides also havean aspartic acid residue in the +Z chelating position. Possession of a Z-axisacid pair however, does not necessarily confer magnesium sensitivity, as seenwith CaM:3Z(NDN), CaM:3Z(NES), and CaM:4Z(DES) (Figures 13 and 16). It isinteresting to note that the magnesium-sensitive CaM:4Z(DDN) differs from themagnesium-insensitive CaM:3Z(NDN) only by the residue in the +Y chelatingposition.TFE-induced conformation is observed to a large degree in all 3 peptides.CaM:4Z(DDS) has the greatest structural change ([]2 = 23,103) followed byCaM:4Z(DES) and CaM:4Z(DDN) ([]2 = 21,204 and 16,421 respectively,Table 7).The CD-monitored calcium titration curves of figure 17 show that eventhough CaM:4Z(DES) has data points parallel to CaM:4Z(DDN) andCaM:4Z(DDS), it fails to produce a sigmoidal curve. The Ka of CaM:4Z(DES) isin the order of 32 mM (p[Ca]f05= 1.5). CaM:4Z(DDS) has a 19-fold greateraffinity for calcium than CaM:4Z(DDN) (Ka = 29.2 vs. 542 iiM, Table 7). Thisdifference in calcium affinity is approximately the same when comparingCaM:3Z(NDS) with CaM:3Z(NDN) (section 11.4., Table 6). The only differencebetween the CaM:4Z and CaM:3Z peptides is the presence of an aspartateresidue in the +Y chelating position of CaM:4Z peptides, as compared to anasparagine residue in the CaM:3Z peptides. This suggests that aspartate in the+Y position produces peptides having greater affinities for calcium than peptideswith an asparagine in the same position. This is further supported by the 2-foldgreater calcium affinity of CaM:4Z(DDS) and CaM:4Z(DDN) compared to theirthree acid residue counterparts CaM:3Z(NDS), and CaM:3Z(NDN) respectively.84RESULTS1.21.11 .00.90.80.70.6f 0.50.40.30.20.10.0—0.1—0.2 -0 2 3 6p[Ca]Figure 17. CD-Monitored Calcium Titration of the CaM:4Z Peptides.OCaM:4Z(DDN) (peptide concentration = 0.65 mg/mL)•CaM:4Z(DDS) (peptide concentration = 0.51 mg/mL)VCaM:4Z(DES) (peptide concentration = 0.68 mg!mL)f is the ratio of the calcium-induced change in ellipticity at 222 nm to the maximumchange in ellipticity that can be elicited by calcium at 222 nm.Cell, path length is 0.1 cm.I I 1- I I1 4 585RESULTSLastly, as with the CaM:4X, and CaM:3Z peptides, there appears to be acorrelation between calcium affinity and the degree of calcium-inducedconformation among the CaM:4Z peptides.Figure 18 illustrates the CD-monitored magnesium titration ofCaM:4Z(DDS) and CaM:4Z(DDN). The magnesium dissociation constants forCaM:4Z(DDS) and CaM:4Z(DDN) are 17 and 54 mM respectively (Table 7).6. CaM:4XZ PeptidesFigure 19 shows that of the CaM:4XZ peptides, CaM:4XZ(NED) andCaM:4XZ(NEE) both exhibit the least amount of structure in the apo state ([e]222= 2,979 ± 24 and 2,365 ± 87 respectively, Table 8), whereas CaM:4XZ(NDD)and CaM:4XZ(NDE) show a significantly greater degree of initial structure ([]2= 6,673 ± 98 and 4,452 ± 61 respectively). This observation is also reflectedupon saturation of the peptides with calcium. In fact, CaM:4XZ(NED) andCaM:4XZ(NEE) both fail to show any noticeable change in conformation in thepresence of calcium, whereas CaM:4XZ(NDD) and CaM:4XZ(NDE), show a 12and 13 residue increase in the x-helical content respectively corresponding to[e]2’s of 17,167 ± 183 and 15,395 ± 92 respectively.Saturation of the peptides with magnesium fails to induce anyconformational change in any of the CaM:4XZ peptides. In all cases, thenumber of calculated residues in the cL-helical regions of the peptides remainsconstant in the absence, as well as in the presence, of magnesium except in thecase of CaM:4XZ(NED), where there is a decrease of one residue (Table 8).All of the CaM:4XZ peptides acquire a degree of cL-helical structure in thepresence of TEE. CaM:4XZ(NED) and CaM:4XZ(NEE) which show no responseto calcium or magnesium have an increase in their a-helical content by 11 and15 residues respectively (Table 8). In the presence of TEE, CaM:4XZ(NDD) and86RESULTS1.21.11.0 -0.9 -0.8 -0.7 -0.3- N!o ‘15 Z02.5 3.0p[CajFigure 18. CD-Monitored Magnesium Titration of CaM:4Z(DDS) and CaM:4Z(DDN).OCaM:4Z(DDS) (peptide concentration = 0.58 mg!mL)•CaM:4Z(DDN) (peptide concentration = 0.65 mg/mL)f is the ratio of the calcium-induced change in ellipticity at 222 nm to the maximumchange in ellipticity that can be elicited by calcium at 222 nm.Cell path length is 0.1 cm.874 2 0—2—4—6—8—10—12—14—16—18—20—22—24—26—28—30210220230240250210220230240250210220230240250210220230240250WAVLENGTH(nm)CaM:4XZ(NED)CaM:4XZ(NEE)a) 0 6 6 C) 