Author's personal copy - Duquesne UniversityProtein dimerization is essential for cellular processes...

This article was published in an Elsevier journal. The attached copyis furnished to the author for non-commercial research and

education use, including for instruction at the author’s institution,sharing with colleagues and providing to institution administration.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/copyright

http://www.elsevier.com/copyright

Author's personal copy

Stalking metal-linked dimers

Kristina O. Pazehoski a, Tyler C. Collins b, Robert J. Boyle b,Michael I. Jensen-Seaman c, Charles T. Dameron d,*

a Division of Natural Sciences, University of Pittsburgh at Greensburg, PA 15601, USAb Department of Chemistry and Biochemistry, Duquesne University, Pittsburgh, PA 15282, USA

c Department of Biological Sciences, Duquesne University, Pittsburgh, PA 15282, USAd Department of Chemistry, Mathematics and Physical Sciences, Saint Francis University, Loretto, PA 15940, USA

Received 20 July 2007; received in revised form 2 October 2007; accepted 18 October 2007Available online 28 November 2007

Abstract

Protein dimerization is essential for cellular processes including regulation and biosignalling. While protein–protein interactions canoccur through many modes, this review will focus on those interactions mediated through the binding of metal ions to the proteins.Selected techniques used to study protein–protein interactions, including size exclusion chromatography, mass spectrometry, affinitychromatography, and frontal zone chromatography, are described as applied to the characterization of the Enterococcus hirae proteinCopY. CopY forms a homodimer to control the expression of proteins involved in the homeostasis of cellular copper levels. At the centerof the CopY dimerization interaction lies a metal binding motif, –CxCxxxxCxC–, capable of binding Zn(II) or Cu(I). The binding ofmetal to this cysteine hook motif, one within each monomer, is critical to the dimerization interaction. The CopY dimer is also stabilizedby hydrophobic interactions between the two monomers. The cysteine hook metal binding motif has been identified in numerous otheruncharacterized proteins across the biological spectrum. The prevalence of the motif gives evidence to the biological relevance of thismotif, both as a metal binding domain and as a dimerization motif.� 2007 Elsevier Inc. All rights reserved.

Keywords: Copper; Copper metabolism; Dimerization; Frontal zone chromatography; Metal binding sites; Protein–protein interactions

1. Introduction

The study of protein interactions is of growing impor-tance in proteomics, with applications ranging from theexamination of individual protein complexes to character-izing networks of protein interactions. Dimerization, thesimplest of protein–protein interactions, is critical to bio-signalling pathways where extracellular messages stimulateintracellular regulatory events to occur. The stimuli origi-nating from outside the cell either induce or inhibit thedimerization of protein complexes where the interactionacts as a regulatory switch. The monomer to dimer associ-ation may be regulated within the pathway or may be a

part of the regulatory mechanism. The dimerization maybe reversible as in the heterologous interaction betweentrypsin and its peptide trypsin inhibitor, or irreversible asin the interaction of an antibody binding to its antigen[1]. Dimerization interactions can occur through directinteractions between proteins or be mediated through con-formational changes induced by cofactor binding. Ca2+

binding, for instance, alters the conformation of cadherinto shift the monomer–dimer equilibrium towards dimerformation [2].

Metals and metalloids are intimately involved in the reg-ulation of genes and proteins involved in their uptake, uti-lization and detoxification. The regulation occurs duringtranscription, translation, and directly at the protein level.For instance, at the transcriptional level metals regulate theexpression of import/export pumps, metallochaperonesand the metallothioneins. Regulation through the sensing

0162-0134/$ - see front matter � 2007 Elsevier Inc. All rights reserved.

doi:10.1016/j.jinorgbio.2007.10.027

* Corresponding author. Tel.: +1 412 396 1894; fax: +1 412 396 5683.E-mail address: [email protected] (C.T. Dameron).

www.elsevier.com/locate/jinorgbio

Available online at www.sciencedirect.com

Journal of Inorganic Biochemistry 102 (2008) 522–531

JOURNAL OFInorganicBiochemistry


of the intracellular copper levels by regulatory proteins hasbeen widely studied. The control mechanisms often featurecopper responsive transcription factors or repressors; inthese proteins the copper, in its cuprous form (Cu(I)), hasstructural and regulatory roles. Copper(I)-responsivetranscription factors are used to sense changes in the intra-cellular metal concentrations and correspondingly up- ordown-regulate the transcription of copper(I) acquisition,distribution, and sequestration genes. Eukaryotic Cu(I)-responsive transcription factors include AceI and MacI ofthe organism Saccharomyces cerevisiae. Ace1 activatesthe transcription of copper sequestration and detoxifica-tion genes under conditions of copper excess [3–5], whileMac1 activates the Ctr1 and Ctr3 high-affinity copperuptake transporters under conditions of copper depletion[6,7]. In a similar vein, copper(I) has been shown to regu-late the CopY repressors in bacteria. CopY and its operonwere initially described in Enterococcus hirae; completesequencing of �500 bacterial genomes shows it to be foundin about 40 related organisms and is phylogeneticallyrelated to the Bla/Mec repressor proteins (Fig. 1). TheCopY repressor regulates the cop operon that is the back-bone of copper uptake, use and efflux pathways in manybacteria. Recent studies, as highlighted below, show thatthe system is dependent both on the zinc induced dimeriza-tion of CopY and on the changes copper(I) binding inducesin the structure.

