The development of chromatography for the characterization of protein interactions: a personal perspective.

This article reviews the progress of a personal endeavour to develop chromatography as a quantitative procedure for the determination of reaction stoichiometries and equilibrium constants governing protein interactions. As well as affording insight into an aspect of chromatography with which many protein chemists are unfamiliar, it shows the way in which minor adaptations of conventional chromatographic practices have rendered the technique one of the most powerful methods available for the characterization of interactions. That pathway towards quantification is followed from the introduction of frontal gel filtration for the study of protein self-association to the characterization of ligand binding by the biosensor variant of quantitative affinity chromatography.

A potential role for the analytical ultracentrifuge in the experimental measurement of protein valence.

A method is described whereby sedimentation velocity is combined with equilibrium dialysis to determine the net charge (valence) of a protein by using chromate as an indicator ion for assessing the extent of the Donnan redistribution of small ions. The procedure has been used in experiments on bovine serum albumin under slightly alkaline conditions (pH 8.0, I 0.05) to illustrate its application to a system in which the indicator ion and protein both bear net negative charge and on lysozyme under slightly acidic conditions (pH 5.0, I 0.10) to illustrate the situation where chromate is a counterion.

Quantitative affinity chromatography.

The objective of this review is to summarize developments in the use of quantitative affinity chromatography to determine equilibrium constants for solute interactions of biological interest. Affinity chromatography is an extremely versatile method for characterizing interactions between dissimilar reactants because the biospecificity incorporated into the design of the affinity matrix ensures applicability of the method regardless of the relative sizes of the two reacting solutes. Adoption of different experimental strategies, such as column chromatography, simple partition equilibrium experiments, solid-phase immunoassay, and biosensor technology, has led to a situation whereby affinity chromatography affords a means of characterizing interactions governed by an extremely broad range of binding affinities--relatively weak interactions (binding constants below 10(3) M(-1)) through to interactions with binding constants in excess of 10(9) M(-1). In addition to its important role in solute separation and purification, affinity chromatography thus also possesses considerable potential for investigating the functional roles of the reactants thereby purified.

Interpretation of the reversible inhibition of adenosine deaminase by small cosolutes in terms of molecular crowding.

Published results on the inhibitory effects of small cosolutes on adenosine deamination by adenosine deaminase [Kurz, L. C., Weitkamp, E., and Frieden, C. (1987) Biochemistry 26, 3027-3032; Dzingeleski, G., and Wolfenden, R. (1993) Biochemistry 32, 9143-9147] have been reexamined. Results for sucrose, dioxane, methanol, and ethanol are shown to be qualitatively consistent with thermodynamic interpretation in terms of molecular crowding effects arising from the occurrence of a minor increase in enzyme volume and/or asymmetry during the kinetic reaction--a conformational transition that could be either preexisting or ligand induced. Direct evidence for the existence of the putative isomeric transition is provided by active enzyme gel chromatography on Sephadex G-100, which demonstrates a negative dependence of enzyme elution volume upon substrate concentration and is therefore consistent with substrate-mediated conformational changes that favor a larger (or more asymmetric) isomeric state of the enzyme. There are thus experimental grounds for adopting the present description of the inhibitory effects of unrelated cosolutes on the kinetics of adenosine deamination by adenosine deaminase in terms of thermodynamic nonideality.

Studies of solute self-association by sedimentation equilibrium: allowance for effects of thermodynamic non-ideality beyond the consequences of nearest-neighbor interactions.

A sedimentation equilibrium study of alpha-chymotrypsin self-association in acetate-chloride buffer, pH 4.1 I 0.05, has been used to illustrate determination of a dimerization constant under conditions where thermodynamic non-ideality is manifested beyond the consequences of nearest-neighbor interactions. Because the expressions for the experimentally determinable interaction parameters comprise a mixture of equilibrium constant and excluded volume terms, the assignment of reasonable magnitudes to the relevant virial coefficients describing non-associative cluster formation is essential for the evaluation of a reliable estimate of the dimerization constant. Determination of these excluded volume parameters by numerical integration over the potential-of-mean-force is shown to be preferable to their calculation by approximate analytical solutions of the integral for this relatively small enzyme monomer with high net charge (+10) under conditions of low ionic strength (0.05 M).

