Interfacial thermodynamics of protein adsorption, ion co-adsorption and ion binding in solution. I. Phenomenological linkage relations for ion exchange in lysozyme chromatography and titration in solution

In this paper we discuss the thermodynamics of ion binding in solution, protein adsorption and ion co-adsorption. The emphasis is on charge regulation effects. To this end, we introduce phenomenological linkage relations from which the ion binding can be calculated from the electrolyte dependency of proton titration curves and the co-adsorption from the electrolyte dependency of protein adsorption isotherms. The linkage relations are derived from classical interfacial thermodynamics, and thus offer an alternative approach as compared to the mass balance equations which are currently used in biotechnology, and Record et al.'s 1978 analysis of Wyman's Binding Polynomial for protein interactions. The co-adsorption theory is an extension of our previous analysis of ion binding in solution, which we include here for comparison of the ion co-adsorption with the ion binding in solution. The theory is applied to the chromatography of lysozyme on the strong cation exchanger ‘mono S’ and to the proton titration of lysozyme in solution. In the accompanying Part 2 of this paper the results are interpreted with a simple model.     

Electrostatic interactions at charged lipid membranes. Calcium binding

A quantitative study of calcium-ion binding by the negatively-charged phospholipid methylphosphatidic acid is presented. Experimental results are compared with the predictions of the Gouy-Chapman theory, taking into account both the ions bound at the membrane surface and the ions held in the diffuse layer. This theory suffices to explain the titration of the calcium/lipid system, but fails to explain completely the behaviour of the ordered-fluid transition temperature, which shows a splitting that according to electrostatic theory alone should not occur. The dependence of the calcium-lipid binding constant. upon 1: 1 electrolyte concentration is correctly predicted by the theory; the latter however gives equations which can only be solved numerically. A simple, approximate equation is therefore given (in the text, eq. 34) for the prediction of the degree of calcium binding to a negatively-charged lipid membrane.     

Protein surface-distribution and protein-protein interactions in the binding of peripheral proteins to charged lipid membranes.

The binding of native cytochrome c to negatively charged lipid dispersions of dioleoyl phosphatidylglycerol has been studied over a wide range of ionic strengths. Not only is the strength of protein binding found to decrease rapidly with increasing ionic strength, but also the binding curves reach an apparent saturation level that decreases rapidly with increasing ionic strength. Analysis of the binding isotherms with a general statistical thermodynamic model that takes into account not only the free energy of the electrostatic double layer, but also the free energy of the surface distribution of the protein, demonstrates that the apparent saturation effects could arise from a competition between the out-of-plane binding reaction and the lateral in-plane interactions between proteins at the surface. It is found that association with nonlocalized sites results in binding isotherms that display the apparent saturation effect to a much more pronounced extent than does the Langmuir adsorption isotherm for binding to localized sites. With the model for nonlocalized sites, the binding isotherms of native cytochrome c can be described adequately by taking into account only the entropy of the surface distribution of the protein, without appreciable enthalpic interactions between the bound proteins. The binding of cytochrome c to dioleoyl phosphatidylglycerol dispersions at a temperature at which the bound protein is denatured on the lipid surface, but is nondenatured when free in solution, has also been studied. The binding curves for the surface-denatured protein differ from those for the native protein in that the apparent saturation at high ionic strength is less pronounced. This indicates the tendency of the denatured protein to aggregate on the lipid surface, and can be described by the binding isotherms for nonlocalized sites only if attractive interactions between the surface-bound proteins are included in addition to the distributional entropic terms. Additionally, it is found that the binding capacity for the native protein is increased at low ionic strength to a value that is greater than that for complete surface coverage, and that corresponds more closely to neutralization of the effective charge (determined from the ionic strength dependence), rather than of the total net charge, on the protein. Electron spin resonance experiments with spin-labeled lipids indicate that this different mode of binding arises from a penetration or disturbance of the bilayer surface by the protein that may alleviate the effects of in-plane interactions under conditions of strong binding.     

