Remeasuring HEWL pK(a) values by NMR spectroscopy: methods, analysis, accuracy, and implications for theoretical pK(a) calculations

Site-specific pK(a) values measured by NMR spectroscopy provide essential information on protein electrostatics, the pH-dependence of protein structure, dynamics and function, and constitute an important benchmark for protein pK(a) calculation algorithms. Titration curves can be measured by tracking the NMR chemical shifts of several reporter nuclei versus sample pH. However, careful analysis of these curves is needed to extract residue-specific pK(a) values since pH-dependent chemical shift changes can arise from many sources, including through-bond inductive effects, through-space electric field effects, and conformational changes. We have re-measured titration curves for all carboxylates and His 15 in Hen Egg White Lysozyme (HEWL) by recording the pH-dependent chemical shifts of all backbone amide nitrogens and protons, Asp/Glu side chain protons and carboxyl carbons, and imidazole protonated carbons and protons in this protein. We extracted pK(a) values from the resulting titration curves using standard fitting methods, and compared these values to each other, and with those measured previously by 1H NMR (Bartik et al., Biophys J 1994;66:1180–1184). This analysis gives insights into the true accuracy associated with experimentally measured pK(a) values. We find that apparent pK(a) values frequently differ by 0.5–1.0 units depending upon the nuclei monitored, and that larger differences occasionally can be observed. The variation in measured pK(a) values, which reflects the difficulty in fitting and assigning pH-dependent chemical shifts to specific ionization equilibria, has significant implications for the experimental procedures used for measuring protein pK(a) values, for the benchmarking of protein pK(a) calculation algorithms, and for the understanding of protein electrostatics in general.     

Cause of abnormal acidity of lysozyme ionogenic active center groups at cell wall lysis

pH-Dependence of the kinetic parameters of Micrococcus lysodeicticus cell lysis under the action of the protein hen egg lysozyme at the pH 6.9-10.0 at 25 and 37 degrees C has been investigated. The pKb effective values for the lysozyme catalytic activity controlling group have been calculated. The DeltaHion value indicates that this group is the carboxyl one though its pK (9.15 at 25 degrees C) is found far for the limit of the carboxyl groups pK values. The cause of this abnormal pK values is supposed to be the strong negative charge of the bacterial cell wall. As a result the enzyme that catalyzes the hydrolysis ofcopolymer N-acetylglucosamine--N-acetylmuramic acid acts in the high acidity microenvironment.     

Discrete charge calculations of potentiometric titrations for globular proteins: sperm whale myoglobin, hemoglobin alpha chain, cytochrome c.

The modified Tanford-Kirkwood theory of Shire et al. for intramolecular electrostatic interactions has been applied to hydrogen ion equilibria of sperm whale ferrimyoglobin, human hemoglobin α-chain and horse cytochrome c. The model employs two sets of parameters derived from the crystalline protein structures, first, the atomic coordinates of charged amino acid residues and, second, static accessibility factors to reflect their solvent exposure. In addition, a consistent set of intrinsic pK values (pK_int) for the individual groups is employed. The theoretical pK values at half-titration for individual groups in each protein correspond to the available observed pK values, and the theoretical titration curves compare closely with experimental potentiometric curves.     

Proton-magnetic-resonance studies of the lysine residues of ribonuclease A.

The amino groups of ribonuclease A (RNase-A) have been methylated with formaldehyde and borohydride to provide observable resonances for proton magnetic resonance (PMR) studies. Although enzymatic activity is lost, PMR difference spectroscopy and PMR studies of thermal denaturation show native conformation is largely preserved in methylated RNase-A. Resonances corresponding to the NH2-terminal alpha-amino and 10 xi-amino N-methyl groups are titrated at 220 MHz to obtain pK values. After correction for the effects of methylation, using values previously derived from model compound studies, a pK of 6.6 is found for the alpha-amino group, a pK of 8.6 for the xi-amino group of lysine-41 and pK values ranging from 10.6 to 11.2 for the other lysine xi-amino groups. Interactions between lysine-7 and lysine-41 or between the alpha-amino and xi-amino groups of lysine-1 have been proposed to account for deviations from simple titration behaviour. The correct continuities for the titration curves of the histidine H-2 proton resonances have been confirmed by selective deuteration of the H-2 protons. Titration curves for the H-2 proton resonances of histidine-12 and histidine-119 of methylated RNase-A show deviations from the titration curves for the native enzyme, indicating some alteration of the active-site conformation. In the presence of phosphate, titration curves for the H-2 proton resonances of histidine-12 and histidine-119 of methylated RNase-A indicate binding of phosphate at the active site, but these curves continue to show deviations from the titration behaviour of native RNase-A. The titration curve for the N-methyl resonance of lysine-41 is perturbed considerably by the presence of phosphate, which indicates a possible catalytic role for lysine-41.     