0) a)Figure19.CDSpectraShowingMeanResidueEllipticityofCaM:4XZ(NDD),CaM:4XZ(NDE), CaM:4XZ(NED) andCaM:4XZ(NEE).1.Apopeptideinaqueousbuffer.2.Calcium-saturatedpeptideintheaqueousbuffer.3.Apopeptideinhydrophobicbuffer.4.Magnesium-saturatedpeptideintheaqueousbuffer.Aqueousbufferis100mMMOPS,150mMKCI,1mMNTA,pH7.2.Hydrophobicbuffer iseithera1:1or1:3(v/v)mixtureofaqueousbuffer inTFE.Cellpathlengthis0.1cm.Table8CorrespondenceBetweentheNumber ofResiduesina-HelicalRegionsandMeanResidueEllipticityValuesoftheCaM:4XZPeptides—[01222(Degcm2/dmol)(a)CalculatedResidues(C)4XZ-Peptides—Ca2.i-Ca2+Mg2+(b)+TFE(b)—Ca2+Ca2+Mg2+TFEKcaKMG(tiM)(p.M)CaM:4XZ(NDD)6,673±9817,167±1836,79220,90282082419.1±0.2CaM:4XZ(NDE)4,452±6115,395±924,56720,274518524430±15CaM:4XZ(NED)2,979±242,999±732,83113,08944315CaM:4XZ(NEE)2,365±872,337±61232115,05733318aMeanresidueellipticityat222nmisexpressedinmillidegrees±S.D.(n=3).bThedataindicatedaretheresultofasingleexperiment.CProceduretoestimatethehelicalcontentdescribedbyChenetaL,1974.00RESULTSCaM:4XZ(NDE) have [O]2’s of 20,902 and 20,274, respectively, whichcorresponds to both peptides having 24 residues in their a-helices.Figure 20 illustrates the CD-monitored calcium titration curves ofCaM:4XZ(NDD) and CaM:4XZ(NDE). Even though the conformationalcharacteristics of both peptides are similar (Table 8), CaM:4XZ(NDD) has a 23-fold greater affinity for calcium than CaM:4XZ(NDE) (KCa = 19.1 versus 430 itMrespectively, Table 8). CaM:4XZ(NDD) has the greatest affinity for calcium of allthe peptides in this study.Unlike CaM:4XZ(NDD) and CaM:4XZ(NDE), CaM:4XZ:(NED) andCaM:4XZ(NEE) fail to show the typical sigmoidal titration curves (Figure 21).This indicates a lack of calcium affinity by these peptides. Failure to produce asigmoidal curve upon calcium titration was also seen previously withCaM:3Z(NES) and CaM:4Z(DES) (Figures 14 and 17). It is interesting to notethat all four peptides have a glutamate residue in the +Z chelating position.90RESULTS1.21.11 .00.90.80.70.60.50.40..30.20.10.0—0.1—0.20 6p[Ca]Figure 20. CD-Monftored Calcium Titration of CaM:4XZ(NDD) and CaM:4XZ(NDE).OCaM:4XZ(NDD) (peptide concentration = 0.60 mglmL)•CaM:4XZ(NDE) (peptide concentration = 0.71 mg/mL)f is the ratio of the calcium-induced change in ellipticity at 222 nm to the maximumchange in ellipticity that can be elicited by calcium at 222 nm.Cell path length is 0.1 cm.I I I I I1 2 3 4 591RESULTS20f—i—2—3—40p[Ca]Figure 21. CD-Monitored Calcium Titration of CaM:4XZ(NED) and CaM:4XZ(NEE).OCaM:4XZ(NED) (peptide concentration = 0.80 mglmL)•CaM:4XZ(NEE) (peptide concentration = 0.95 mg/mL)f is the ratio of the calcium-induced change in ellipticity at 222 nm to the maximumchange in ellipticity that can be elicited by calcium at 222 nm.Cell path length is 0.1 cm.1 2 3 4 592RESULTSIII. DIMERIZATION1. CD AnalysesFigure 22 shows the CD spectra of CaM:4Z(DDS), CaM:4X(DSD) and a1:1 mixture of the two peptides. In each case, a spectrum is presented reflectingthe apo-peptide as well as a spectrum showing the response of the peptide(s) tocalcium. The CD spectra, along with the data of table 9 show the conformationalsimilarities between CaM:4Z(DDS) and CaM:4X(DSD). In the apo state,CaM:4Z(DDS) and CaM:4X(DSD) have 6 and 5 residues respectively making upthe x-helical regions of the peptide corresponding to [e]2221sof 4,700 ± 20 and4,653 ± 77 respectively. Addition of calcium results in CaM:4Z(DDS) having 18residues in the x-helical region whereas CaM:4X(DSD) has 19 residues in the chelical region (Table 9). Comparing these results with the results discussedpreviously for CaM:4Z(DDS) and CaM:4X(DSD) (sections 11.5 and 11.3respectively), good agreement is found for CaM:4X(DSD). Although theprevious ellipticity values recorded for CaM:4Z(DDS) in the absence andpresence of calcium (5,090 ± 229 and 21,707 ± 140 respectively, Table 7) aresomewhat greater than the values found in this section (4,700 ± 20 and 15,064 ±69, Table 9), the actual number of residues contributing to the ct-helical portionof the peptide is identical for the apo-peptide (6 residues, Tables 7, and 9) andonly differ by one residue for the calcium-saturated peptides (19 vs. 18 residues,Tables 7 and 9).When the peptides are mixed together in a 1:1 ratio, the CD spectraappears similar to that of either of the individual peptides (Figure 22). The meanresidue ellipticity values in the absence and presence of calcium are 4,851 ±150 and 15,815 ± 64 respectively, which falls within the error limits of bothCaM:4Z(DDS) and