PocketMatch (version 2.0): A parallel algorithm for the detection of structural similarities between protein ligand binding-sites

Knowledge of protein-ligand interactions is essential to understand several biological processes and important for applications ranging from understanding protein function to drug discovery and protein engineering. Here, we describe an algorithm for the comparison of three-dimensional ligand-binding sites in protein structures. A previously described algorithm, PocketMatch (version 1.0) is optimised, expanded, and MPI-enabled for parallel execution. PocketMatch (version 2.0) rapidly quantifies binding-site similarity based on structural descriptors such as residue nature and interatomic distances. Atomic-scale alignments may also be obtained from amino acid residue pairings generated. It allows an end-user to compute database-wide, all-to-all comparisons in a matter of hours. The use of our algorithm on a sample dataset, performance-analysis, and annotated source code is also included.

Molecular Dissection of Mycobacterium tuberculosis Integration Host Factor Reveals Novel Insights into the Mode of DNA Binding and Nucleoid Compaction

Background: mIHF belongs to a subfamily of proteins, distinct from E. coli IHF. Results: Functionally important amino acids of mIHF and the mechanism(s) underlying DNA binding, DNA bending, and site-specific recombination are distinct from that of E. coli IHF. Conclusion: mIHF functions could contribute beyond nucleoid compaction. Significance: Because mIHF is essential for growth, the molecular mechanisms identified here can be exploited in drug screening efforts. The annotated whole-genome sequence of Mycobacterium tuberculosis revealed that Rv1388 (Mtihf) is likely to encode for a putative 20-kDa integration host factor (mIHF). However, very little is known about the functional properties of mIHF or the organization of the mycobacterial nucleoid. Molecular modeling of the mIHF three-dimensional structure, based on the cocrystal structure of Streptomyces coelicolor IHF duplex DNA, a bona fide relative of mIHF, revealed the presence of Arg-170, Arg-171, and Arg-173, which might be involved in DNA binding, and a conserved proline (Pro-150) in the tight turn. The phenotypic sensitivity of Escherichia coli ihfA and ihfB strains to UV and methyl methanesulfonate could be complemented with the wild-type Mtihf but not its alleles bearing mutations in the DNA-binding residues. Protein-DNA interaction assays revealed that wild-type mIHF, but not its DNA-binding variants, binds with high affinity to fragments containing attB and attP sites and curved DNA. Strikingly, the functionally important amino acid residues of mIHF and the mechanism(s) underlying its binding to DNA, DNA bending, and site-specific recombination are fundamentally different from that of E. coli IHF. Furthermore, we reveal novel insights into IHF-mediated DNA compaction depending on the placement of its preferred binding sites; mIHF promotes DNA compaction into nucleoid-like or higher order filamentous structures. We therefore propose that mIHF is a distinct member of a subfamily of proteins that serve as essential cofactors in site-specific recombination and nucleoid organization and that these findings represent a significant advance in our understanding of the role(s) of nucleoid-associated proteins.

Molecular Dissection of Mycobacterium tuberculosis Integration Host Factor Reveals Novel Insights into the Mode of DNA Binding and Nucleoid Compaction

The annotated whole-genome sequence of Mycobacterium tuberculosis indicated that Rv1388 (Mtihf) likely encodes a putative 20 kDa integration host factor (mIHF). However, very little is known about the functional properties of mIHF or organization of mycobacterial nucleoid. Molecular modeling of the mIHF three-dimensional structure, based on the cocrystal structure of Streptomyces coelicolor IHF-duplex DNA, a bona fide relative of mIHF, revealed the presence of Arg170, Arg171, and Arg173, which might be involved in DNA binding, and a conserved proline (P150) in the tight turn. The phenotypic sensitivity of Escherichia coli Delta ihfA and Delta ihfB strains to UV and methylmethanesulfonate could be complemented with the wild-type Mtihf, but not its alleles bearing mutations in the DNA-binding residues. Protein DNA interaction assays revealed that wild-type mIHF, but not its DNA-binding variants, bind with high affinity to fragments containing attB and attP sites and curved DNA. Strikingly, the functionally important amino acid residues of mIHF and the mechanism(s) underlying its binding to DNA, DNA bending, and site-specific recombination are fundamentally different from that of E. coli IHF alpha beta. Furthermore, we reveal novel insights into IHF-mediated DNA compaction depending on the placement of its preferred binding sites; mIHF promotes compaction of DNA into nucleoid-like or higher-order filamentous structures. We hence propose that mIHF is a distinct member of a subfamily of proteins that serve as essential cofactors in site-specific recombination and nucleoid organization and that these findings represent a significant advance in our understanding of the role(s) of nucleoid-associated proteins.