Kinetics of local helix formation in poly-L-glutamic acid studied by time-resolved photoacoustics: neutralization reactions of carboxylates in aqueous solutions and their relevance to the problem of protein folding

Photoactivatable caged protons have been used to trigger proton transfer reactions in aqueous solutions of acetate, glutamate, and poly-L-glutamic acid, and the volumetric and enthalpic changes have been detected and characterized by means of time-resolved photoacoustics. Neutralization of carboxylates in aqueous solutions invariably results in an expansion of the solution due to the disappearance of two charges and is accompanied by little enthalpic change. The reactions occur with thermally activated, apparent bimolecular rates on the order of 10(10) M(-1)s(-1). In the case of aqueous solutions of poly-L-glutamic acid at pH around the pK(a) of the coil-to-helix transition, diffusional binding of a proton by carboxylates is followed by a sequential reaction with rate 1.06 (+/- 0.05) x 10(7)s(-1). This step is not thermally activated in the temperature range we have investigated and is likely related to local formation of hydrogen bonds near the protonation site. This structural event may constitute a rate-limiting step in helix propagation.     

Ultrasonic absorption studies of protein-buffer interactions

The acoustic absorption of protein solutions in the presence of phosphate and other buffering ions has been studied in the physiological pH range. Buffers containing hydroxyl residues as titratable groups cause a pronounced increase of protein sound absorption, which is attributed to relaxation processes of proton transfer reactions between buffer ions and accessible imidazole and -amino groups of the protein surface. Amino group based buffers like Good's buffers do not induce additional sound absorption. Measurement of the ultrasonic absorption as a function of pH and of buffer concentration, and corresponding parameter fitting of the equation describing proton transfer relaxation processes has been used to evaluate equilibrium parameters. For the imidazole group of the amino acid histidine a pK value of 6.22 and for the imidazole group of the protein lysozyme a pK value of 5.71 have been determined. In hemoglobin the ligand-linked pK changes have been monitored by recording ultrasonic titration curves.     

Numerical method to analyze pH-profile from initial velocity data and semi-empirical procedure to identify catalytic residues of enzymes using pK and heat of ionization

A numerical computing method to estimate pK and heat of ionization directly from initial velocity data was developed for systematic analysis of rate parameters. Values of rate and thermodynamic parameters were determined together with standard deviations, without using Arrhenius and van't Hoff plots. Distributions of pK and heat of ionization of small-molecular derivatives of aliphatic carboxylates, imidazoles, thiols, and aliphatic amines as possible models for catalytic residues of enzymes were analyzed from a thermodynamic point of view. Each group was classified in a parallelogram without overlapping, except for a part of the imidazole and thiol groups. It was shown that values of pK and heat of ionization of carboxylates and histidine residues involved in the catalytic activity of enzymes determined from pH-profile experiments can be classified into similar, though somewhat extended, areas to those of the corresponding small-molecular derivatives. Identification of catalytic residues using values of pK and heat of ionization is proved to be a reliable method when the procedure is properly used.     