CaM:4X(DSD) (Table 9).934 2 0—2—4—6—8—10—12—14—16—18—20—22—24—26—28—30210220230240250210220230240250210220230240250WAVLENGTH(nm)Figure22.CDSpectraShowingMeanResidueEllipticityofCaM:4Z(DDS),CaM:4X(DSD), anda1:1CombInationofCaM4Z(DDS) andCaM:4X(DSD).1.Apopeptldeinaqueousbuffer.2.Calcium-saturatedpeptideIntheaqueousbuffer.Aqueousbufferis100mMMOPS, 150mMKCI,1mMNTA,pH7.2.Hydrophobicbufferiseithera1:1or1:3(v/v)mixtureofaqueousbufferInTFE.CellpathlengthIs0.1cm.CaM:4Z(DDS)CaM:4X(DSD)CaM:4Z(DDS)+CaM:4X(DSD)ci) 0 2 2 C.) C) ci) C., xTable9CorrespondenceBetweentheNumberofResiduesina-HelicalRegionsandMolarEllipticityValues inCaM:4Z(DDS), CaM:4X(DSD)anda1:1CombinationofCaM:4Z(DDS)andCaM:4X(DSD).—[1222(Degcm2ldmol)(a)CalculatedResiduesPeptides—Ca2+Ca2—Ca2+Ca2Ka(.tM)CaM:4Z(DDS)4,700±2015,064±6961835.4±3CaM:4X(DSD)4,653±7716,344±120519398±65CaM:4Z(DDS)4,851±15015,815±6461941.0±15tCaM:4X(DSD)425.4±196taMeanresidueellipticityat222nmisexpressedinmillidegrees±S.D.(n3).tKa’ScalculatedusingamodifiedPerrinandSayce(1967),nonlinearregressioncomputerprogramwrittenbyDr.TatsuyaFujimiya(seesectionIlL5.b.ofMaterialsandMethods).RESULTSThe sigmoidal curves of figure 23 represent the CD-monitored calciumtitration of CaM:4Z(DDS), CaM:4X(DSD) and the 1:1 combination of twopeptides. The p[CaJ05 values of CaM:4Z(DDS) and CaM:4X(DSD) individuallyare 4.45 and 3.40 respectively corresponding to Ka’S of 35.4 and 398 iiMrespectively (Figure 23 and Table 9). The Kca for CaM:4X(DSD) is within theerror limits of the Ka previously determined for the peptide on its own (Ka =407 ± 27, Table 5). Similarly, the Kca of CaM:4Z(DDS) in the mixture is veryclose to the previously determined Kca of 29.2 ± 1.0 mM (Table 7).Extrapolation of the p[Ca]f05 values for the 1:1 mixture is attained using acomputer derived nonlinear regression two component competition equation.The p[Ca]105 values obtained are 3.37 and 4.39 corresponding to Kca’s of 425.4± 196 and 41.0 ± 15 1iM respectively (Table 9). These values are in agreementwith the calculated Ka values of the peptides titrated individually.The Hill plot of figure 24 presents the data in a linearized form by plottinglog (f/I -f) versus p[Ca]. The Hill plot allows an assessment of the nature of themechanism of calcium binding in terms of cooperative or noncooperativebinding. The Hill coefficients (slope at the midpoint of calcium binding) are 1.23and 1.28 for CaM:4Z(DDS) and CaM:4X(DSD) respectively and 1.09 for the 1:1mixture. Since the Hill coefficients are all close to one, a cooperative bindingmechanism can be ruled out.2. Polyacrylamide Gel Electrophoresis of CaM:4Z(DDS), CaM:213,CaM:314, and TnC.Figures 25 and 26 show the polyacrylamide gels in which CaM:3/4,CaM:4Z(DDS), CaM:2/3 and TnC Ill are run. The conditions of theelectrophoresis are as mentioned previously (section lll.6.b of Material and961.21.11 .00.90.80.70.6f0.50.40.30.20.10.0—0.1—0.26p[Ca]Figure23.CD-MonitoredCalciumTitrationofCaM:4Z(DDS),CaM:4X(DSD),anda1:1CombinationofCaM:4Z(DDS)andCaM:4X(DSD).OCaM:4Z(DDS)(peptideconcentration=0.71mg/mL)•CaM:4X(DSD)(peptideconcentration=0.76mg/mL)VCaM:4Z(DDS)+CaM:4X(DSD)(peptideconcentration0.74mgImL)fistheratioof thecalcium-inducedchangeInellipticityat222nmtothemaximumchangeInelllptlcltythatcanbeelicitedbycalciumat222nm.Cellpathlengthis0.1cm.II2345RESULTS3V2— 0.0——I4— —— .•——0’•0’— ——_o_.—— —— — 0.•— —-— ‘j4.w U-——._o ——-‘•—- o-—0— o_—I- ———2—5—4—3p[Ca]Figure 24. Hill Plot of CaM:4Z(DDS), CaM:4X(DSD) and a 1:1 Combination ofCaM:4Z(DDS) and CaM:4X(DSO).VCaM:4Z(DDS) (peptide concentration = 0.71 mg/mI)•CaM:4X(DSD) (peptide concentration 0.76 mg/mL)OCaM:4Z(ODS) + CaM:4X(DSD) (peptide concentration = 0.74 mg!mL)f is the ratio of the calcium-induced change in ellipticity at 222 nm to the maximumchange in ellipticity that can be elicited by calcium at 222 nm.Cell path length is 0.1 cm.98RESULTS(—)Figure 25. Polyacrylamide Gel Electrophoresis of CaM:3/4 and CaM:4Z(DDS).Lane A: CaM:4Z(DDS) (conc. 1.0 mg!mL) with 0.25 M calciumLane B: CaM:4Z(DDS) (conc. 1.0 mgfmL)Lane C: CaM:4Z(DDS) (conc. 0.5 mglmL) with 0.25 M calciumLane D: CaM:4Z(DDS) (conc. 0.5 mg/mL)Lane E: CaM:3/4 (conc. 0.9 mg/mL) with 0.25 M calciumLane F: CaM:3/4 (conc. 0.9 mglmL)(+)99RESULTS(—)Figure 26. Polyacrylamide Gel Electrophoresis of CaM:2/3, CaM:4Z(DDS), and TnC Ill.Lane A: TnC Ill (conc. 3.0 mg/mL) with 0.25 M calciumLane B: TnC Ill (conc. 3.0 mg/mL)Lane C: CaM:4Z(DDS) (conc. 3.0 mg/mL) with 0.25 M calciumLane D: CaM:4Z(DDS) (conc. 3.0 mglmL)Lane E: CaM:2/3 (conc. 2.2 mglmL) with 0.25 M calciumLane F: CaM:2!3 (conc. 2.2 mglmL)(+)100RESULTSMethods). CaM:4Z(DDS), with a molecular weight of 3721, net charge of —4,and a mass to charge ratio of —930.25, demonstrates the least amount ofmigration. The rf value of CaM:4Z(DDS) is 0.42 regardless of peptideconcentration and whether or not calcium is present (Table 10, Figures 25 and26).Table 10Physical Characteristics of CaM:4Z(DDS), Troponin C Site III, CaM:213,and CaM:314 Used in Polyacrylamide Gel ElectrophoresisPeptide(s) I Molecular Net Charge MIZ ifWeight (daltons)CaM:4Z(DDS) 3721—4—930.25 0.42TnC III 3904—6 —650.67 0.61CaM:2/3 7726—9—858.44 0.49CaM:3/4 7842—12—653.50 0.67CaM:2/3 has the second smallest rf value (0.49, Table 10). Like CaM:4Z(DDS),the ri value is not affected by the presence of calcium. CaM:2/3 has a molecularweight of 7726 daltons, a net charge of —9 and a mass to charge ratio of —858.44(Table 10). Even though CaM:3/4 has a molecular weight similar to CaM:2/3(7842 vs. 7726 daltons respectively, Table 10), the migration of CaM:3/4 isapproximately 1.4 times greater than CaM:2!3 (Figures 25 and 26). The threeextra net negative charges associated with CaM:3/4 (compared to CaM:213)results in CaM:314 having a smaller mass to charge ratio than CaM:213 (—653.50vs. —858.44 respectively, Table 10). As with CaM:4Z(DDS) and CaM:2/3, themobility of CaM:3/4 is not affected by calcium (Figure 25). Lastly, TnC Ill, whichhas a molecular weight in proximity to CaM:4Z(DDS) (3904 vs. 3721 daltonsrespectively, Table 10) has an if value of 0.61 which is intermediate to CaM:2/3101RESULTSand CaM:3/4. TnC Ill has a net charge of —6 and a mass to charge ratio of —650.67 (Table 10). Once again, calcium does not influence the mobility of thepeptide.102DISCUSSIONDISCUSSIONI. STRUCTURE/ACTIVITY RELATIONSHIP OF CALCIUM AFFINITY OF THEHELIX-LOOP-HELIX CALCIUM BINDING MOTIFThe acid pair hypothesis is an attempt at an explanation for the variousaffinities for cations demonstrated by the hlh-calcium binding motifs found in anumber of calcium binding proteins including calmodulin, troponin C,panialbumin, oncomodulin, and calbindin. The structural similarity and aminoacid sequence homology among the calcium binding sites in these proteinsbelies the diversity of cation affinities and selectivities demonstrated by thesesites. The acid pair hypothesis assumes that the cation affinity of the hlhcalcium binding motif can be explained in terms of the number and location ofacidic amino acid residues in positions 1, 3, 5, 9 and 12 of the twelve-residueioop region of the motif. These five positions are located at the +X, +Y, +Z, —xand —Z coordinates of an octahedral coordination shell (Figure 27). RegardlessC-TERMINAL HEL(-xN-TERMINAL 1-IELLXFigure 27. Schematic Representation of the Twelve Residue Calcium Binding LoopRegion of the Helix-Loop-Helix Motif.-z103DISCUSSIONof cation affinity, the +X and —Z coordinate positions (i.e., positions 1 and 12 ofthe loop) are invariably aspartic and glutamic acid, respectively. This meansthat variation of the three residues in positions 3, 5 and 9 may provide a meansof changing the cation affinity of all hlh-calcium binding motifs. There is oneother chelating oxygen atom in the loop region at position 7 which provides the —Y coordinate of the octahedral coordination sphere. This atom is invariably thepeptide carbonyl oxygen of the amino acid in position 7 and therefore the sidechain of this residue does not interact directly with the cation and is notconsidered relevant for cation affinity by the acid pair hypothesis. It does notimmediately follow that all residues in the calcium binding proteins, the sidechains of which are not involved in cation chelation, do not affect cation affinity.It is our contention that the acid pair hypothesis is required, but not sufficient, fora complete description of cation affinity to the hlh-calcium binding motif.1. Evaluation of Serine and Asparagine in the +Z Chelating Position ofX-Axis Acid Paired HLH- Calcium Binding Motifs.The +Z coordinating position (position 5) is most frequently occupied byan aspartic acid residue that is paired with the invariant glutamic acid residue ofthe —Z position (position 12, Figure 27). The second and third most frequentlyoccurring residues in this position are serine and asparagine, respectively(Marsden et a!., 1990). During studies on hlh-calcium binding motifs that have asingle X-axis acid pair, the residue in position 5 cannot