DNA Elasticity from Short DNA to Nucleosomal DNA

Active biological processes like transcription, replication, recombination, DNA repair, and DNA packaging encounter bent DNA. Machineries associated with these processes interact with the DNA at short length (<100 base pair) scale. Thus, the study of elasticity of DNA at such length scale is very important. We use fully atomistic molecular dynamics (MD) simulations along with various theoretical methods to determine elastic properties of dsDNA of different lengths and base sequences. We also study DNA elasticity in nucleosome core particle (NCP) both in the presence and the absence of salt. We determine stretch modulus and persistence length of short dsDNA and nucleosomal DNA from contour length distribution and bend angle distribution, respectively. For short dsDNA, we find that stretch modulus increases with ionic strength while persistence length decreases. Calculated values of stretch modulus and persistence length for DNA are in quantitative agreement with available experimental data. The trend is opposite for NCP DNA. We find that the presence of histone core makes the DNA stiffer and thus making the persistence length 3-4 times higher than the bare DNA. Similarly, we also find an increase in the stretch modulus for the NCP DNA. Our study for the first time reports the elastic properties of DNA when it is wrapped around the histone core in NCP. We further show that the WLC model is inadequate to describe DNA elasticity at short length scale. Our results provide a deeper understanding of DNA mechanics and the methods are applicable to most protein-DNA complexes.

Molecular mechanisms of interleukin-2 gene inducibility: developmental control and combinatorial action of transcription factors

Interleukin-2 is one of the lymphokines secreted by T helper type 1 cells upon activation mediated by T-cell receptor (TCR) and accessory molecules. The ability to express IL-2 is correlated with T-lineage commitment and is regulated during T cell development and differentiation. Understanding the molecular mechanism of how IL-2 gene inducibility is controlled at each transition and each differentiation process of T-cell development is to understand one aspect of T-cell development. In the present study, we first attempted to elucidate the molecular basis for the developmental changes of IL-2 gene inducibility. We showed that IL-2 gene inducibility is acquired early in immature CD4- CD8-TCR- thymocytes prior to TCR gene rearrangement. Similar to mature T cells, a complete set of transcription factors can be induced at this early stage to activate IL-2 gene expression. The progression of these cells to cortical CD4^+CD8^+TCR^(1o) cells is accompanied by the loss of IL-2 gene inducibility. We demonstrated that DNA binding activities of two transcription factors AP-1 and NF-AT are reduced in cells at this stage. Further, the loss of factor binding, especially AP-1, is attributable to the reduced ability to activate expression of three potential components of AP-1 and NF-AT, including c-Fos, FosB, and Fra-2. We next examined the interaction of transcription factors and the IL-2 promoter in vivo by using the EL4 T cell line and two non-T cell lines. We showed an all-or-none phenomenon regarding the factor-DNA interaction, i.e., in activated T cells, the IL-2 promoter is occupied by sequence-specific transcription factors when all the transcription factors are available; in resting T cells or non-T cells, no specific protein-DNA interaction is observed when only a subset of factors are present in the nuclei. Purposefully reducing a particular set of factor binding activities in stimulated T cells using pharmacological agents cyclosporin A or forskolin also abolished all interactions. The results suggest that a combinatorial and coordinated protein-DNA interaction is required for IL-2 gene activation. The thymocyte experiments clearly illustrated that multiple transcription factors are regulated during intrathymic T-cell development, and this regulation in tum controls the inducibility of the lineage-specific IL-2 gene. The in vivo study of protein-DNA interaction stressed the combinatorial action of transcription factors to stably occupy the IL-2 promoter and to initiate its transcription, and provided a molecular mechanism for changes in IL-2 gene inducibility in T cells undergoing integration of multiple environmental signals.

Design, synthesis and characterization of sequence-specific DNA-cleaving metalloproteins

This thesis describes research pursued in two areas, both involving the design and synthesis of sequence specific DNA-cleaving proteins. The first involves the use of sequence-specific DNA-cleaving metalloproteins to probe the structure of a protein-DNA complex, and the second seeks to develop cleaving moieties capable of DNA cleavage through the generation of a non-diffusible oxidant under physiological conditions.