Beryllium binding at neutral pH: the importance of the Be-O-Be motif

Beryllium speciation at physiological conditions is critical to understanding chronic beryllium disease (CBD). The MHC-class II receptor alleles that have been linked to CBD have more than six carboxylates in a short 20 amino acid segment of the binding pocket and it has been suggested that beryllium may bind within the MHC-class II receptor via the carboxylates. Previous reports also show that citric acid binds beryllium significantly stronger than similar carboxylate ligands such as tartaric acid and is one of the few ligands that can compete with hydrolysis to solubilize beryllium across the entire pH range at molar concentrations. We have characterized the binding of Be to citric acid and shown using a combination of NMR, mass spectrometry and ligand competition studies that Be2L and Be4L2 species dominate. A Be-O-Be linkage with the bridging oxygen coming from the aliphatic alcohol is critical to the stability of the complex. We show through competition experiments that the most stable Be-O-Be arrangement has one Be in a five-member ring and the other Be in a six-member ring. The unusual deprotonation of an aliphatic alcohol (pK(a) = 18) at neutral pH has significant ramifications on the potential interactions of Be with biological ligands such as carbohydrates and Ser and Thr residues.     

Chloride--bicarbonate exchange in red blood cells: physiology of transport and chemical modification of binding sites

About 80% of the CO2 formed by metabolism is transported from tissues to lungs as bicarbonate ions in the water phases of red cells and plasma. The catalysed hydration of CO2 to bicarbonate takes place in the erythrocytes but most of the bicarbonate thus formed must be exchanged with extracellular chloride to make full use of the carbon dioxide transporting capacity of the blood. The anion transport capacity of the red cell membrane is among the largest ionic transport capacities of any biological membrane. Exchange diffusion of chloride and bicarbonate is nevertheless a rate-limiting step for the transfer of CO2 from tissues to lungs. Measurements of chloride and bicarbonate self-exchange form the basis for calculations that demonstrate that the ionic exchange processes cannot run to complete equilibration at capillary transit times less than 0.5 s. The anion exchange diffusion is mediated by a large transmembrane protein constituting almost 30% of the total membrane protein. The kinetics of exchange diffusion must depend on conformational changes of the protein molecule, associated with the binding and subsequent translocation of the transported anion. We have characterized the nature of anion-binding sites facing the extracellular medium by acid-base titration of the transport function and modification of the transport protein in situ with group-specific amino acid reagents. Anion binding and translocation depend on the integrity and the degree of protonation of two sets of exofacial groups with apparent pK values of 12 and 5, respectively. From the chemical reactivities towards amino acid reagents it appears that the groups whose pK = 12 are guanidino groups of arginyl residues, while the groups whose pK = 5 are likely to be carboxylates of glutamic or aspartic acid. Our studies suggest that the characteristics of anion recognition sites in water-soluble proteins and in the integral transport proteins are closely related.     

Mathematical modelling of charge properties of protein molecules]

A procedure of theoretical determination of the dependence of protein molecule charge on the medium pH has been developed. The suggested procedure allows calculating the protein pI value, the molecule charge at the definite pH value, as well as the corresponding values for the protein molecule. Calculations for insulin, apo A-I and apo A-II molecules have been carried out. Calculated pI values are equal to 5.25, 5.64 and 4.86, respectively. A comparison of the theoretical curves and experimental data allows obtaining information of the molecule structure. Carboxyl groups with abnormally high pK values are discovered, that, probably, indicates to the direct interaction of two COOH-groups. A supposition is made on the most probable arrangement of the functional fragments in apo A-I and apo A-II molecules.     

Proton titration curve of yeast iso-1-ferricytochrome c. Electrostatic and conformational effects of point mutations.

The proton titration curves of yeast iso-1-ferricytochrome c and selected point mutants of this protein have been determined between pH 3 and 11 at 10 and 25 degrees C with a computer-controlled titration system. Initial titration of the wild-type protein to acidic pH followed by subsequent titrations to alkaline and then acidic pH demonstrates hysteresis, with one more group (28.7) titrating between pH 11 and 3 than originally titrated (27.7) between pH 3 and 11. Initial titration to alkaline pH, however, resulted in observation of the same number of groups in both directions of titration (28.7 vs 28.6). At 10 degrees C, 7.5 fewer groups were found to titrate over the same range of pH. Titration curves obtained for six cytochrome c mutants modified at Arg-38, Phe-82, Tyr-48, and Tyr-67 were analyzed by subtraction of the corresponding titration curve for the wild-type protein to produce difference titration curves. In most cases, the effects of these mutations as revealed in the difference titration curves could be accounted for as either the result of introduction of an additional group titrating within this pH range, the result of a change in the pK of a titrating residue, and/or the result of a change in the pK for either the first acidic or the first alkaline protein conformational transition. In addition to demonstration of the electrostatic consequences of the mutations in cytochrome c studied here, this study establishes the general usefulness of precise proton titration curve analysis in the characterization of variant proteins produced through recombinant genetic techniques.     