The intention of this review is two fold. First, it is hopedthat it can serve as a guide to those investigating the rolesthat metal ions play in monomer–oligomer equilibria. Weare not attempting to review all of the many varied pro-cesses used to study protein–protein association, but onlyselected techniques used within the investigator’s labora-tory to study a series of related native and synthetic fusionscontaining the –CxCxxxxCxC– metal binding motif (x =any amino acid). The motif serves, as described below, tohook the monomers together and has been named the cys-teine hook motif. Sufficient detail is provided to illustrateeach technique and how it is applied to the study of ahomodimeric protein in a rapid equilibrium with its mono-mer components. For in depth description of the tech-niques the reader will be directed to specific monographson the topic. The secondary goal of this review is to illumi-nate a new metal requiring dimerization motif firstdescribed in a bacterial regulatory protein, CopY. Themotif’s description is unique due to the pronounced effectthat bound Zn(II) or Cu(I) has on the monomer–dimerequilibrium of proteins with which it is associated.

1.1. Regulation of the cop operon and CopY

One of the best understood copper homeostasis systemsis that of the Gram-positive bacterium Enterococcus hirae

[8–15]. The control of copper concentration in this bacte-rium is regulated by the cop operon, which contains thefour genes copY, copZ, copA, and copB. The copY geneencodes a protein that acts as a repressor of the cop operon.

The expression of all the genes on the cop operon is con-trolled by this repressor, which is sensitive to cellular levelsof copper [16]. CopZ encodes a copper chaperone that car-ries copper throughout the cell and shields the cytoplasmfrom the toxic effects of the copper ion, while ensuringthe copper is delivered to the correct target [16,17]. ThecopA and copB genes encode two P-type copper ATPasesthat act as Cu(I) import and export pumps, respectively[18]. The four proteins encoded by the genes on the cop

operon interact in specific ways to regulate levels of intra-cellular copper (Fig. 2).

The metallated state of the Enterococcus hirae CopYmetal binding site not only affects the DNA bindingcapability of the protein, but has also been shown to influ-ence the dimerization [19]. CopY, a member of the Bla/Mec repressor family, is the archetypical prokaryotic cop-per regulated repressor [20]. In the DNA binding state,CopY binds one Zn(II) ion in a –CxCxxxxCxC– motif.The binding site is located at the extreme C-terminus ofthe protein [19]. Zn(II)CopY binds to DNA as a homodi-mer across a 27-base pair inverted repeat, identifiedby DNase I footprinting experiments [21]. Subsequentsite-directed mutagenesis studies indicate that two invertedrepeat sequences, TACAnnTGTA, positioned 61 and 30base pairs before the transcription start site of the cop

operon, are crucial to the protein–DNA interaction[21,22]. A very similar DNA motif serves as a binding sitefor BlaI (from Bacillus licheniformis) and MecI (fromStaphylococcus aureus), structurally characterized [23–25]proteins from the b-lacatamase family. The N-terminusof CopY exhibits significant sequence similarity to the b-lactmase family [19,21] as well as to the DNA bindingregion of the phage k cro repressor [26]. A diglutaminemotif (QQ) of the cro repressor, also found in CopY, hasbeen shown to specifically interact with the ACA sequencewithin the TACAnnTGTA DNA motif [26]. While it hasbeen suggested that CopY may be binding to DNA inthe same fashion [27], the QQ residues are not at all con-served across bacterial homologs.

1.2. The structure of CopY

The C-terminal portion of CopY, containing the metalbinding motif, also plays a role in DNA binding, althoughit most likely does not physically interact with the DNA. Itis clear that the metals, 1 zinc(II) or 2 copper(I)s, areligated in a conserved –CxCxxxxCxC– motif [16]. DNAbinding depends on the type of metal bound to CopY, asdemonstrated by electrophoretic band shift assays [16,19].To maintain DNA binding activity Zn(II) must be boundto the protein [21]. Within the Bla/Mec family of repressorsthe affinity for DNA is strongly linked to the formation ofdimers [20]. Cu(I) ions displace the Zn(II) in the CopYmetal binding site, prompting the protein to lose affinityfor DNA [16]. The change in coordination potentiallychanges the protein conformation in order to affect theDNA binding state of the CopY protein. The 2Cu(I)

K.O. Pazehoski et al. / Journal of Inorganic Biochemistry 102 (2008) 522–531 523


protein stays in its dimeric form but is presumed to have analtered conformation. X-ray absorption analysis of the 2Cu(I)CopY complex suggests that each Cu(I) ion is coordi-nated in a trigonal geometry with sulfur ligands. The X-rayabsorption data verifies the cuprous oxidation state of thebound copper, and indicates a Cu–S bond length of 2.26 A,

which is typical of trigonally coordinated Cu(I). The dataalso give evidence of a Cu–Cu interaction at a distanceof 2.69 A, further supporting the notion that 2Cu(I) ionsare bound by the –CxCxxxxCxC– metal binding site. Acompact Cu(I)2S4 core appears to form upon Cu(I) coordi-nation by CopY, and two of the four sulfur ligands pro-

Fig. 1. CXCX(4–6)CXC motif in CopY homologs. Phylogenetic tree of amino acid sequences similar to E. hirae CopY, obtained by BLASTing (cutoff e-value of 0.00001) CopY to 496 complete bacterial genome sequences. Peptides containing all four cysteines in the –CXCX(4–6)CXC– motif are followed bya closed circle; those with three cysteines are followed by a closed triangle; while those with two cysteines are indicated by an open triangle. True CopYhomologs appear to be restricted to the Lactobacillales, and are related to a larger group of transcriptional repressors of b-lactamases.

524 K.O. Pazehoski et al. / Journal of Inorganic Biochemistry 102 (2008) 522–531


vided by the cysteine residues are shared between the twobound Cu(I) [28].