be an acidic residue so itis replaced in our models by either serine or asparagine. The effects of serineor asparagine in this position on cation affinity are found to be different, andtherefore, for comparison purposes, the residues are not freely interchangeable.CaM:3X(NSD), which has a serine residue in the +Z chelating position shows a104DISCUSSION29-fold lower affinity for calcium than CaM:3X(NND) which has an asparagineresidue in the same position (Kca = 15.4 mM vs. Ka = 524 tiM, respectively,Table 11). Similarly, CaM:4X(DSD) shows a 10-fold lower affinity for calciumthan CaM:4X(DND) (Kca = 407 jiM VS. Kca = 42.1 jiM, respectively, Table II).All other variables being equal, the asparagine residue in the +Z position resultsin a model motif with significantly greater cation affinity than an otherwiseidentical model motif with serine in that position.The differences in calcium affinities may be attributed to differentconformations of the calcium binding sites that are influenced by the 4 and ‘vangles of the residue occupying the +Z position. The commonly observed wangle for serine range from 150-180°, which is considerably greater than the ‘vangle of asparagine and aspartic acid residues which ranges from 75-135°(Matthews, 1993). In the absence of calcium, both CaM:3X(NSD) andCaM:4X(DSD) show very little cL-helical structure (Figures 9 and 11; Table 11).Five residues make up the x-helical portion of both apo-peptides compared tosix to eight residues in the hlh-calcium binding motifs showing high affinity forcalcium (see CaM:4X(DND), CaM:3Z(NDS) and CaM:4Z(DDS), Table 11). It ispossible that a certain amount of preformed structure is required by hlh-calciumbinding motifs prior to binding the calcium cation and a serine residue in the +Zposition does not provide the preformed structure needed for a high calciumaffinity motif. The detrimental effect of serine in the +Z chelating position onapo-peptide structure is further demonstrated in the presence of trifluoroethanol.High affinity peptides show 27-28 residues of &-helical structure in the presenceof the structure forming solvent whereas CaM:3X(NSD) and CaM:4X(DSD) show21-22 residues of cL-helical structure in the same solvent (Figures 9 and 11;Table 11).105901C)oI00)—1CD CD0.—CD 0)•ri0.=CDCD0)F%)==0.-.‘(ftCDCn>((0CDCDCDx0.-o-.CDZIoo°oo°°ooo°o 0)0)0)0)0)0)0)0)0)0)0)0)0)0NNNNNNXXxxxXxx- ..—.-‘——‘-‘—.——.-%-%—.r-,—OQQzZZQQDQZzZZ’rnmmQQmQQZZCZZO)1,)’mI]jg 9W9!J3J9r.3---Cfl0--400)-.1F’.)0)0)40) 0CD0)COF’.)-0)-- 040(flC7ICo—J.)F%)00)010)O)0% *H*HHFfFf*F’.)-‘.3F’)-k-.h%)F’)--CDF’.)0(DO)COC,)0Cii.(71F’)(ØO)kC04C0(31-F’)F’.)-.-—O)-JUDC,),)Ci)CD---4C.)0)Cii.F’.)(,)F’)F’)-0 C)CO0)0---C.)0)0)Cii0)Cii+— -.1-4000-C,)F’)-.--cn0**HHFf**Ifj•f0)0)-Si-0)-F.) -Ci)0).-COF’.)--00CD-.10Ci)0LCi)(Il-0-Si(0——I’. —h+3 Ci).0)—J..CO--4(31-F’)(iiCiiF’)..0F’)c,0)CO0)0)(0(71—10)CDCOF’.)CD0)CDCO0.0F’)F’3.O)0)F’)F’)0)O0)C--F’)F’)F’)--F’)-—F’)F’)—‘-F’)---+C31(i)00O)0)-JC,)OsJ0)..j0)(D(D-IOOF’)C001F’)(DO)Ui(O0-..j Ui0)-1000F’)COCii-.JF’)01C,)-.J01(D0)010)-CD0..3.01-40)-jCD-4UiocIC)C)> rC•) -—C-F’)-F’)hF’)•F’)0)0001C.)C,)-1.0).CD0)-(310)m -—C.)Ci)(310)0)0)0)Cii.-4C,)0).-4.(31(D-—0+3F’3F’) CO01...010)puJF’,1mC,)-!“C.)01(00)4..1(n0)--4 0Ui!O00.D(,)fl,0k0F’)F’)0)00—0)-Je0(Ti0)0) 01010)00-Cfl!111111111111111111R14§1a)CDCDOC.,—I.CD‘CDCD(Q0•C.,a,Im-NOISSflOS1ODISCUSSIONThe calcium binding sites in troponin C and calmodulin that contain serinein the +Z chelating position of the loop region do not help in the interpretation ofthe present data. Rabbit skeletal troponin C site II is a low affinity calciumbinding site and contains a serine residue in the +Z position. While this is aplausible explanation for the low affinity of this site in skeletal TnC based on thisstudy, other studies using synthetic peptide analogs of skeletal TnC have shownthat the low affinity of this site may be due to the non-chelating glutamic acidresidue in position 2 of the loop region (Shaw et al., 1991). These authors,working with the chicken skeletal TnC sequence, substituted the chelatingresidues of the high affinity site Ill (SCIII, Figure 28) with the chelating residuesof the low affinity site II (LII, Figure 28), and found that this new hybrid site (LIIL,Figure 28) had an affinity for calcium on par with the affinity of a synthetic modelof the high affinity site Ill. This suggests that the non-chelating residues of siteIll are responsible for the high calcium affinity of skeletal TnC site Ill. The hybridpeptide (LIIL, Figure 