Molecular dynamics effects on protein electrostatics

Electrostatic calculations have been carried out on a number of structural conformers of tuna cytochrome c. Conformers were generated using molecular dynamics simulations with a range of solvent simulating, macroscopic dielectric formalisms, and one solvent model that explicitly included solvent water molecules. Structures generated using the lowest dielectric models were relatively tight, with side chains collapsed on the surface, while those from the higher dielectric models had more internal and external fluidity, with surface side chains exploring a fuller range of conformational space. The average structure generated with the explicitly solvated model corresponded most closely with the crystal structure. Individual pK values, overall titration curves, and electrostatic potential surfaces were calculated for average structures and structures along each simulation. Differences between structural conformers within each simulation give rise to substantial changes in calculated local electrostatic interactions, resulting in pK value fluctuations for individual sites in the protein that vary by 0.3-2.0 pK units from the calculated time average. These variations are due to the thermal side chain reorientations that produce fluctuations in charge site separations. Properties like overall titration curves and pH dependent stability are not as sensitive to side chain fluctuations within a simulation, but there are substantial effects between simulations due to marked differences in average side chain behavior. These findings underscore the importance of proper dielectric formalism in molecular dynamics simulations when used to generate alternate solution structures from a crystal structure, and suggest that conformers significantly removed from the average structure have altered electrostatic properties that may prove important in episodic protein properties such as catalysis.     

Engineering the pH-dependence of kinetic parameters of maize glutathione S-transferase I by site-directed mutagenesis

The optimisation of enzymes for particular application or conditions remains an important target in all protein engineering endeavours. Here, we report a successful strategy for altering the pH-profile of kinetic parameters and to define in detail the molecular mechanism of maize glutathione S-transferase I (GST I). To accomplish this, selected residues from the glutathione binding site (His40, Ser11, Lys41, Asn49, Gln53 and Ser67) were mutated to Ala, and the pH-dependence of the catalytic parameters V(max), and V(max)/K(GSH)(m) of the mutated forms were analysed. The pH-dependence of V(max) for the wild-type enzyme exhibits two transitions in the acidic pH range with pK(a1) of 5.7 and pK(a2) of 6.6. Based on thermodynamic data, site-directed mutagenesis and UV deference spectroscopy, it was concluded that pK(a1) corresponds to GSH carboxylates, whereas the pK(a2) has a conformational origin of the protein. The pH-dependence of V(max)/K(GSH)(m) for the wild-type enzyme exhibits a single transition with pK(a) of 6.28 which was attributed to the thiol ionisation of bound GSH. These findings complement the conclusions about the catalytic mechanism deduced from the crystal structure of the enzyme and provide the basis for rationally designing engineered forms of GST I with valuable properties.     

Ion-pair formation as a source of enhanced reactivity of the essential thiol group of D-glyceraldehyde-3-phosphate dehydrogenase.

The reactivity and the mode of activation of the essential--SH group (Cys-149) of D-glyceraldehyde-3-phosphate dehydrogenase have been studied by means of a spectrophotometric method [Polgár, L., FEBS Lett. 38, 187-190 (1974)], capable of detecting the dissociated form of the thiol group in proteins. Alkylations of Cys-149 of NAD-free D-glyceraldehyde-3-phosphate dehydrogenase with iodoacetamide and iodoacetate were investigated. The corrected absorbance change on alkylation at 250 nm (which is a direct parameter of the dissociation of the thiol group) and the alkylation rate were determined as a function of pH. The pH profiles of both dissociation and alkylation rate of Cys-149 conform to doubly sigmoid curves. All these curves implicate two ionizing groups (pK1 equals 5.5, pK2 equals 8.2). It is concluded that there are two reactive forms of the--SH group in the apoenzyme between pH 5 and 10. One reactive form corresponds to the free mercaptide ion. The other can be identified with an ion-pair composed of a mercaptide ion and some base, possibly the imidazolium group of His-176. The ion-pair has lower molar absorption coefficient and nucleophilicity than the free mercaptide ion. The two reactive forms are transformed into each other with pK2 equals 8.2. The ion-pair decomposes to a nondissociated thiol group and a protonated base with pK1 equals 5.5. In the presence of NAD, only the pH-rate profile of alkylation of D-glyceraldehyde-3-phosphate dehydrogenase was measured (at 370 nm). Using iodoacetamide as alkylating agent we also obtained a doubly sigmoid curve. A slight downward shift on pK1 and an upward shift in pK2 indicate that the ion-pair exists in a somewhat wider pH-range in the enzyme-coenzyme complex. An increase in the ionic strength of the reaction mixture from 0.09 to 0.45 M does not abolish the doubly sigmoid character of the curves determined either in the presence or in the absence of NAD.     