The general purpose of these investigations into CopYand related proteins described below is to fully characterizethe structure and function of the –CxCxxxxCxC– dimeriza-tion motif. Recent studies have shown that the C-terminalsequence is sufficient to induce dimerization of normallymonomeric proteins (K.O. Pazehoski, unpublished data).In addition to the –CxCxxxxCxC– motif at the extremeC-terminus, which has been shown to bind metals and isimportant to the regulation of the cop operon [19,29,30],adjacent to the cysteine motif on the N-terminal side is anon-leucine zipper aliphatic repeat. We have hypothesizedthat both the metal binding sequence and the aliphaticrepeat contribute to the formation of dimers. The dimeriza-tion data below illustrates the pathway used to initiate thecharacterization of metal induced dimerization. The datasupports the hypothesis that both the aliphatic repeatand the metal binding motif are important to the dimeriza-tion process.

The determination of the roles of both the aliphaticrepeat sequence and the cysteine-rich metal binding motifin CopY dimerization would be greatly facilitated if thethree-dimensional structure of the C-terminal region wasknown. Earlier attempts at investigating the metal bindingdomain structure by NMR were initiated through the prep-aration of a truncated version of CopY. The shortenedprotein included the 38 amino acids at the C-terminus ofCopY, and was termed ‘‘Ymbs” (CopY metal binding site).An accurate NMR solution structure was unable to beobtained due to a flexible, unstructured region encom-

passed by the amino acids before the metal bindingdomain. Solubility issues at high protein concentrationswere also a concern, as is the case with the full CopY pro-tein. Secondary structure prediction, performed throughthe PSI-PRED protein structure prediction server [31,32],suggests that these N-terminal residues of Ymbs may forman a-helical structure. To make NMR studies on Ymbsmore feasible, the truncate was fused to another proteinand classified as a ‘‘solubility enhancement tag” (SET)for further studies. The immunoglobulin binding domainB1 of the streptococcal protein G (abbreviated as GB1)was chosen as the SET because of its high stability, solubil-ity, known monomeric structure, lack of cysteine residuesand its small size (56 amino acids) (Fig. 3). GB1 has been

Fig. 2. Copper homeostasis in Enterococcus hirae. Several proteins are required to control the intracellular copper concentration in E. hirae. Copper(I) isimported into the cell through the CopA ATPase, which possesses an N-terminal cytoplasmic metal binding domain. The copper chaperone protein,CopZ, is proposed to interact with the CopA metal binding domain [65]. CopZ then delivers Cu(I) to the repressor protein, CopY [16]. Cu(I) replacesZn(II) in the metal binding site of CopY, and the repressor protein releases from the promoter DNA of the cop operon. The copY, copZ, copA and copB

genes of the operon are then transcribed. Excess Cu(I) can be exported by the CopB ATPase [66]. The fates of the displaced Zn(II) and the CopZchaperone after it has delivered its Cu(I) are unknown.

Fig. 3. Construction of the GB1-Ymbs38 fusion protein. The knownthree-dimensional structure of GB1 is shown (PDB# 2GB1). Although thestructure of the CopY fragment, Ymbs38, has not been solved, acombination of secondary structure prediction [67] and preliminaryNMR results allow for a hypothesized structure to be prepared. Theamino acid sequence is shown under the corresponding structuralcomponents. Hydrophobic amino acids that are postulated to participatein dimerization are shown in bold italic letters. The cysteines that areresponsible for metal coordination are underlined (highlighted in ball andstick form in the proposed structure).



used in several instances as a SET to aid in NMR studies ofproteins that are usually very difficult to solubilize at thehigh concentrations required by NMR. GB1 is particularlysuitable for both dimerization studies and structural analy-sis of Ymbs because of its known monomeric formation, itsinability to bind metal ions, and its extensive structuralcharacterization by NMR methods [33–38].

1.3. Qualitative identification and characterization of dimers

Virtually all dimer and higher order protein multimerscan dissociate into their monomeric constituents. The asso-ciation constant and other factors determine how easily themonomer–dimer relationship can be visualized. In manycases the protein complexes appear to be essentially stableover the time scale of the experiment and can be easily iso-lated in their polymeric form. Less recognized are thosemonomeric–dimeric (or polymeric) proteins that are in arapid equilibrium. At the typical concentrations and timeframes that these proteins are observed, it may not be clearthat multiple species are present in the solution. The behav-ior of many proteins falls into this category. If it is suspectedthat the protein or proteins being studied can form dimersor multimers, how can it be demonstrated? Once identified,what can be done to determine which domains and residuesare involved in the interaction? How can the physical prop-erties, association constants and rate constants that controlthe interaction be determined? The following sections,Qualitative Techniques and Quantitative Techniques, illus-trate a route taken in the investigation of one metal bindinghomodimeric regulatory protein.

2. Experimental methods and results

2.1. Qualitative techniques

Both qualitative and quantitative methods are useful inthe study of dimerizing and oligomerizing systems. Qualita-tive investigations often lend themselves to the developmentof screening methods that enable the protein domainsinvolved and specific residues to be identified. The qualita-tive methods described below are based on size exclusion(gel filtration) chromatography, mass spectrometry andaffinity interactions. The last method lends itself to thedevelopment of simple screening methods.