An ESR study of the effect of an electrostatic field on binding of divalent cations to the surface of serum low-density lipoproteins

The ESR technique has been used to study binding of Mn(II) ions to low-density lipoprotein (LDL) in solutions of various electrolyte ionic strengths. A model of the binding has been proposed which describes all the observations in electrolytes of ten different concentrations in terms of two types of binding sites and two corresponding sets of intrinsic binding parameters (n1 = 8, Kd1 = 1.31 X 10(-3) mol X l-1 and n2 = 170, Kd2 = 5.71 X 10(-2) mol X l-1). These parameters, together with the values of the potential (phi 0) responsible for binding of the ions to specific charged sites on the surface, reproduce the observed binding curves well in all the systems studied. The phi 0 values are obtained as an appropriate solution of the Poisson-Boltzmann equation.     

The interaction of various lanthanide ions and some anions with phosphatidylcholine vesicle membranes. A 31P NMR study of the surface potential effects

The interaction of various lanthanide ions with vesicles of phosphatidylcholine from egg yolk has been followed by 31P NMR at 30 degrees C. From known magnetic properties of these ions, separation of the paramagnetic shift into a pure contact and a pseudo-contact part was carried out. Binding curves for the contact contribution (F curves) were obtained from vesicles in solutions of sodium salts with monovalent anions over a wide concentration range. These curves should be insensitive to any conformational effects due to ion binding. Indication of a conformational change in the lipid head group at low ion binding was obtained by studying the ratio between the contact and the pseudo-contact contributions. Besides the adsorption of lanthanide ions, specific anion binding to the surface was introduced to account for the enhanced chemical shifts (Cl- < Br- < NO3-). The results were analyzed in terms of the theory for the diffuse double layer (Gouy-Chapman-Grahame) with equilibrium conditions for the adsorbing cations and anions. Simulations of the titration curves furnished parameters for the ion-lipid interactions. The synergism between the cations and anions follows from the potential effects. Comparison of results with lanthanide ions and Ca2+ indicates that the anion adsorption probably depends on the nature of the adsorbed cation. Lanthanide ion binding to L-glycerophosphorylcholine is not influenced by sodium salts. The binding constant for this complex is weaker than with phosphatidylcholine. The chemical shifts for the lanthanide ion complexes with these two phosphorus compounds seem to be about the same.     

Transport effects on the kinetics of protein-surface binding.

A detailed model is presented for protein binding to active surfaces, with application to the binding of avidin molecules to a biotin-functionalized fiber optic sensor in experiments reported by S. Zhao and W. M. Reichert (American Chemical Society Symposium Series 493, 1992). Kinetic data for binding in solution are used to assign an intrinsic catalytic rate coefficient k to the biotin-avidin pair, deconvoluted from transport and electrostatic factors via application of coagulation theory. This intrinsic chemical constant is built into a reaction-diffusion analysis of surface binding where activity is restricted to localized sites (representing immobilized biotin molecules). The analysis leads to an effective catalytic rate coefficient keff characterizing the active surface. Thereafter, solution of the transport problem describing absorption of avidin molecules by the macroscopic sensor surface leads to predictions of the avidin flux, which are found to be in good agreement with the experimental data. The analysis suggests the following conclusions. 1) Translational diffusion limitations are negligible for avidin-biotin binding in solution owing to the small (kinetically limiting) value k = 0.00045 m/s. 2) The sparse distribution of biotin molecules and the presence of a repulsive hydration force produce an effective surface-average catalytic rate coefficient keff of order 10(-7) m/s, much smaller than k. 3) Avidin binding to the fiber optic sensor occurs in an intermediate regime where the rate is influenced by both kinetics and diffusion.     