pK values of histidine residues in ribonuclease Sa: effect of salt and net charge

The primary goal of this study was to gain a better understanding of the effect of environment and ionic strength on the pK values of histidine residues in proteins. The salt-dependence of pK values for two histidine residues in ribonuclease Sa (RNase Sa) (pI=3.5) and a variant in which five acidic amino acids have been changed to lysine (5K) (pI=10.2) was measured and compared to pK values of model histidine-containing peptides. The pK of His53 is elevated by two pH units (pK=8.61) in RNase Sa and by nearly one pH unit (pK=7.39) in 5K at low salt relative to the pK of histidine in the model peptides (pK=6.6). The pK for His53 remains elevated in 1.5M NaCl (pK=7.89). The elevated pK for His53 is a result of screenable electrostatic interactions, particularly with Glu74, and a non-screenable hydrogen bond interaction with water. The pK of His85 in RNase Sa and 5K is slightly below the model pK at low salt and merges with this value at 1.5M NaCl. The pK of His85 reflects mainly effects of long-range Coulombic interactions that are screenable by salt. The tautomeric states of the neutral histidine residues are changed by charge reversal. The histidine pK values in RNase Sa are always higher than the pK values in the 5K variant. These results emphasize that the net charge of the protein influences the pK values of the histidine residues. Structure-based pK calculations capture the salt-dependence relatively well but are unable to predict absolute histidine pK values.     

1H nuclear-magnetic-resonance studies of porcine lutropin and its alpha and beta subunits.

The titration curves of the histidine residues of porcine lutropin and its isolated alpha and beta subunits have been determined by following the pH-dependence of the imidazole C-2 proton resonances. The isolated alpha subunit contains a buried histidine, whose C-2 proton does not exchange with solvent, and which has the unusually low pK of 3.3. In the native hormone all the histidine residues have relatively normal pK values (between 5.7 and 6.2). The four histidine C-2 proton resonances have been assigned to specific residues in the amino-acid sequence, by means of deuterium and tritium exchange experiments on the alpha subunit and its des(92-96) derivative. The histidine with a pK of 3.3 is identified as His-alpha87. The effects of pH on tyrosine and methyl proton resonances show that the titration of His-87 in the isolated alpha subunit is accompanied by a significant conformational change which involves loosening of the protein structure but which is not a normal unfolding transition. The role of conformational changes in the generation of biological activity by subunit association in the glycoprotein hormones is discussed.     

Role of calcium as trigger in thermal beta-lactoglobulin aggregation

Divalent calcium ions have been suggested to be involved in intermolecular protein-Ca2+-protein cross-linking, intramolecular electrostatic shielding, or ion-induced protein conformational changes as a trigger for protein aggregation at elevated temperatures. To address the first two phenomena in the case of beta-lactoglobulin, a combination of chemical protein modification, calcium-binding, and aggregation studies was used, while the structural integrity of the modified proteins was maintained. Although increasing the number of carboxylates on the protein by succinylation results in improved calcium-binding, calcium appears to be less effective in inducing protein aggregation. In fact, the larger the number of carboxylates, the higher the concentration of calcium that is required to trigger the aggregation. Lowering the number of negative charges on the protein surface via methylation of carboxylates reduces calcium-binding properties, but calcium-induced aggregation at low concentration is improved. Monovalent sodium ions cannot take over the specific role of calcium. The relation between net surface charge and number of calcium ions bound required to trigger the aggregation suggests that calcium needs to bind site specific to carboxylates with a threshold affinity. Subsequent site-specific screening of surface charges results in protein aggregation, driven by the partial unfolding of the protein at elevated temperatures, which is then facilitated by the absence of electrostatic repulsion.     