2.2. Size exclusion

Size exclusion (gel filtration) methods can be used togather evidence for monomer–dimer-polymeric behaviorof isolated proteins. Classically, gel filtration is used as aconvenient way to estimate the apparent molecular weightof a protein. The volume through which the protein issieved is more appropriately related to the hydrodynamicradius of the protein, which depends on a number of fac-tors including the density, shape, and the molecular weight.It is not uncommon for the calculated molecular weight of

a protein to differ significantly from the apparent size(molecular weight) as estimated from gel filtration. Whilethe discrepancy may suggest that a dimer or multimer ispresent, in many cases the estimated weight lies betweenthat of a monomer and a dimer. Very often the apparentsize is rationalized by suggesting the protein has adoptedan elongated conformation that would migrate with ahigher apparent size due to the shape of the molecule. Aless investigated but very real possibility is that the pro-tein’s observed size is the result of the rapid equilibriumbetween the monomer and dimer. If a rapid equilibriumis responsible for the apparent size then changing the con-centration of protein applied to the gel filtration columnmay change its apparent size. Discrepancies in the pub-lished molecular weight of CopY and concentration depen-dent differences in apparent size during the course of itspurification suggested that the observed size was actuallythe averaged position of the monomer and dimer in rapidequilibrium.

Fig. 4 highlights the changes in apparent size of thefusion-protein GB1-Ymbs38 when fixed volumes of increas-ing concentration were resolved on a Shodex KW803 equil-ibrated in 50 mM Tris, pH 7.8, 150 mM NaCl, 0.05% b-ME, at a flow rate of 1 mL/min. A shift towards later elu-tion times is evident upon decrease of concentration ofGB1-Ymbs38. A similar analysis conducted on standardmonomeric proteins shows no change in apparent size.

2.3. Electrospray ionization mass spectroscopy

Electrospray ionization mass spectrometry (ESI-MS)has been employed relatively recently as a method to iden-tify oligomerization of proteins in solution. In 1993, themethod was used to detect non-covalently associated leu-cine zipper dimers [39]. Since then, ESI-MS has been usedto detect a multitude of other protein dimers [40–42],including the oxygenase domain of nitric oxide synthase[43]. Although the ionization process involves the transferof a protein from an aqueous to a gaseous phase, somedimeric species are able to survive the process [44–46].Because the near absence of solvent upon transfer to thegas phase weakens hydrophobic interactions, ESI-MS isnot a suitable method to determine the stability or the spe-cific binding energy of protein dimers [47]. Regardless, themethod is very useful for the straightforward identificationof homodimeric and higher order protein complexes.

The ability of non-covalent protein complexes to survivethe ionization process is aided by using conditions of lowtemperature and low electrospray voltage [39,41–44,47,48].Also important to the process is the difference between thecone and extractor voltages, referred to as the declusteringvoltage, DCS. The GB1-Ymbs38 protein was subjected toESI-MS with a cone voltage of 20 V and an extractor volt-age of 10 V (DCS = 10 V). A voltage of 2 kV at the tip ofthe inlet capillary was used to generate the electrospray.The temperature at the ion spray interface was kept at40 �C, while protein samples (approximately 250 lM) were



delivered at a rate of 50 lL/min in 10 mM ammonium ace-tate, pH 8.0. The resulting spectra demonstrate that the for-mation of non-covalently linked dimers is dependent on the

presence of the metal loaded Ymbs38 portion (Fig. 5A).GB1 alone (Fig. 5B) lacks evidence of dimerization. Thepresence of dimers is indicated by the appearance of weakersignals, all of which possess odd charges.

2.4. Affinity resin binding assay

Affinity purification of proteins is facilitated by the fusionof six histidine residues (termed ‘‘6 � histidine tag”) to oneterminus of the protein through molecular biology tech-niques. In the case of GB1-Ymbs38, the 6 � histidine tagis attached to the N-terminus. The histidine residues coordi-nate very strongly to the Ni2+ ions that are immobilized onan affinity resin. Other proteins in the mixture applied ontothe resin pass through without interacting with the affinityresin. The retained 6 � histidine tagged protein can beeluted by the introduction of a high concentration of imid-azole after sufficient washing to remove all of the non-inter-acting contaminating proteins. Protein–protein interactionscan be detected with this technique by mixing the 6 � histi-dine tagged protein with an untagged version of the sameprotein. If the untagged protein specifically interacts withthe 6 � histidine tagged protein, it will be retained on theaffinity resin, and subsequently co-elute with the 6 � histi-dine tagged protein upon application of imidazole [49].Co-elution is typically revealed by subjecting the eluant toSDS-PAGE (Fig. 6A–C). The assay has been successfullyused to detect homodimerization in the eukaryotic TATAbinding protein (TBP) involved in transcription ofnuclear-encoded genes [50,51], among others [41]. Detectionof association between unlike proteins has also been demon-strated [52]. An untagged GB1-Ymbs38 containing Zn(II) isretained on the affinity resin when mixed with a 6 � histidinetagged Zn(II) GB1-Ymbs38, providing another piece of evi-dence pointing toward the presence of dimers (Fig. 6D, lane3). The Ymbs38 portion must be present for dimerization tooccur, as the untagged GB1-Ymbs38 is not retained on theaffinity resin when mixed with a 6 � histidine tagged GB1(Fig. 6D, lane 10). It should be noted that the untagged pro-tein was prepared by thrombin digestion of the tagged pro-tein. Due to a second thrombin cleavage site present in thenative sequence of GB1, the untagged protein appears as adouble band on the electrophoresis gel.