28), discussed by these authors has a loop sequence verysimilar to CaM:4X(DSD) (Figure 28) which has a low affinity for calcium (Kca =407 jiM). Therefore, the suggestion that replacing the lysine residue in the non-chelating position 2 of the loop region with glutamic acid is the cause of lowcation affinity in skeletal TnC site II does not correlate with the low affinity ofCaM:4X(DSD). The major difference between the amino acid sequence of theloop regions of the hybrid peptide (LIIL, Figure 28), and CaM:4X(DSD) is theresidue type in the positions 10 and 11 of the loop. The hybrid peptide containsisoleucine and glutamate while CaM:4X(DSD) contains two alanine residues. Inview of our results indicating that a serine in the +Z position is detrimental tocalcium binding, we can only speculate that the isoleucine and glutamateresidues in positions 10 and 11, respectively, of the hybrid peptide compensatefor the detrimental action of the serine in the +Z position of this peptide, whereas107LOOP123456789101112N-TERMINALHELIXC-TERMINALHELIXXYZ-Y-X-zKCaSCIIIKSEEELANAFRIFQKJAGYIIEELGEILRATG3MLIILKSEEELANAFRIFKQAGTl.lELGEILRATG8iiMLIIKSEEELANAFRIF2EGGTI2FELGEILRATG3mMCaM:4X(DSD)SEEEIREAFRVF2K2GGYIPAALRHVLTNLG4O7iiMCDsiteParvalbuminSADDVKKAFAlIQQ,KGFIEDLKLFLQNFKO.4-2nM(Carp4.25)Figure28.AminoAcidSequenceofSCIII,LIIL,LII,CaM:4X(DSD),andtheCDSiteofParvalbumin(Carp4.25).TheN-andC-terminalhelicalregionsflankingthecalcium-bindingloopregionareindicatedbyhorizontalbars.Theloopregionisindicatedbyahorizontalbar,andthesequencepositionsintheloopregionarenumbered1through12startingwiththeN-terminalresidue.Theaminoacidresidueswithsidechainsinteractingwithcalciumareinboldtextandunderlined,andthepositionofthechelatingresidueintheoctahedralarrangementofdentatesisindicated.The—Ypositionisassumedtobethepeptidecarbonyloxygenoftheresidueinposition7oftheloopregion.DISCUSSIONthe alanine residues in the corresponding position of the loop region ofCaM:4X(DSD), do not compensate for the serine residue, It is difficult toreconcile the difference in cation affinities between Shaw’s hybrid peptide andCaM:4X(DSD) and it remains to do further studies to solve this problem.Further complications arise when examining other natural calcium bindingproteins containing the hlh-calcium binding motif. Parvalbumins contain acalcium binding site with serine in the +Z chelating position, and this site (Figure28) has high calcium affinity (Kca = 0.4 -2.0 r1M) (Moeschler eta!., 1980; Haiechet al., 1979). In view of the detrimental effect of serine in the +Z position oncalcium binding, there may be a compensation in the parvalbumin site for thisanticipated detrimental action of serine. Inspection of the sequences ofCaM:4X(DSD) and the CD calcium binding site of carp parvalbumin (Figure 28)suggests two possibilities. The glutamic acid in the —X position of theparvalbumin site II is a strong candidate for the role of compensator and isaddressed later (Section 1.3 of Discussion). The residues in positions 10 and 11of the loop region of carp parvalbumin are also possible differences that may insome way compensate for the serine residue in the +Z position; however, asmentioned for the hybrid peptide above, this is only speculation and will have tobe investigated.2. Evaluation of Serine and Asparagine in the —X Chelating Position ofZ-Axis Acid Paired HLH-Calcium Binding Motifs.Aspartic acid is the most common residue at the —X position paired withthe invariant aspartate at the +X position to give an X-axis acid pair. Serine,glutamic acid, and asparagine are the second, third, and fourth most common —X109DISCUSSIONligands, respectively. For studies on the hlh-calcium binding motifs with a singleZ-axis acid pair, we require a non-acidic residue in the —X position, whichexcludes aspartate and glutamate from consideration. Again, studies usingsynthetic motifs with serine or asparagine in this position indicate that the effectsof these residues in this position on calcium affinity are not identical and,therefore, for comparison purposes, these residues are not freelyinterchangeable. Replacement of a serine residue in the —X chelating positionof CaM:3Z(NDS) with an asparagine residue to give CaM:3Z(NDN) results in a17-fold decrease in affinity for calcium (Kca = 58.8 iM VS. Kca = 1000 riM, Table11). Similarly, CaM:4Z(DDN) with an asparagine residue in the —X chelatingposition has a 19-fold lower affinity for calcium than CaM:4Z(DDS), which has aserine residue in the —X