Generalized binding phenomena in an allosteric macromolecule

A general macromolecular partition function is developed in terms of chemical ligand activity, temperature and pressure for systems described by an array of species which are characterized by their state of allosteric conformation and ligand stoichiometry. The effects of chemical ligand binding, enthalpy change, and volume change are treated in a parallel manner. From a broad viewpoint all of these effects can be regarded as specific cases of generalized binding phenomena. This approach provides a general method for analyzing calorimetric and ligand binding experiments. Several applications are given: (1) Thermal scanning data for tRNAphe (P.L. Privalov and V.V. Filimonov, J. Mol. Biol. 122 (1978) 447) are shown to fit a general model with six conformational states. By application of linkage theory it is shown that sodium chloride is expelled as the molecule denatures. (2) The results of calorimetric titrations on the arabinose binding protein (H. Fukada, J.M. Sturtevant and F.A. Quiocho, J. Mol. Biol. 258 (1983) 13193) are shown to fit a simple two-state allosteric model. (3) A thermal binding curve is simulated for an unusual respiratory protein, trout I hemoglobin (B.G. Barisas and S.J. Gill, Biophys. Chem. 9 (1979) 235), in order to illustrate both the similarities and differences between enthalpy and chemical ligand binding processes.     

Oxygen binding to sickle cell hemoglobin.

The extent of oxygen binding and light scattering of concentrated solutions of hemoglobin S have been determined as a function of oxygen partial pressure using a thin film optical cell. Nearly reversible oxygen binding is observed as witnessed by the small hysteresis found between slow deoxygenation and reoxygenation runs. High co-operativity is noted from unusually large concentration-dependent Hill coefficients when aggregated hemoglobin S is present. The application of linkage theory with the inclusion of non-ideal solution properties permits a test of various simple models for oxygen binding to both the monomer (α₂β₂^s) and polymer (aggregated) phase. It is concluded that oxygen binding to the polymer is either negligible or small under present experimental conditions. Phase diagrams of the solution concentration in equilibrium with polymer phase as a function of oxygen partial pressure are derived using best fit values of polymer parameters.     

An analogue of the binding polynomial for the case of ligands binding to an aggregating macromolecule

It is known that when a macromolecule exists as a mixture of forms with different molecular weights in equilibrium in dilute solution with ligands then the saturation functions cannot be derived by differentiating a binding polynomial. However, we demonstrate that in such circumstances a saturation function can be expressed in terms of the binding polynomials for each of the polymeric forms of macromolecule and integration of this using the conservation equation affords a function, Q, which is the analogue of the binding polynomial for a non-aggregating system. The existence of Q leads to heterotropic linkage relationships.     

Dissecting the N-terminal myosin binding site of human cardiac myosin-binding protein C. Structure and myosin binding of domain C2

Myosin-binding protein C (MyBP-C) binds to myosin with two binding sites, one close to the N terminus and the other at the C terminus. Here we present the solution structure of one part of the N-terminal binding site, the third immunoglobulin domain of the cardiac isoform of human MyBP-C (cC2) together with a model of its interaction with myosin. Domain cC2 has the beta-sandwich structure expected from a member of the immunoglobulin fold. The C-terminal part of the structure of cC2 is very closely related to telokin, the myosin binding fragment of myosin light chain kinase. Domain cC2 also contains two cysteines on neighboring strands F and G, which would be able to form a disulfide bridge in a similar position as in telokin. Using NMR spectroscopy and isothermal titration calorimetry we demonstrate that cC2 alone binds to a fragment of myosin, S2Delta, with low affinity (kD = 1.1 mM) but exhibits a highly specific binding site. This consists of the C-terminal surface of the C'CFGA' beta-sheet, which includes Glu(301), a residue mutated to Gln in the disease familial hypertrophic cardiomyopathy. The binding site on S2 was identified by a combination of NMR binding experiments of cC2 with S2Delta containing the cardiomyopathy-linked mutation R870H and molecular modeling. This mutation lowers the binding affinity and changes the arrangement of side chains at the interface. Our model of the cC2-S2Delta complex gives a first glimpse of details of the MyBP-C-myosin interaction. Using this model we suggest that most key interactions are between polar amino acids, explaining why the mutations E301Q in cC2 and R870H in S2Delta could be involved in cardiomyopathy. We expect that this model will stimulate future research to further refine the details of this interaction and their importance for cardiomyopathy.     