2.5. Quantitative techniques

Frontal zone size exclusion chromatography, a tech-nique developed in the early 1960s [53], is an approachablemethod of determining association constants for proteinsin which the monomer and dimer are in rapid equilibrium.Analytical ultracentrifugation, frontal zone chromatogra-phy, and isothermal titration calorimetry (ITC) methods,when exercised fully, have the ability to provide associationconstants, free energies and other useful physical parame-ters. The apparent size being monitored by frontalzone chromatography, ultracentrifugation, dynamic lightscattering methods can be readily observed. In an ITC

Fig. 4. The effects of loading concentrations on the elution of monomericstandards and monomer–dimers in rapid equilibrium. Small zone injec-tions of non-associating protein standards versus the self-associating GB1-Ymbs38. Panel A: Protein standards – Blue dextran (2000 kDa), bovineserum albumin (66 kDa), and carbonic anhydrase (29 kDa). Varyingconcentrations of 1 mg/mL (solid line), 0.5 mg/mL (dashed line), and0.1 mg/mL (dotted line). Panel B: Protein standards – Ovalbumin(45 kDa) and RNase A (13.7 kDa). Varying concentrations of 1 mg/mL(solid line), 0.5 mg/mL (dashed line), and 0.1 mg/mL (dotted line). PanelC: GB1-Ymbs38 – Varying concentrations of 160 lM (solid line), 145 lM(dashed line), and 130 lM (dotted line). The elution volume of the proteinstandards remains constant as loading concentration is decreased. A shifttowards later elution times is evident upon decrease of concentration ofGB1-Ymbs38. All samples were run on the Shodex KW803 equilibrated in50 mM Tris, pH 7.8, 150 mM NaCl, 0.05% b-ME, at a flow rate of 1 mL/min.



experiment, the data can sometimes be described by morethan one model system and therefore requires that theexperimenter have a clearer understanding of the phenom-enon being measured. On the other hand, the ITC is anextremely powerful method that enables a number of phys-ical parameters to be determined in a single experiment.Analytical centrifugation, isothermal titration calorimetryand dynamic light scattering are typically not readily avail-able and the equipment is expensive. By contrast frontalzone methods are considerably more accessible and lessexpensive. They are, however, only useful for those pro-teins which have a rapid equilibrium between their mono-meric and dimeric forms. Frontal zone methods onlyrequire appropriate size exclusion media, a pump and asuitable detector for analysis, but do suffer from the needfor relatively large amounts of soluble protein.

2.6. Frontal zone chromatography

Analysis of association–dissociation equilibria in frontalzone gel filtration chromatography involves application of a

sufficiently large volume of a protein solution to a gel filtra-tion column to ensure that a plateau region is attained in theelution profile [53]. Self-associating proteins that are inrapid equilibrium between monomer and dimer forms exhi-bit elution profiles with defining characteristics, the mostrecognizable of which is non-symmetrical advancing andtrailing edges. A very sharp advancing edge occurs as higherconcentrated protein, making up the section of the advanc-ing boundary immediately before the plateau region, over-takes the less concentrated protein at the very front edgeof the sample that enters the column resin. Higher concen-trated protein has a proportionally larger amount ofdimeric species, and thus will travel faster through the col-umn resin than the less concentrated section at the leadingedge. The same occurrence causes a long trailing edge afterthe plateau region, as the less concentrated and thereforemore prominently monomeric protein migrates slowly. Inaddition to the unsymmetrical elution profile, self-associat-ing proteins exhibit progressively later elution volumes asthe concentration of loaded protein is decreased. The laterelution volume can be attributed to a higher fraction of

Fig. 5. ESI mass spectrum of Zn(II)GB1-Ymbs38 and of GB1. Panel A. Zn(II)GB1-Ymbs38, monomer ions are the most abundant, and their chargestates are indicated with italicized numbers. The direct identification of the homodimer complex is represented by the less intense signals indicated withbold numbers. The molecular mass derived from the +17 and +19 charged ions is 25,260 Da, while the mass derived from the +9 and +10 charged ions is12,625 Da (Calculated monomeric molecular mass of GB1-Ymbs38 = 12,760 Da.). The deviation of the mass derived from the spectrum from the actualmass is due to a calibration error. Panel B. GB1, monomeric charged ions are labeled with italicized numbers. The molecular mass derived from the +6 and+7 charge states is 8258 Da (Calculated molecular mass of GB1 with the 6 � histidine tag = 8387 Da.).



monomeric species in the diluted protein solution as a resultof mass action [54]. Monomer–dimer equilibrium constantscan be obtained from the analysis of elution volume datafrom a series of frontal zone chromatography trials [53–64].

2.7. Frontal zone analysis of GB1-Ymbs38

Of the four GB1-Ymbs38 variants prepared, only theprotein in which both the metal-binding and the hydropho-bic amino acids were removed showed loss of self-associa-tion by the frontal zone chromatography technique(Fig. 7). The sample shifts in size by only a minimalamount as the protein concentration changes, and at thehighest loading concentration, elutes at an apparent molec-ular weight of only 2 kDa higher than its calculated molec-ular weight of 12.8 kDa. For variants in which only one ofthe two contributing factors was removed, the binding ofmetal was shown to be the more critical factor in CopYdimerization. The mutation of the hydrophobic leucineand isoleucine residues to serines does not appear to affectthe self-association when Zn(II) is bound (Fig. 7). Despitethis observation, the hydrophobic residues do indeed play arole in the CopY–CopY interaction, perhaps providing sta-