chelating position (Kca = 542 iiM VS. Kca = 29.2 1tM,respectively, Table 11). CaM:4Z(DDN) with four acid residues in chelatingpositions has a 9-fold lower calcium affinity than CaM:3Z(NDS) with only threeacid residues in chelating positions (Kca = 542 j.tM VS. Kca = 58.8 tM, Table 11).This may be due to the detrimental effect of the asparagine in the —X chelatingposition of CaM:4Z(DDN).Our studies with the synthetic peptides indicate that the models with aserine residue in position 9 have a significantly greater affinity for calcium thanotherwise identical peptides with asparagine in the same position. This may bethe result of serine participating in a reciprocal backbone-side-chain hydrogen-bonding interaction with the glutamate in the —Z position. This type ofarrangement in the N-terminus of an x-helix is referred to as a “capping box” andmay be significant in increasing the stability of a protein (Serrano and Fersht,1989; Bell eta!., 1992; Harper and Rose, 1993; Matthews, 1993). In a cappingbox, the side chain of the Ncap residue (N-terminal residue of the c-heIix) formsa hydrogen bond with the backbone NH of the N+3 residue, and reciprocally, the110DISCUSSIONside chain of the N+3 residue forms an H-bond with the backbone NH of theNcap residue. The occurrence of a capping box is dependent in part, on thenon-helical dihedral angles of the N-terminal residue (N-cap) of the ct-helix, theangle in particular. In an assessment of previously determined capping boxes,the angle of the N-cap residue was in the range of 167° ± 5° that coincideswith the ii angle of serine (150°-I 80°) (Harper and Rose 1993; Matthews, 1993).This suggests the possibility that CaM:3Z(NDS) and CaM:4Z(DDS) possesscapping boxes that could induce structure-stabilizing properties that aremanifested by an increase in affinity for calcium. Support for this hypothesiscomes from crystal structures showing capping boxes in the high affinity calciumbinding sites of calbindin and troponin C (Szebenyi and Moffat, 1986; Satyshuretal., 1988).The detrimental effects of asparagine in the —x chelating position (N-cap)may also be explained in terms of the capping box. Even though asparagine inthe N-cap position of a-helices has been observed with high frequency, it isenergetically unfavorable relative to serine (Serrano and Fersht, 1989; Bell eta!., 1992). Figure 29 helps to explain the observed difference in stabilizationeffects between ct-helices when either serine or asparagine is in the N-capposition. In the case of serine (Figure 29a), the y-hydroxyl accepts a hydrogenbond (dotted) from the backbone amide of residue N3 (equivalent to the —Zposition in our hlh-model) and forms a capping box. The relative position of thishydrogen-bond donor and acceptor is largely determined by the backbone ‘vangle and the side chain torsion angle Xi• Asparagine, on the other hand, hasan additional degree of freedom associated with its side chain (rotation aroundX2; Figure 29b), which must be taken into consideration. This factor, along witha smaller ii angle and larger side chain, relative to serine, creates anunfavorable contact (arrowheads) between asparagine’s side-chain oxygen atom111DISCUSSIONN-capSecN-capAsnFigure 29. Illustration of the Difference in Backbone Geometry When Either Serine orAsparagine is Found in the N-cap Position. (Taken from Bell eta!., 1992).and the 13-carbon of residue N2 (Figure 29b). It is conceivable that such closecontacts occur in CaM:3Z(NDN) and CaM:4Z(DDN) resulting in reducedaffinities for calcium. However, it should also be mentioned that close contactsdo not always occur when asparagine forms the N-cap of an cL-helix (Bell et aL,1992). The investigators suggest that the key to avoiding these unfavorableclose contacts may be dependent on the N’ angle of the N-cap residue (Bell etal., 1992). Finally, the fact that the calcium-saturated peptides with anasparagine in the —X chelating position display a smaller degree of cL-helicalxlab112DISCUSSIONcontent compared to peptides with a serine residue in the same position,provides further support for the suggestion that the capping boxes are absent inCaM:3Z(NDN) and CaM:4Z(DDN).3. Evaluation of Glutamate and Aspartate in the —X Chelating Positionof X-Axis Acid Paired HLH-Calcium Binding Motifs.To investigate the influence of direct chelation by glutamic acid in the —xchelating position of an hlh-motif on calcium affinity, peptides with glutamic acidin the —x position are compared to peptides