Chapter One provides a brief review of the literature concerning sequence-specific DNA-binding proteins. Chapter Two summarizes the results of affinity cleaving experiments using leucine zipper-basic region (bZip) DNA-binding proteins. Specifically, the NH_2-terminal locations of a dimer containing the DNA binding domain of the yeast transcriptional activator GCN4 were mapped on the binding sites 5'-CTGACTAAT-3' and 5'ATGACTCTT- 3' using affinity cleaving. Analysis of the DNA cleavage patterns from Fe•EDTA-GCN4(222-281) and (226-281) dimers reveals that the NH_2-termini are in the major groove nine to ten base pairs apart and symmetrically displaced four to five base pairs from the central C of the recognition site. These data are consistent with structural models put forward for this class of DNA binding proteins. The results of these experiments are evaluated in light of the recently published crystal structure for the GCN4-DNA complex. Preliminary investigations of affinity cleaving proteins based on the DNA-binding domains of the bZip proteins Jun and Fos are also described.

Chapter Three describes experiments demonstrating the simultaneous binding of GCN4(226-281) and 1-Methylimidazole-2-carboxamide-netropsin (2-ImN), a designed synthetic peptide which binds in the minor groove of DNA at 5'-TGACT-3' sites as an antiparallel, side-by-side dimer. Through the use of Fe•EDTA-GCN4(226-281) as a sequence-specific footprinting agent, it is shown that the dimeric protein GCN4(226-281) and the dimeric peptide 2- ImN can simultaneously occupy their common binding site in the major and minor grooves of DNA, respectively. The association constants for 2-ImN in the presence and in the absence of Fe•EDTA-GCN4(226-281) are found to be similar, suggesting that the binding of the two dimers is not cooperative.

Chapter Four describes the synthesis and characterization of PBA-β-OH-His- Hin(139-190), a hybrid protein containing the DNA-binding domain of Hin recombinase and the putative iron-binding and oxygen-activating domain of the antitumor antibiotic bleomycin. This 54-residue protein, comprising residues 139-190 of Hin recombinase with the dipeptide pyrimidoblamic acid-β-hydroxy-L-histidine (PBA-β-OH-His) at the NH2 terminus, was synthesized by solid phase methods. PBA-β-OH-His-Hin(139- 190) binds specifically to DNA at four distinct Hin binding sites with affinities comparable to those of the unmodified Hin(139-190). In the presence of dithiothreitol (DTT), Fe•PB-β-OH-His-Hin(139-190) cleaves DNA with specificity remarkably similar to that of Fe•EDTA-Hin(139-190), although with lower efficiency. Analysis of the cleavage pattern suggests that DNA cleavage is mediated through a diffusible species, in contrast with cleavage by bleomycin, which occurs through a non-diffusible oxidant.

噬菌体434单链阻遏蛋白的高亲和力DNA结合序列的设计和筛选

一． 设计和筛选单链阻遏蛋白的高亲和力DNA结合序列 　　单链阻遏蛋白RRTRES是噬菌体434阻遏蛋白的衍生物，它是噬菌体434阻 遏蛋白的N端DBD(1-69位氨基酸)组成的共价二聚体。这个单链分子有两个DBD，一个是野生型噬菌体434的DBD-R，另一个是突变了的DBD - RTRES，二者用重组接头以头接尾的方式连接起来。在RTRES的α3-螺旋中．1、1、2、5位氨基酸与DNA识别紧密相关，它们分别为T、R、E、S。为了筛选出突变的RTRES的DNA结合位点，设计了核心序列为CATACAAGAAAGNNNNNNTTTATG随机DNA库，通过RRTRES与随机DNA库的体外结合和循环筛选。将筛选到的群体克隆并测序。通过与单链阻遏蛋白RRTRES的亲和力测定，对每一个筛选到的序列进行特性分析。结果表明，当结合位点（上述划线部分）为TTAC或TTCC时为最适操纵区序列。它们与单链阻遏蛋白RRTRES的亲和力很高，Kd值在1-10pM的范围。其中随机部分为TTTACG的操纵区与RRTRES的亲和力最高，Kd值约为lpM;当结合位点为TTAC时，平均Kd值为3pM：当结合位点为 TTCC时，Kd值在5-lOpM之间。天然噬菌体434阻遏蛋白与其操纵区的亲和力的Kd值在nM数量级，与之相比，所筛选操纵区的亲和力明显提高。此外，亲和力大小还受到结合位点两侧的碱基的影响，特别是5'位碱基的影响。 　　表达纯化同源双突变的单链阻遏蛋白RTRESRTRE'根据RTRES的以上识别特一点，设计了一系列新的操纵区序列，它们的共有序列为GTAAGAAARNTTACN，或GGAAGAAARNTTCCN，并测定它们与RTRES RTRE之间的结合特异性。结果表明，它们可被RTRES RTRES特异识别，且亲和力也很高，Kd值在5-40pM之间。其中GTAAGAAAGTTTACG与RTRES RTRES之间结合的Kd值约为5pM。同样，表达了异源双突变的单链阻遏蛋白R*RTRES，然而它与 设计的相关操纵区的亲和力并不很高，Kd值约为lOOpM。利用本工作中的随机筛选和合理设计的原则，得到了新的具有特异性识别和高亲和力的蛋白一DNA相互作用。这个方法可望用于其他DBP的新的结合特异性的筛选。 二． 非同位素的方法筛选单链阻遏蛋白的最佳DNA结合序列初探 　　克隆和表达了带半胱氨酸尾的单链阻遏蛋白，利用已包被了马来酰胺的活性板可以与自由巯基结合的特性，将蛋白固定在活性板表面。体外筛选RTRES RTRES的最佳DNA结合序列，得到了一些与RTRES RTRES结合的序列，但Kd值nM数量级。此方法需进一步优化。