Fig. 6. HIS-select resin binding assay. Panel A: A 6 � histidine tagged (6 � his) version of the protein is mixed with an untagged protein and incubated at37 �C for 1 h to allow for subunit (monomer) exchange. Panel B: The protein mixture is applied to the HIS-Select affinity resin. The histidine-tag facilitatesthe strong adhesion of the protein to the Ni2+ resin. Any untagged protein that has dimerized with tagged protein will also adhere to the resin, while anyexcess protein or untagged dimer will be washed through. Panel C: Imidazole is added to elute the histidine-tagged protein from the resin. The eluants areanalyzed by SDS-PAGE for the presence of the untagged protein, indicative of protein dimerization. Panel D: SDS-PAGE of affinity resin binding assay.The 6 � histidine tag was removed from the protein by cleavage with thrombin. Untagged GB1-Ymbs38 appears as a double band due to a secondthrombin cleavage site within the GB1 amino acid sequence. The Ymbs38 portion remains intact. Lane 1: Mixture of 6 � his tagged GB1-Ymbs38 withuntagged GB1-Ymbs38 before application to the affinity resin. Lane 2: Wash fraction. Lane 3: Eluant of tagged GB1-Ymbs38 with untagged GB1-Ymbs38 assay mixture. Lane 4: EZ run molecular weight standard (Fisher Scientific). Lane 5: Untagged GB1-Ymbs38 before application to the affinityresin. Lane 6: Wash fraction. Lane 7: Eluant of untagged GB1-Ymbs38. Lane 8: Mixture of 6 � his tagged GB1 with untagged GB1-Ymbs38 beforeapplication to the affinity resin. Lane 9: Wash fraction. Lane 10: Eluant of tagged GB1 with untagged GB1-Ymbs38 assay mixture.

Fig. 7. Concentration dependence of elution volume: frontal exclusionchromatography. GB1-Ymbs variants were loaded onto the KW 803HPLC column (Shodex) pre-equilibrated at 4 �C with 50 mM Tris,150 mM NaCl, 0.1% b-mercaptoethanol, pH 7.8. Zn(II)GB1-Ymbs (d),Zn(II)GB1-Ymbs Leu/Ile-tp-Ser mutant (s) and apo-GB1-Ymbs (N)retain the ability to dimerize, evidenced by the measurable shift inmolecular weight as the protein concentration changes. Apo-GB1-YmbsLeu/Ile-to-Ser mutant (4) exists as a monomer at all protein concentra-tions tested, with a maximum apparent molecular weight of 14.8 kDa atthe highest concentration.



bilization to the dimer. Zn(II)GB1-Ymbs38, Zn(II)GB1-Ymbs38 Leu/Ile-to-Ser mutant and apo-GB1-Ymbs38retain the ability to dimerize, evidenced by their measur-able shifts in molecular weight as the protein concentrationchanges. Apo-GB1-Ymbs38 Leu/Ile-to-Ser mutant exists asa monomer at all protein concentrations tested, with amaximum apparent molecular weight of 14.8 kDa at thehighest concentration.

2.8. Analysis of association constants

Quantitative data in the form of equilibrium constantscan be obtained from the elution volume data gathered infrontal zone chromatography experiments [54,55,57,61–64]. The analysis requires the assignment of magnitudesfor the elution volumes of monomer, VM, and of polymer(dimer), VP. The monomeric VM is assigned by extrapola-tion of the elution volumes measured for the mutated formof apo GB1-Ymbs38 to zero protein concentration. Thedimeric VP is determined by plotting the results of theZn(II)GB1-Ymbs38 according to the logarithmic trans-form of the law of mass action:

log½ð1� fMÞCt� ¼ logðnKaÞ þ n logðfMCtÞ ð1Þ

where fM is the fraction of monomer, Ct is the base molarprotein concentration, n is the reaction stoichiometry, andKa is the association constant. Understanding that the frac-tion of monomer can be calculated by the equation:

fM ¼ ðV W � V PÞ=ðV M � V PÞ ð2Þ

where VW is the observed elution volume of the sample,values of VP can be estimated, inserted into the aboveequation, and subsequently plotted according to Eq. (1).Attainment of the appropriate VP value is confirmed uponEq. (1) appearing as a linear plot with the expected reactionstoichiometry (2 for the GB1-Ymbs38 dimer) as the slopeof the line. The equilibrium association constant Ka canthen be determined by calculating the fM for each loadedprotein sample according to Eq. (2), plotting fM as a func-tion of protein concentration, then fitting the plotted datato the monomer–dimer model equation of:

fM ¼ ð�1þ ð1þ ð8KaCtÞÞ1=2Þ=ð4KaCtÞ ð3Þwhere Ka is the sole curve-fitting parameter (Fig. 8). Table 1displays typical values obtained for dissociation constants(Kd = 1/Ka) of the GB1-Ymbs38 variants by this method.

3. Discussion and conclusions

Based on the qualitative and quantitative methods, boththe cysteine hook motif (–CxCxxxxCxC–) and the adjacentaliphatic repeat (Fig. 3) portions of the C-terminal domainof CopY contribute to the dimerization, but which of theseconstitute the bulk of the specificity and stability is notclear and is still under investigation. Hypothetically, thecysteine hook motif is believed to constitute the bulk ofthe specificity and a large measure of the thermodynamicstability. The aliphatic repeat is believed to primarily con-tribute to the thermodynamic stability. To deconvolute themechanism of the protein’s dimerization, combinations ofbiophysical and molecular biology methods, qualitativeand quantitative, are being applied to the native protein,mutants of the native protein, and fusion proteins contain-ing portions of the CopY protein. The set of methods usedto attack the cysteine hook motif described above illus-trates one pathway towards conducting these types of stud-ies. These methods and other will be applied as theinvestigation continues. Ultimately, the intention is to fullyunderstand the dimerization so that the motif, whichappears in many other organisms and in a myriad of con-texts, can be exploited to identify dimerizing/regulatoryproteins and be used in biotechnology applications.

Acknowledgement

This work was supported by NIH 1 R15 DK067060-01A2 (CTD).

References

[1] J.D.S. Klemm, Stuart L. Crabtree, R. Gerald, Ann. Rev. Immunol.16 (1998) 569–592.