Membrane composition affects the reversibility of annexin A2t binding to solid supported membranes: a QCM study

By means of the quartz crystal microbalance (QCM) technique, we investigated the interaction of porcine heterotetrametric annexin A2t with solid supported lipid membranes. Dissociation and rate constants of annexin A2t binding to various lipid mixtures were determined as a function of Ca2+ concentrations in solution. In contrast to what has been observed for annexin A1, the binding affinity and kinetics of annexin A2t binding are not influenced by cholesterol. In the experimental setup chosen, the annexin A2t binding is strictly Ca2+-dependent and only affected by the amount of phosphatidylserine (PS) in the membrane and the Ca2+ concentration in solution. By Ca2+-titration experiments at constant annexin A2t concentration, we investigated the reversibility of annexin A2t adsorption and desorption. Surprisingly, Ca2+-titration curves display a significant hysteresis. Protein desorption curves starting from annexin A2t bound to the membrane at 1 mM CaCl2 exhibit high cooperativity with half-maximum Ca2+ concentrations in the submicromolar range. However, protein adsorption curves starting from an EGTA-containing solution with soluble annexin A2t always show two inflection points upon addition of Ca2+ ions. These two inflection points may be indicative of two protein populations differently bound to the solid-supported membrane. The ratio of these two annexin A2t populations depends on the amount of PS molecules and cholesterol in the membrane as well as on the Ca2+ concentration. We propose a model discussing the results obtained in terms of two binding sites differing in their affinity due to lipid rearrangement.     

Isolation of collagen binding proteins from embryonic chicken corneal epithelial cells

In the present paper, we report the isolation and characterization of embryonic corneal membrane glycoproteins that demonstrate specific affinity for collagen. Two collagen binding proteins have been isolated: a novel 70-kDa protein and a 47-kDa protein which is apparently similar to that reported by Kurkinen et al. (Kurkinen, M., Taylor, A., Garrels, J. I., and Hogan, B. (1984) J. Biol. Chem. 259, 5915-5922). Both proteins label metabolically with [35S]methionine and [3H] glucosamine. 125I iodination of cell surface proteins revealed that the two collagen binding proteins are expressed on the epithelial cell surface. The 70-kDa protein appears to be an integral membrane protein, whereas the 47-kDa protein can be removed from membranes by alkali treatment. The isolated proteins exhibit binding to native type IV collagen as well as heat-denatured type I collagen. It seems likely that we have isolated, at least in part, the cell surface receptor or receptor complex that binds collagen to the basal surface of epithelia.     

Isolation of ribosomal protein-RNA complexes by nitrocellulose membrane filtration: equilibrium binding studies.

E. coli ribosomal proteins are retained by nitrocellulose filters. In contrast, 16S RNA passes through nitrocellulose filters. We have found that specific protein-RNA complexes involving single proteins also pass through nitrocellulose filters. Thus, by utilizing radioactively labeled r-proteins, nitrocellulose filtration can be used to study directly and sensitively the stoichiometry of r-protein-RNA association. The filtration process maintains near equilibrium conditions, making it applicable to weak as well as strong protein-RNA associations. We have used nitrocellulose filtration to obtain saturation binding curves for the association of proteins S4, S7, S8 and S20 with 16S RNA. In each case, the stoichiometry of binding was one mole of protein or less per mole of RNA. The stoichiometry of protein S8 binding to 16S RNA measured by filtration is comparable to that observed by sucrose gradient centrifugation. Association constants for the binding of proteins S4, S8 and S20 to 16S RNA have been determined by analysis of the saturation binding curves and were found to range from .3-6 X 10(7)M-1.     