Electrochemiluminescent detection based on solid-phase extraction at tris(2,2 '-bipyridyl)ruthenium(II)-modified ceramic carbon electrode

A sensitive electrochemiluminescent detection scheme by solid-phase extraction at Ru(bpy)(3)(2+)-modified ceramic carbon electrodes (CCEs) was developed. The as-prepared Ru(bpy)(3)(2+)-modified CCEs show much better long-term stability than other Nafion-based Ru(bpy)(3)(2+)-modified electrodes and enjoy the inherent advantages of CCEs. The log-log calibration plot for dioxopromethazine is linear from 1.0 x 10(-9) to 1.0 x 10(-4) mol L-1 using the new detection scheme. The detection limit is 6.6 x 10(-10) mol L-1 at a signal-to-noise ratio of 3. The new scheme improves the sensitivity by similar to 3 orders of magnitude, which is the most sensitive Ru(bpy)(3)(2+) ECL method. The scheme allows the detection of dioxopromethazine in a urine sample within 3 min. Since Ru(bpy)(3)(2+) ECL is a powerful technique for determination of numerous amine-containing substances, the new detection scheme holds great promise in measurement of free concentrations, investigation of protein-drug interactions and DNA-drug interactions, pharmaceutical analysis, and so on.

A binding site for the cyclic adenosine 3',5'-monophosphate-response element-binding protein as a regulatory element in the grp78 promoter.

The 78-kDa glucose-regulated protein (GRP78) is ubiquitously expressed in many cell types. Its promoter contains multiple protein-binding sites and functional elements. In this study we examined a high affinity protein-binding site spanning bp -198 to -180 of the rat grp78 promoter, using nuclear extracts from both B-lymphoid and HeLa cells. This region contains a sequence TGACGTGA which, with the exception of one base, is identical to the cAMP-response element (CRE). Site-directed mutagenesis reveals that this sequence functions as a major basal level regulatory element in hamster fibroblast cells and is also necessary to maintain high promoter activity under stress-induced conditions. By gel mobility shift analysis, we detect two specific protein complexes. The major specific complex I, while immunologically distinct from the 42-kDa CRE-binding protein (CREB), binds most strongly to the grp site, but also exhibits affinity for the CRE consensus sequence. As such, complex I may consist of other members of the CREB/activating transcription factor protein family. The minor specific complex II consists of CREB or a protein antigenically related to it. A nonspecific complex III consists of the Ku autoantigen, an abundant 70- to 80-kDa protein complex in HeLa nuclear extracts. By cotransfection experiments, we demonstrate that in F9 teratocarcinoma cells, the grp78 promoter can be transactivated by the phosphorylated CREB or when the CREB-transfected cells are treated with the calcium ionophore A23187. The differential regulation of the grp78 gene by cAMP in specific cell types and tissues is discussed.

Thermodynamic analysis of ligand-induced changes in protein thermal unfolding applied to high-throughput determination of ligand affinities with extrinsic fluorescent dyes.