[2] B.O. Nagar, M. Ikura, J.M. Rini, Nature 380 (1996) 360–364.[3] D.J. Thiele, Mol. Cell. Biol. 8 (1988) 2745–2752.[4] C.F. Evans, D.R. Engelke, D.J. Thiele, Mol. Cell. Biol. 10 (1990)

426–429.

Fig. 8. Determination of equilibrium association constants. The weightfraction of monomer plotted as a function of the protein concentrationapplied to the chromatography column allows for determination of anassociation constant, Ka, upon fitting the data to a monomer–dimerstoichiometric model (solid line).

Table 1Dissociation constants of dimeric GB1-Ymbs38 variants

GB1-Ymbs38 variant Dissociation constant Kd (M)a

Zn(II) GB1-Ymbs38 4.4 (±0.2) � 10�5

Zn(II) GB1-Ymbs38 Ile/Leu mutant 6.3 (±0.6) � 10�5

Apo – GB1-Ymbs38 4.1 (±0.3) � 10�5

Apo – GB1-Ymbs38 Ile/Leu mutant 71.4 (±2.0) � 10�5

a Dissociation constants are calculated from the reciprocal of the asso-ciation constant, Ka, which is directly derived from the non-linear leastsquares best fit of the fraction of monomer versus protein concentrationdata (Fig. 8).



[5] E.B. Gralla, D.J. Thiele, P. Silar, J.S. Valentine, Proc. Natl. Acad.Sci. 88 (1991) 8558–8562.

[6] S. Labbe, Z. Zhu, D.J. Thiele, J. Biol. Chem. 272 (1997) 15951–15958.[7] Y. Yamaguchi-Iwai, M. Serpe, D. Haile, W. Yang, D.J. Kosman,

R.D. Klausner, A. Dancis, J. Biol. Chem. 272 (1997) 17711–17718.[8] Z.H. Lu, P. Cobine, C.T. Dameron, M. Solioz, J. Trace Elem. Exp.

Med. 12 (1999) 347–360.[9] M. Solioz, A. Odermatt, R. Krapf, FEBS Lett. 346 (1994) 44–47.

[10] M. Solioz, K.D. Bissig, Schweiz. Med. Wochenschr. 128 (1998) 1175–1180.

[11] H. Wunderli-Ye, M. Solioz, Adv. Exp. Med. Biol. 448 (1999) 255–264.[12] M. Solioz, J.V. Stoyanov, Macro and Trace Elements, in: M. Anke,

R. Muller, U. Schafer, M. Stoeppler (Eds.), How Cells Deal WithCopper: A Lesson from Bacteria, Schubert-Verlag, Leipzig, 2002, pp.485–492.

[13] M. Solioz, J.V. Stoyanov, FEMS Microbiol. Rev. 27 (2003) 183–195.[14] P.A. Cobine, C.E. Jones, W.A. Wickramasinghe, M. Solioz, C.T.

Dameron, in: E.J. Massaro (Ed.), Handbook of Copper Pharmacologyand Toxicology, Humana Press Inc., Totowa, NJ, 2002, pp. 177–188.

[15] R. Wimmer, C.T. Dameron, M. Solioz, in: E.J. Massaro (Ed.),Handbook of Copper Pharmacology and Toxicology, Humana PressInc., Totowa, NJ, 2002, pp. 527–542.

[16] P. Cobine, W.A. Wickramasinghe, M.D. Harrison, T. Weber, M.Solioz, C.T. Dameron, FEBS Lett. 445 (1999) 27–30.

[17] R. Wimmer, T. Herrmann, M. Solioz, K. Wuthrich, J. Biol. Chem. 99(1999) 22597–22603.

[18] A. Odermatt, H. Suter, R. Krapf, M. Solioz, J. Biol. Chem. 268(1993) 12775–12779.

[19] P.A. Cobine, C.E. Jones, C.T. Dameron, J. Inorg. Biochem. 88 (2002)192–196.

[20] P. Gregory, R. Lewis, S. Curnock, K. Dyke, Mol. Micro. 24 (1997)1025–1037.

[21] D. Strausak, M. Solioz, J. Biol. Chem. 272 (1997) 8932–8936.[22] R. Portmann, K.R. Poulsen, R. Wimmer, M. Solioz, BioMetals 19

(2006) 61–70.[23] R. Garcia-Castellanos, A. Marrero, G. Mallorqui-Fernandez, J.

Potempa, M. Coll, F.X. Gomis-Rueth, J. Biol. Chem. 278 (2003)39897–39905.

[24] R. Garcia-Castellanos, G. Mallorqui-Fernandez, A. Marrero, J.Potempa, M. Coll, F.X. Gomis-Rueth, J. Biol. Chem. 279 (2004)17888–17896.

[25] V. Melckebeke Helene, C. Vreuls, P. Gans, P. Filee, G. Llabres, B.Joris, J.-P. Simorre, J. Mol. Biol. 333 (2003) 711–720.

[26] J.E. Anderson, M. Ptashne, S.C. Harrison, Nature 326 (1987) 846–852.

[27] H. Wunderli-Ye, M. Solioz, Biochem. Biophys. Res. Commun. 259(1999) 443–449.

[28] P. Cobine, G.N. George, C.E. Jones, W.A. Wickranasinghe, M.Solioz, C.T. Dameron, Biochemistry 41 (2002) 5822–5829.