Interpreting the effects of small uncharged solutes on protein-folding equilibria.

Proteins are designed to function in environments crowded by cosolutes, but most studies of protein equilibria are conducted in dilute solution. While there is no doubt that crowding changes protein equilibria, interpretations of the changes remain controversial. This review combines experimental observations on the effect of small uncharged cosolutes (mostly sugars) on protein stability with a discussion of the thermodynamics of cosolute-induced nonideality and critical assessments of the most commonly applied interpretations. Despite the controversy surrounding the most appropriate manner for interpreting these effects of thermodynamic nonideality arising from the presence of small cosolutes, experimental advantage may still be taken of the ability of the cosolute effect to promote not only protein stabilization but also protein self-association and complex formation between dissimilar reactants. This phenomenon clearly has potential ramifications in the cell, where the crowded environment could well induce the same effects.

Ultracentrifugal studies of the effect of molecular crowding by trimethylamine N-oxide on the self-association of muscle glycogen phosphorylase b.

The suitability of sedimentation equilibrium for characterizing the self-association of muscle glycogen phosphorylase b has been reappraised. Whereas sedimentation equilibrium distributions for phosphorylase b in 40 mM Hepes buffer (pH 6.8) supplemented with 1 mM AMP signify a lack of chemical equilibrium attainment, those in buffer supplemented additionally with potassium sulfate conform with the requirements of a dimerizing system in chemical as well as sedimentation equilibrium. Because the rate of attainment of chemical equilibrium under the former conditions is sufficiently slow to allow resolution of the dimeric and tetrameric enzyme species by sedimentation velocity, this procedure has been used to examine the effects of thermodynamic nonideality arising from molecular crowding by trimethylamine N-oxide on the self-association behaviour of phosphorylase b. In those terms the marginally enhanced extent of phosphorylase b self-association observed in the presence of high concentrations of the cosolute is taken to imply that the effects of thermodynamic nonideality on the dimer-tetramer equilibrium are being countered by those displacing the T<==>R isomerization equilibrium for dimer towards the smaller, nonassociating T state. Because the R state is the enzymically active form, an inhibitory effect is the predicted consequence of molecular crowding by high concentrations of unrelated solutes. Thermodynamic nonideality thus provides an alternative explanation for the inhibitory effects of high concentrations of glycerol, sucrose and ethylene glycol on phosphorylase b activity, phenomena that have been attributed to extremely weak interaction of these cryoprotectants with the T state of the enzyme.

Theoretical aspects of the quantitative characterization of ligand binding.

Living organisms grow, differentiate, reproduce, and respond to their environment via specific and integrated interactions between biomolecules. The investigation of molecular interactions therefore constitutes a major area of biochemical study, occupying a ubiquitous and central position between molecular physiology on the one hand and structural chemistry on the other. While specificity resides in the details of structural recognition, the dynamic interplay between biomolecules is orchestrated precisely by the thermodynamics of the biomolecular equilibria involved. A common set of physicochemical principles applies to all such phenomena, irrespective of whether the interaction of interest involves an enzyme and its substrate or inhibitor, a hormone or growth factor and its receptor, an antibody and its antigen, or, indeed, the binding of effector molecules that modulate these interactions. The binding affinity, binding specificity, number of binding sites per molecule, as well as the enthalpic and entropic contributions to the binding energy are common parameters that assist an understanding of the biochemical outcome. This unit aims to provide an overview of the design and interpretation of binding experiments.

Analysis of thermodynamic non-ideality in terms of protein solvation.

The effects of thermodynamic non-ideality on the forms of sedimentation equilibrium distributions for several isoelectric proteins have been analysed on the statistical-mechanical basis of excluded volume to obtain an estimate of the extent of protein solvation. Values of the effective solvation parameter delta are reported for ellipsoidal as well as spherical models of the proteins, taken to be rigid, impenetrable macromolecular structures. The dependence of the effective solvated radius upon protein molecular mass exhibits reasonable agreement with the relationship calculated for a model in which the unsolvated protein molecule is surrounded by a 0.52-nm solvation shell. Although the observation that this shell thickness corresponds to a double layer of water molecules may be of questionable relevance to mechanistic interpretation of protein hydration, it augurs well for the assignment of magnitudes to the second virial coefficients of putative complexes in the quantitative characterization of protein-protein interactions under conditions where effects of thermodynamic non-ideality cannot justifiably be neglected.