having an indirect chelating asparticacid residue in the —X position. The residue occupying the +Z chelating positionof these peptides is asparagine. In the h!h-motifs having three acid residues inchelating positions, replacement of the aspartic acid in the —X position ofCaM:3X(NND) with glutamic acid to give CaM:3X(NNE), results in a 6-folddecrease in the calcium affinity (Ka = 524 iiM vs Kca = 3140 pM, Table 11). Amore pronounced decrease in calcium affinity is observed in the hlh-motifshaving 4 acidic residues in chelating positions. Replacing the aspartic acidresidue in the —X position of CaM:4X(DND) by a glutamic acid to giveCaM:4X(DNE) results in a 46-fold decrease in calcium affinity (Kca = 42.1 iM VSKca = 1950 p.M respectively, Table II). Comparison of the affinities ofCaM:4X(DND) (Kca = 42.1 j.tM) with CaM:3X(NND) (Ka = 524 pM) andCaM:3X(NNE) (Kca = 3140 1iM) indicates that the additional acid residue foundin the +Y chelating position of CaM:4X(DND) results in an increased affinity forcalcium. This is predicted by the acid pair hypothesis which states that calciumaffinity of an hlh-calcium binding motif increases with the number of acidresidues found in the chelating position up to a maximum of four (Reid, 1987).However, in the case of CaM:4X(DNL), the acid pair hypothesis does not explain113DISCUSSIONits lower affinity compared to CaM:3X(NND) (Ka = 1950 and 524 1iMrespectively, Table Ii). These results are interesting since it was expected thatthe direct chelation of a glutamic acid residue in the —x position with the calciumcation would result in an increased affinity for calcium. With very few exceptions,the occurrence of glutamic acid in the —X chelating position of hlh-motifscoincides with a serine residue in the +Z position (Marsden et a!., 1990). Theappearance of these two residues in the +Z and —X positions of calcium bindingsites may lead to a hydrogen-bonded interaction which could serve to stabilize afavorable conformation and create a charge distribution system which couldconceivably increase the electron density on the serine oxygen and therebyincrease the interaction of this residue with the cationCH/Ca / CH Glu (-X)I &CC12‘CHSer (÷Z)Figure 30. Postulated Glu!Ser Side Chain Hydrogen-Bonding and Interaction withCalcium.(Figure 30). This postulated positive contribution of serine towards cationbinding of the hlh-motif is in direct opposition to the demonstrated detrimentaleffect toward calcium binding of this residue in the +Z position previouslydiscussed (Section 1.1 of Discussion). It is possible that the postulated114DISCUSSIONhydrogen-bonded system between serine in the +Z position and glutamate in the—X position may overcome the detrimental effect of these residues that isexhibited when they are not found together. To investigate this occurrence, hlhmotifs were prepared with a serine in the +Z position and a glutamic acid in the —X position.The combined effects of serine in the +Z position and glutamic acid in the—x position are examined such that the number of acidic residues in chelatingpositions remain constant. In peptides having 3 acid residues in chelatingpositions, the replacement of asparagine in the +Z position of CaM:3X(NNE)with serine to give CaM:3X(NSE), results in a 6-fold decrease in calcium affinity(Kca = 3140 1iM VS. Kr,a = 19.6 mM, Table 11). Likewise, in a peptide with 4acidic residues in chelating positions, replacing the asparagine in the +Zposition of CaM:4X(DNE) (Ka= 1950 j.tM, Table 11) by a serine residue to giveCaM:4X(DSE) (Kca = 2806 1iM, Table 11), also results in a decrease in calciumaffinity of only 1.4-fold. As predicted by the acid pair hypothesis, the addition ofan anionic residue in the +Y position of CaM:3X(NSE) to give CaM:4X(DSE)results in an increase in calcium affinity (Kca = 19.6 mM vs. 1++ 2>>jD--C—. 0 X-IC)C?).--i...)-C?)F.,)C?)C?)01F.)-————————>rID--‘—.CDCDCD0——CD=•0— ——.CDDCDsC)XXIIII>00 ++CD00ci)-l——CDx-‘—xlCD>>C,—.CD-1C)()A-.-ci)L)r..)ci)Ci)Ui--UiPJ—————————3mr%)-0-01%.)F%31%.)Ci)Ci).00.CD)I-o—S—.U)CDXXIIiC)>++-II CDCDG)oQø-4>—-‘—CDZCl)<0X-‘XCD-I••C)-.r-crr.cC,01r--—————————I3m•.h-CG)(31--.CDXZ-bDè)èo-:CoCDC)0,CD--C3icoCo010--Ie..--.b.(IlC,)0)01-CDF3.—;,———————————I3m (D.C,0)0)CD—JLi.)1..)CO-)CDCD-.1CDcCo0-h)0)C,).—1..0).-.1--I—.C,)-C..)‘3C..).01(31(31.3C,)C)—;————————————I3mL)(DXCOC310)-0)CDC..)0)M--40--40COCo0)010)(31CDCD0)—J-4—z.C..)-C.)-I’.)C..)0)—1F..)..--l———————3mC)C,)OF.)C.),•U1-1FJ00CD--I0)0I—J0Z—j0)Co--0-r..CoF.)CDC C. C.)-1C.;;;