The quantification of protein-ligand interactions is essential for systems biology, drug discovery, and bioengineering. Ligand-induced changes in protein thermal stability provide a general, quantifiable signature of binding and may be monitored with dyes such as Sypro Orange (SO), which increase their fluorescence emission intensities upon interaction with the unfolded protein. This method is an experimentally straightforward, economical, and high-throughput approach for observing thermal melts using commonly available real-time polymerase chain reaction instrumentation. However, quantitative analysis requires careful consideration of the dye-mediated reporting mechanism and the underlying thermodynamic model. We determine affinity constants by analysis of ligand-mediated shifts in melting-temperature midpoint values. Ligand affinity is determined in a ligand titration series from shifts in free energies of stability at a common reference temperature. Thermodynamic parameters are obtained by fitting the inverse first derivative of the experimental signal reporting on thermal denaturation with equations that incorporate linear or nonlinear baseline models. We apply these methods to fit protein melts monitored with SO that exhibit prominent nonlinear post-transition baselines. SO can perturb the equilibria on which it is reporting. We analyze cases in which the ligand binds to both the native and denatured state or to the native state only and cases in which protein:ligand stoichiometry needs to treated explicitly.

Molecular Dynamics of HIV1-Integrase in Complex with 93del - A Structural Perspective on the Mechanism of Inhibition

HIV1 integrase is an important target for the antiviral therapy. Guanine-rich quadruplex, such as 93del, have been shown to be potent inhibitors of this enzyme and thus representing a new class of antiviral agents. Although X-ray and NMR structures of HIV1 integrase and 93del have been reported, there is no structural information of the complex and the mechanism of inhibition still remains unexplored. A number of computational methods including automated protein-DNA docking and molecular dynamics simulation in explicit solvent were used to model the binding of 93del to HIV1 integrase. Analysis of the dynamic behaviour of the complex using principal components analysis and elastic network modelling techniques allow us to understand how the binding of 93del aptamer and its interactions with key residues affect the intrinsic motions of the catalytic loops by stabilising them in catalytically inactive conformations. Such insights into the structural mechanism of inhibition can aid in improving the design of anti-HIV aptamers.

Computational design with flexible backbone sampling for protein remodeling and scaffolding of complex binding sites

Dissertation presented to obtain the Doutoramento (Ph.D.) degree in Biochemistry at the Instituto de Tecnologia Qu mica e Biol ogica da Universidade Nova de Lisboa

Peroxisome proliferator-activated receptor structures: ligand specificity, molecular switch and interactions with regulators.

Peroxisome proliferator-activated receptors (PPARs) compose a family of nuclear receptors that mediate the effects of lipidic ligands at the transcriptional level. In this review, we highlight advances in the understanding of the PPAR ligand binding domain (LBD) structure at the atomic level. The overall structure of PPARs LBD is described, and important protein ligand interactions are presented. Structure-activity relationships between isotypes structures and ligand specificity are addressed. It is shown that the numerous experimental three-dimensional structures available, together with in silico simulations, help understanding the role played by the activating function-2 (AF-2) in PPARs activation and its underlying molecular mechanism. The relation between the PPARs constitutive activity and the intrinsic stability of the active conformation is discussed. Finally, the interactions of PPARs LBD with co-activators or co-repressors, as well as with the retinoid X receptor (RXR) are described and considered in relation to PPARs activation.

GREBP, un nouveau facteur de transcription contrôlant l’expression de la guanylate cyclase A, récepteur de l’ANP, via l’élément de réponse au cGMP