Three-dimensional structure of RK-1: a novel alpha-defensin peptide.

NMR spectroscopy and simulated annealing calculations have been used to determine the three-dimensional structure of RK-1, an antimicrobial peptide from rabbit kidney recently discovered from homology screening based on the distinctive physicochemical properties of the corticostatins/defensins. RK-1 consists of 32 residues, including six cysteines arranged into three disulfide bonds. It exhibits antimicrobial activity against Escherichia coli and activates Ca(2+) channels in vitro. Through its physicochemical similarity, identical cysteine spacing, and linkage to the corticostatins/defensins, it was presumed to be a member of this family. However, RK-1 lacks both a large number of arginines in the primary sequence and a high overall positive charge, which are characteristic of this family of peptides. The three-dimensional solution structure, determined by NMR, consists of a triple-stranded antiparallel beta-sheet and a series of turns and is similar to the known structures of other alpha-defensins. This has enabled the definitive classification of RK-1 as a member of this family of antimicrobial peptides. Ultracentrifuge measurements confirmed that like rabbit neutrophil defensins, RK-1 is monomeric in solution, in contrast to human neutrophil defensins, which are dimeric.

Synthesis and evaluation of affinity adsorbents for glycoproteins: an artificial lectin.

A combination of rational design based on mimicking natural protein-carbohydrate interactions and solid-phase combinatorial chemistry has led to the identification of an affinity ligand which displays selectivity for the mannose moiety of glycoproteins. The ligand, denoted 18/18 and comprising a triazine scaffold bis-substituted with 5-aminoindan, has been synthesised in solution, characterised by TLC, 1H-NMR and MS. When immobilised to amine-derivatised agarose at concentrations >24 micromol/g moist weight gel, ligand 18/18 selectively binds glucose oxidase. The adsorbed enzyme was quantitatively eluted with 0.5 M alpha-D-methyl-mannoside and to a lesser extent with the equivalent glucoside. An investigation of the comparative retention times of saccharidic solutes showed that significant retardation was observed for alpha-D-mannose, mannobiose and mannan, with little or no evidence for selective retention of other saccharides, with the exception of alpha-L-fucose. Interestingly, alpha-L-fucose and alpha-D-mannose share an identical configuration of the hydroxyl groups on C-2, C-3 and C-4. Analysis of Scatchard plots from partition equilibrium studies on the interaction of glucose oxidase and the p-nitrophenyl-glycosides of D-mannose, D-glucose, L-fucose and D-galactose with immobilised 18/18 establish that the affinity constants (K(AX)) for the enzyme, the glycosides of mannose, glucose and fucose, and the p-nitrophenyl-galactoside are 4.3x10(5) M(-1), 1.9X10(4) M(-1) and 1.2x10(4) M(-1) respectively. 1H-NMR studies on the interaction of alpha-D-methyl-mannoside with ligand 18/18 in solution confirm the involvement of the hydroxyl group in the C-2 position. Molecular modelling suggests the formation of four hydrogen bonds between the hydroxyl groups at positions C-2, C-3 and C-4 of alpha-D-methyl-mannoside and the bridging and ring nitrogen atoms of the triazine scaffold, with aromatic stacking of a second ligand against the carbohydrate face. The greater specificity of ligand 18/18 for mannose and glucose than for galactose parallels that exhibited by concanavalin A.

Analysis of sedimentation equilibrium distributions reflecting nonideal macromolecular associations.