Curves of ligand binding. The use of hyperbolic functions for expressing titration curves

1. The dependences of the concentrations of the non-ligated, uni-ligated and bi-ligated forms of a molecule that binds two molecules of ligand are expressed as functions of the logarithm of free ligand concentration by means of hyperbolic functions. Expressions are also given for the saturation both of an individual site and of the molecule as a whole. This form of expression allows derivation of the following points. 2. The sharpness of bell-shaped curves of concentration of the uni-ligated form is analysed in terms of the heights of their points of inflexion; these can rise to 1/ radical2 of the curve. 3. A single group can exhibit a doubly sigmoid saturation curve if this group and another have comparable affinities for a ligand, and if ligand binding at one of them diminishes the affinity at the other. If the molecular pK values pK(1) and pK(2) for the first and second molecules of ligand are called pK*+/-logm, so that K*(2)=K(1)K(2) and m(2)=K(1)/K(2), then the doubly sigmoid curve can be represented by the sum of two independent one-site saturation curves, in general of unequal height, of pK values pK*+/-log(1/2)[m+ radical(m(2)-4)]. The error in such representation is small either if the mutual interaction between the groups (i.e. m) is large, or if the groups have very similar affinities for the ligand. 4. The sum of two one-site saturation curves, again of pK values of pK*+/-log(1/2)[m+ radical(m(2)-4)] but of equal heights, gives a precise value for the total saturation, provided that the binding of one molecule does not promote the binding of a second, i.e. providing that m>/=2. Hence determinations of saturation cannot distinguish interacting and possibly identical sites from independent and different ones.     

The transfer of platelet factor 4 from its proteoglycan carrier to natural and synthetic polymers

1. Transfer of dansyl-platelet factor 4 complexed with a series of glycosaminoglycans to heparin has been detected and studied by measuring changes in the anisotropy of the dansyl fluorescence. The protein was most easily transferred from chondroitin sulphate and least easily from heparan sulphaet. 2. The transfer of the dye-labelled protein from its biological chondroitin 4-sulphate proteoglycan carrier to natural and synthetic anionic polymers was similarly followed. The transfer to heparin and dermatan sulphate was shown to be the same whether 3 mM Ca²⁺ or 8 mM EGTA was present in the solution. 3. The shapes of the binding curves of the dansyl-factor to the polymers have been compared at I = 0.4 M. 4. The observed changes in anisotropy of dye fluorescence have been correlated with the charge density and the stereochemistry of the charged groups of the natural polymers. Large complexes are observed with polymers of high negative charge/weight ratios. Less charged polymers containing disaccharide units of iduronic acid and glucosamine N-sulphate will also form large complexes at I = 0.15 M. 5. It is demonstrated that the release of a platelet factor 4 proteoglycan complex in vivo would result in the transfer of the protein to heparin, moderate quantities of either dermatan or heparan sulphates would not prevent this transfer.     

Studies of the histidine residues of human and bovine glycoprotein hormones by nuclear magnetic resonance

Titration curves of the histidine residues in lutropin, thyrotropin, follitropin and chorionic gonadotropin have been assigned using imidazole C-2 proton nuclear magnetic resonance spectra and their estimated pK values determined. Spectra of reassociated hormone preparations, in which one or the other of their two subunits (alpha or beta) have had their accessible histidines exchanged with deuterium, permitted assignment of C-2 resonance to specific residues. Similar titration curves were found for residues which are conserved from one hormone to another. However, these conserved histidines do not have identical pK values, indicating that differences in the conformation or microenvironment around these residues occur in these hormones. Changes in some pK values also occur as a function of subunit association. The most dramatic change seen in all cases is the exposure to solvent of histidine alpha-83; in isolated alpha subunits this residue is unavailable for titration over a wide pH range. This change appears to be a general consequence of the association of the two subunits in any of these hormones. The data show that all histidines in the intact hormones are accessible to the environment, including those proposed to be in domains involved in subunit-subunit interaction.