Thermodynamics of metal ion binding. 2. Metal ion binding by carbonic anhydrase variants

The ability to construct molecular motifs with predictable properties in aqueous solution requires an extensive knowledge of the relationships between structure and energetics. The design of metal binding motifs is currently an area of intense interest in the bioorganic community. To date synthetic motifs designed to bind metal ions lack the remarkable affinities observed in biological systems. To better understand the structural basis of metal ion affinity, we report here the thermodynamics of binding of divalent zinc ions to wild-type and mutant carbonic anhydrases and the interpretation of these parameters in terms of structure. Mutations were made both to the direct His ligand at position 94 and to indirect, or second-shell, ligands Gln-92, Glu-117, and Thr-199. The thermodynamics of ligand binding by several mutant proteins is complicated by the development of a second zinc binding site on mutation; such effects must be considered carefully in the interpretation of thermodynamic data. In all instances modification of the protein produces a complex series of changes in both the enthalpy and entropy of ligand binding. In most cases these effects are most readily rationalized in terms of ligand and protein desolvation, rather than in terms of changes in the direct interactions of ligand and protein. Alteration of second-shell ligands, thought to function primarily by orienting the direct ligands, produces profoundly different effects on the enthalpy of binding, depending on the nature of the residue. These results suggest a range of activities for these ligands, contributing both enthalpic and entropic effects to the overall thermodynamics of binding. Together, our results demonstrate the importance of understanding relationships between structure and hydration in the construction of novel ligands and biological polymers.     

Protein GRAB of streptococcus pyogenes regulates proteolysis at the bacterial surface by binding alpha2-macroglobulin.

In the molecular interplay between pathogenic microorganisms and their host, proteolytic mechanisms are believed to play a crucial role. Here we find that the important human pathogen Streptococcus pyogenes (group A Streptococcus) expresses a surface protein with high affinity (Ka = 2.0 x 10(8) M-1) for alpha2-macroglobulin (alpha2M), the dominating proteinase inhibitor of human plasma. The immunoglobulin-binding protein G of group C and G streptococci also contains an alpha2M-binding domain and a gene encoding protein GRAB (protein G-related alpha2M-binding protein) was identified in the S. pyogenes Genome Sequencing data base. The grab gene is present in most S. pyogenes strains and is well conserved. Protein GRAB has typical features of a surface-attached protein of Gram-positive bacteria. It also contains a region homologous to parts of the alpha2M-binding domain of protein G and a variable number of a unique 28-amino acid-long repeat. Using Escherichia coli-produced protein GRAB and synthetic GRAB peptides, the alpha2M-binding region was mapped to the NH2-terminal part of protein GRAB, which is the region with homology to protein G. An isogenic S. pyogenes mutant lacking surface-associated protein GRAB showed no alpha2M binding activity and was attenuated in virulence when injected intraperitoneally in mice. Finally, alpha2M bound to the bacterial surface via protein GRAB was found to entrap and inhibit the activity of both S. pyogenes and host proteinases, thereby protecting important virulence determinants from proteolytic degradation. This regulation of proteolytic activity at the bacterial surface should affect the host-microbe relation during S. pyogenes infections.     

The effect of heme-linked ionizable groups on cyanide binding to methemoglobin

The pH dependence of the kinetics of the binding of cyanide ion to methemoglobins A and S and to guinea pig and pigeon methemoglobins appears to be not directly correlated with the net charges on the proteins. The kinetics can, however, be adequately explained in terms of three sets of heme-linked ionizable groups with pK1 ranging between 4.9 and 5.3, pK2 between 6.2 and 7.9, and pK3 between 8.0 and 8.5 at 20 degrees C. pK1 is assigned to carboxylic acid groups, pK2 to histidines and terminal amino groups, and pK3 to the acid-alkaline methemoglobin transition. Kinetic second order rate constants have also been determined for the binding of cyanide ion by the four sets of methemoglobin species present in solution. The pKi values and the rate constants of methemoglobin S are strikingly different from those of methemoglobin A. This result is explained in terms of different electrostatic contributions to the free energy of heme linkage arising from differences in the environments of ionizable groups at the surfaces of the two molecules.