A rigorous statistical-mechanical approach is adopted to derive general quantitative expressions that allow for the effects of thermodynamic nonideality in equilibrium measurements reflecting interaction between dissimilar macromolecular reactants. An analytical procedure based on these expressions is then formulated for obtaining global estimates of equilibrium constants and the corresponding reference thermodynamic activities of the free reactants in each of several sedimentation equilibrium experiments. The method is demonstrated by application to results from an ultracentrifugal study of an electrostatic interaction between ovalbumin and cytochrome c (Winzor, D. J., M. P. Jacobsen, and P. R. Wills. 1998. Biochemistry. 37:2226-2233). It is demonstrated that reliable estimates of relevant thermodynamic parameters are extracted from the data through statistical analysis by means of a simple nonlinear fitting procedure.

Binding of an RNA trafficking response element to heterogeneous nuclear ribonucleoproteins A1 and A2.

Heterogeneous nuclear ribonucleoprotein (hnRNP) A2 binds a 21-nucleotide myelin basic protein mRNA response element, the A2RE, and A2RE-like sequences in other localized mRNAs, and is a trans-acting factor in oligodendrocyte cytoplasmic RNA trafficking. Recombinant human hnRNPs A1 and A2 were used in a biosensor to explore interactions with A2RE and the cognate oligodeoxyribonucleotide. Both proteins have a single site that bound oligonucleotides with markedly different sequences but did not bind in the presence of heparin. Both also possess a second, specific site that bound only A2RE and was unaffected by heparin. hnRNP A2 bound A2RE in the latter site with a K(d) near 50 nm, whereas the K(d) for hnRNP A1 was above 10 microm. UV cross-linking assays led to a similar conclusion. Mutant A2RE sequences, that in earlier qualitative studies appeared not to bind hnRNP A2 or support RNA trafficking in oligodendrocytes, had dissociation constants above 5 microm for this protein. The two concatenated RNA recognition motifs (RRMs), but not the individual RRMs, mimicked the binding behavior of hnRNP A2. These data highlight the specificity of the interaction of A2RE with these hnRNPs and suggest that the sequence-specific A2RE-binding site on hnRNP A2 is formed by both RRMs acting in cis.

From gel filtration to biosensor technology: the development of chromatography for the characterization of protein interactions.

The objective of this review is to summarize the development of chromatographic techniques for the determination of reaction stoichiometries and equilibrium constants for solute interactions of biological importance. Gel chromatography is shown to offer a convenient means of characterizing solute self-association as well as solute-ligand interactions. Affinity chromatography is an even more versatile method of characterizing interactions between dissimilar reactants because the biospecificity incorporated into the design of the affinity matrix ensures applicability of the method regardless of the relative sizes of the two reactants. Adoption of different experimental strategies such as column chromatography, simple partition equilibrium experiments and biosensor technology has created a situation wherein affinity chromatography affords a means of characterizing the whole range of reaction affinities-from relatively weak interactions (binding constants less that 10(3)M (-1)) to tight interactions with binding constants greater than 10(9)M (-1). In addition to its established prowess as a means of solute separation and purification, chromatography thus also possesses considerable potential for investigation of the functional roles of the purified reactants-an endeavour that requires characterization as well as identification of the interactions responsible for a physiological phenomenon.

Textilinins from Pseudonaja textilis textilis. Characterization of two plasmin inhibitors that reduce bleeding in an animal model.

The incidence of vein-graft occlusion associated with myocardial infarction and thrombosis following the use of the plasmin inhibitor, aprotinin, to reduce blood loss during vascular surgery has prompted the isolation of an alternative kinetically distinct inhibitor of plasmin from the venom of Pseudonaja textilis. This inhibitor has been called textilinin (Txln) and two distinct forms have been isolated from the Brown-snake venom (molecular weight, 6688 and 6692). A comparison of plasmin inhibitor constants for aprotinin and the Txlns 1 and 2 indicated that the former bound very tightly (inhibitor constant, Ki approximately 10(-11) mol/l), while both of the latter bound less tightly (Ki approximately 10(-9) mol/l). Homogeneity of Txlns 1 and 2 was confirmed by sodium dodecyl sulphate-polyacrylamide gel electrophoresis and mass spectrometry. A sequence difference of six amino acids was observed between the two forms of Txln. Txln 1 and 2 showed, respectively, 45 and 43% homology with aprotinin, while there was 58 and 55% homology, respectively, with a plasmin inhibitor from the venom of eastern Taipan, Oxyuranus scutellatus. Both Txlns have six cysteines, like other inhibitors of this group, and homology was determined by alignment of these cysteines. Both have been shown to reduce blood loss by about 60% in a murine tail vein bleeding model. It is proposed that the kinetic profiles of Txln 1 and 2 for plasmin allow the arrest of haemorrhage without the possible threat of thrombosis.