[29] M.S. Solioz, V. Jivko, FEMS Microbiol. Rev. 27 (2003) 183–195.[30] Z.H. Lu, C.T. Dameron, M. Solioz, Biometals 16 (2003) 137–143.[31] L.J. McGuffin, K. Bryson, D.T. Jones, Bioinformatics 16 (2000) 404–

405.[32] D.T. Jones, J. Mol. Biol. 292 (1999) 195–202.[33] A.M. Gronenborn, D.R. Filpula, N.Z. Essig, A. Achari, M. Whitlow,

P.T. Wingfield, G.M. Clore, Science 253 (1991) 657–661.[34] J.J. Barchi, B. Grasberger, A.M. Gronenborn, G.M. Clore, Prot. Sci.

3 (1994) 15–21.[35] D. Idiyatullin, I. Nesmelova, V.A. Daragan, K.H. Mayo, Prot. Sci. 12

(2003) 914–922.

[36] M.K. Frank, F. Dyda, A. Dobrodumov, A.M. Gronenborn, Nat.Struct. Biol. 9 (2002) 877–885.

[37] I.L. Byeon, J.M. Louis, A.M. Gronenborn, J. Mol. Biol. 333 (2003)141–152.

[38] A.M. Gronenborn, G.M. Clore, J. Mol. Biol. 233 (1993) 331–335.[39] Y.-T. Li, Y.-L. Hsieh, J.D. Henion, M.W. Senko, F.W. McLafferty,

B. Ganem, J. Am. Chem. Soc. 115 (1993) 8409–8413.[40] M. Przybylski, C. Maier, K. Hagele, E. Bauer, E. Hanappel, R. Nave,

U. Kruger, K.P. Schafer, in: R.S. Hodges, J.A. Smith (Ed.), PeptideChemistry, Structure and Biology, Primary Structure Elucidation,Surfactant Function and Specific Formation of SupramolecularDimer Structures of Lung Surfactant Associated SP-C Proteins,1994, pp. 338–339.

[41] S. Witte, F. Neumann, U. Krawinkel, M. Przybylski, J. Biol. Chem.271 (1996) 18171–18175.

[42] J.A.T. Hornby, S.G. Condreanu, R.N. Armstrong, H.W. Dirr,Biochemistry 41 (2002) 14238–14247.

[43] J.C. Smith, M. Siu, S.P. Rafferty, J. Am. Soc. Mass. Spectrom. 15(2004) 629–638.

[44] H. Wendt, E. Durr, R.M. Thomas, M. Przybylski, H.R. Bosshard,Prot. Sci. 4 (1995) 1563–1570.

[45] P. Kebarle, L. Tang, Anal. Chem. 65 (1993) 3654–3668.[46] R.D. Smith, K.J. Light-Wahl, Biol. Mass. Spectrom. 22 (1993) 493–

501.[47] J.M. Daniel, S.D. Friess, S. Rajagopalan, S. Wendt, R. Zenobi, Int. J.

Mass. Spectrom. 216 (2002) 1–27.[48] G.L. Glish, R.W. Vachet, Nat. Rev. Drug. Discov. 2 (2003) 140–150.[49] T. Lu, M. Van Dyke, M. Sawadogo, Anal. Biochem. 213 (1993) 318–

322.[50] R.A. Coleman, A.K.P. Taggart, L.R. Benjamin, B.F. Pugh, J. Biol.

Chem. 270 (1995) 13842–13849.[51] C.A. Adams, S.R. Kar, J.E. Hopper, M.G. Fried, J. Biol. Chem. 279

(2004) 1376–1382.[52] F. Giorgino, O. de Robertis, L. Laviola, C. Montrone, S. Perrini,

K.C. McCowen, R.J. Smith, Proc. Natl. Acad. Sci. 97 (2000) 1125–1130.

[53] D.J. Winzor, H.A. Scheraga, Biochemistry 2 (1963) 1263–1267.[54] P.J. Darling, J.M. Holt, G.K. Ackers, Biochemistry 39 (2000) 11500–

11507.[55] W.E. Royer Jr., R.A. Fox, F.R. Smith, D. Zhu, E.H. Braswell, J.

Biol. Chem. 272 (1997) 5689–5694.[56] D.J. Winzor, J. Biochem. Biophys. Meth. 56 (2003) 15–52.[57] E. Nenortas, D. Beckett, Anal. Biochem. 222 (1994) 366–373.[58] D.J. Winzor, H.A. Scheraga, J. Phys. Chem. 68 (1964) 338–343.[59] R. Valdes Jr., G.K. Ackers, Meth. Enzymol. 61 (1979) 125–142.[60] G.K. Ackers, T.E. Thompson, Proc. Natl. Acad. Sci. 53 (1965) 342–

349.[61] R. Valdes Jr., G.K. Ackers, J. Biol. Chem. 252 (1977) 74–81.[62] D. Beckett, K.S. Koblan, G.K. Ackers, Anal. Biochem. 196 (1991)

69–75.[63] A.C. Drohat, E. Nenortas, D. Beckett, D.J. Weber, Prot. Sci. 6 (1997)

1577–1582.[64] G.F. Sanchez-Perez, M. Gasset, J.J. Calvette, M.A. Pajares, J. Biol.

Chem. 278 (2003) 7285–7293.[65] G. Multhaup, D. Strausak, K.D. Bissig, M. Solioz, Biochem.

Biophys. Res. Commun. 288 (2001) 172–177.[66] A. Odermatt, R. Krapf, M. Solioz, Biochem. Biophys. Res. Commun.

202 (1994) 44–48.[67] B.G. Fujitsu Computer Systems, Fujitsu Computer Systems, Beaver-

ton, OR, 2006.


Author's personal copy - Duquesne UniversityProtein dimerization is essential for cellular processes...

Documents

Transcript of Author's personal copy - Duquesne UniversityProtein dimerization is essential for cellular processes...