La découverte du système des peptides natriurétiques (NP), au début des années 80, fut une découverte majeure qui révéla le rôle endocrinien du cœur. Les connaissances sur la relaxation vasculaire, la diurèse et la natriurèse provoquées par ce système ont évolué vers un niveau de complexité insoupçonné à cette époque. Nous savons à présent que les NP sont impliqués dans plusieurs autres mécanismes dont la prolifération cellulaire, l’apoptose, l’inhibition du système rénine-angiotensine-aldostérone (RAAS) et le métabolisme des adipocytes. Le métabolisme des lipides est maintenant devenu une cible de choix dans la lutte contre l’obésité. Cette condition aux proportions pandémiques est un facteur de risque majeur dans l’apparition de l’hypertension et du syndrome métabolique (MetS). La compréhension des mécanismes et des défauts de la voie des NP pourrait avoir un impact positif sur le contrôle du MetS et de l’hypertension. L’expression du récepteur des peptides natriuretiques de type 1 (NPR1/GCA) est contrôlée par plusieurs agents incluant son propre ligand, le peptide natriurétique de l’oreillette (ANP). La découverte d’une boucle de retro-inhibition, dans les années 90, a été un événement majeur dans le domaine des NP. En effet, suite à une stimulation à l’ANP, le NPR1/GCA peut inhiber l’activité transcriptionnelle de son propre gène par un mécanisme dépendant du cGMP. Notre groupe a identifié un élément cis-régulateur responsable de cette sensibilité au cGMP et mon projet consistait à identifier la ou les protéine(s) liant cet élément de réponse au cGMP (cGMP-RE). Nous avons identifié un clone liant le cGMP-RE en utilisant la technique du simple hybride chez la levure et une banque d’ADN complémentaire (ADNc) de rein humain. Ce clone provient d’un ADNc de 1083-bp dont le gène est localisé sur le chromosome 1 humain (1p33.36) et codant pour une protéine dont la fonction était inconnue jusqu’ici. Nous avons nommé cette nouvelle protéine GREBP en raison de sa fonction de cGMP Response Element Binding Protein. Des essais de liaison à l’ADN ont montré que cette protéine possède une affinité 18 fois plus élevée pour le cGMP-RE que le contrôle, tandis que des expériences de retard sur gel (EMSA) ont confirmé la spécificité des interactions protéine-ADN. De plus, l’immuno-précipitation de la chromatine (ChIP) a prouvé que GREBP lie le cGMP-RE dans des conditions physiologiques. La liaison de GREBP au cGMP-RE inhibe l’expression du gène rapporteur luciférase sous contrôle du promoteur de npr1/gca. L’inhibition de GREBP à l’aide d’ARN interférant active le promoteur de npr1/gca. Dans les cellules NCI-H295R, l’ANP stimule l’expression de grebp de 60% après seulement 3 heures et inhibe l’expression de npr1/gca de 30%. GREBP est une protéine nucléaire surtout exprimée dans le cœur et ayant le facteur eIF3F comme partenaire. Les variations nucléotidiques du gène sont plus fréquentes chez les patients hypertendus que chez des patients normotendus ou hypertendus souffrant de MetS. Nous rapportons ici l’existence d’un gène spécifique à l’humain qui agit comme répresseur transcriptionnel de npr1/gca et potentiellement impliqué dans le développement de l’hypertension.

An insight into the mechanism of protein separation by colloidal gas aphrons (CGA) generated from ionic surfactants

Colloidal gas aphrons (CGA), which are surfactant stabilised microbubbles, have been previously applied for the recovery of proteins from model mixtures and a few studies have demonstrated the potential of these dispersions for the selective recovery of proteins from complex mixtures. However there is a lack of understanding of the mechanism of separation and forces governing the selectivity of the separation. In this paper a mechanistic study is carried out to determine the main factors and forces influencing the selectivity of separation of whey proteins with CGA generated from ionic surfactants. Two different separation strategies were followed: (i) separation of lactoferrin and lactoperoxidase by anionic CGA generated from a solution of sodium bis-(2-ethyl hexyl) sulfosuccinate (AOT); (ii) separation of beta-lactoglobulin by cationic CGA generated from a solution of cetyltrimethylammonium bromide (CTAB). Separation results indicate that electrostatic interactions are the main forces determining the selectivity however these could not completely explain the selectivities obtained following both strategies. Protein-surfactant interactions were studied by measuring the zeta potential changes on individual proteins upon addition of surfactant and at varying pH. Interestingly strongest electrostatic interactions were measured at those pH and surfactant to protein mass ratios which were optimum for protein separation. Effect of surfactant on protein conformation was determined by measuring the change in fluorescence intensity upon addition of surfactant at varying pH. Differences in the fluorescence patterns were detected among proteins which were correlated to differences in their conformational features which could in turn explain their different separation behaviour. The effect of conformation on selectivity was further proven by experiments in which conformational changes were induced by pre-treatment of whey (heating) and by storage at 4 degrees C. Overall it can be concluded that separation of proteins by ionic CGA is driven mainly by electrostatic interactions however conformational features will finally determine the selectivity of the separation with competitive adsorption having also an effect. (c) 2006 Elsevier B.V. All rights reserved.