Stoichiometry of complex formation between Copper(I) and the N-terminal domain of the Menkes protein.

The inherent cellular toxicity of copper ions demands that their concentration be carefully controlled. The cellular location of the Menkes ATPase, a key element in the control of intracellular copper, is regulated by the intracellular copper concentration through the N-terminus of the enzyme, comprising 6 homologous subdomains or modules, each approximately 70 residues in length and containing a -Cys-X-X-Cys- motif. Based on the proposal that binding of copper to these modules regulates the Menkes ATPase cellular location by promoting changes in the tertiary structure of the enzyme, we have expressed the entire N-terminal domain (MNKr) and the second metal-binding module (MNKr2) of the Menkes protein in E. coli and purified them to homogeneity. Ultraviolet-visible, luminescence, and X-ray absorption spectroscopy show that copper and silver bind to the single module, MNKr2, with a stoichiometry of one metal ion per module. However, the array of six modules, MNKr, binds Cu(I) to produce a homogeneous conformer with 4 mol equiv of metal ion. The metal ions are bound in an environment that is shielded from solvent molecules. We suggest a model of the Menkes protein in which the Cu(I) binding induces tertiary changes in the organization of the six metal-binding domains.

Interpretation of thermodynamic non-ideality in sedimentation equilibrium experiments on proteins.

This investigation re-examines theoretical aspects of the allowance for effects of thermodynamic non-ideality on the sedimentation equilibrium distribution for a single macromolecular solute, and thereby resolves the question of the constraints that pertain to the definition of the activity coefficient term in the basic sedimentation equilibrium expression. Sedimentation equilibrium results for ovalbumin are then presented to illustrate a simple procedure for evaluating the net charge (valence) of a protein from the magnitude of the second virial coefficient in situations where the effective radius of the protein can be assigned. Finally, published sedimentation equilibrium results on lysozyme are reanalysed to demonstrate the feasibility of employing the dependence of the second virial coefficient upon ionic strength to evaluate both the valence and the effective radius of the non-interacting solute.

Consequences of the non-specific binding of a protein to a linear polymer: reconciliation of stoichiometric and equilibrium titration data for the thrombin-heparin interaction.

Theoretical aspects of the thermodynamic characterization of cooperative protein interactions with non-specific segments of a linear polymer lattice have been re-examined. This reconsideration has not only provided an alternative derivation of recursive expressions for the stoichiometry of random ligand binding prior to elimination of the parking problem but also extended that treatment to include binding with overlap of additional lattice units. The major obstacle to thermodynamic characterization of non-specific protein-polymer interactions is determination of the lattice capacity for ligand, which in turn defines the length of the polymer segment to which the protein binds. Although these parameters are most readily obtained from studies under conditions that ensure essentially stoichiometric interaction, the endpoint of such a titration is likely to reflect the irreversible rather than the equilibrium binding capacity of the lattice for ligand. Consideration of published results for spectrofluorometric titrations of the thrombin-heparin system under stoichiometric conditions in such terms has permitted their reconciliation with results of a later publication on the interaction under equilibrium conditions.

Reassessment of the calibration constant for the IAsys biosensor.

A magnitude of 50 are s ng-1 mm2 has been determined for the calibration constant relating biosensor response to the amount of protein bound to the sensor surface of an IAsys cuvette. These studies entailed enzymatic assessment of the extent of lactate dehydrogenase depletion in the liquid phase arising from enzyme binding to a carboxymethyldextran-coated sensor surface, and also estimation of a maximum biosensor response for the electrostatic interaction of ovalbumin with an aminosilane-coated sensor surface. The latter results required correction for contributions to biosensor response resulting from changes in the refractive index of the liquid phase effected by high protein concentrations.