Roles of distal arginine in activity and stability of Coprinus cinereus peroxidase elucidated by kinetic and NMR analysis of the Arg51Gln, -Asn, -Leu, and -Lys mutants

In heme peroxidases, a distal His residue plays an essential role in the initial two electron oxidation of resting state enzyme to compound I by hydrogen peroxide. A distal Arg residue assists in this process. The contributions of the charge, H-bonding capacity, size, and mobility of this Arg residue to Coprinus cinereus peroxidase (CIP) reactivity and stability have been examined by substituting Arg51 with Gln (retains H-bond donor at N epsilon position), Asn (small size, H-bond donor and acceptor), Leu (similar to Asn, but hydrophobic), and Lys (charge and H-bond donor, but at N zeta position). UV-visible spectroscopy was used to monitor pH-linked heme changes, compound I formation and reduction, fluoride binding, and thermostability. (1)H NMR spectroscopy enabled heme pocket differences in both resting and cyanide-ligated states of the enzymes to be evaluated and compared with wild-type CIP. We found that the H-bonding capacity of distal Arg is key to fast compound I formation and ligand binding to heme, whereas charge is important for lowering the pK(a) of distal His and for the binding and stabilisation of anionic ligands at heme iron. The properties of the distal Arg residue in CIP, cytochrome c peroxidase (CCP) and horseradish peroxidase (HRP) differ significantly in their pH induced transitions and dynamics.     

Homology modeling of a heme protein, lignin peroxidase, from the crystal structure of cytochrome c peroxidase.

A 3-dimensional model of lignin peroxidase (LiP) was constructed based on its sequence homology with other peroxidases, particularly cytochrome c peroxidase, the only protein with a known crystal structure in the peroxidase family. The construction of initial conformations of insertions and deletions was assisted by secondary structure predictions, amphipathic helix predictions, and consideration of the specific protein environment. A succession of molecular dynamics simulations of these regions with surrounding residues as constraints were carried out to relax the bond lengths and angles. Full protein molecular dynamics simulations with explicit consideration of bound waters were performed to relax the geometry and to identify dynamically flexible regions of the successive models for further refinement. Among the important functionally relevant structural features predicted are: (i) four disulfide bonds are predicted to be formed between Cys3 and Cys15, Cys14 and Cys285, Cys34 and Cys120 and Cys249 and Cys317; (ii) a glycosylation site, Asn257, was located on the surface; (iii) Glu40 was predicted to form a salt bridge with Arg43 on the distal side of the heme and was considered as a possible origin for the pH dependence of compound I formation; and (iv) two candidate substrate binding sites with a cluster of surface aromatic residues and flexible backbones were found in the refined model, consistent with the nature of known substrates of LiP. Based on these predicted structural features of the model, further theoretical and experimental studies are proposed to continue to elucidate the structure and function of LiP.     

Constant pH molecular dynamics with proton tautomerism

The current article describes a new two-dimensional lambda-dynamics method to include proton tautomerism in continuous constant pH molecular dynamics (CPHMD) simulations. The two-dimensional lambda-dynamics framework is used to devise a tautomeric state titration model for the CPHMD simulations involving carboxyl and histidine residues. Combined with the GBSW implicit solvent model, the new method is tested on titration simulations of blocked histidine and aspartic acid as well as two benchmark proteins, turkey ovomucoid third domain (OMTKY3) and ribonuclease A (RNase A). A detailed analysis of the errors inherent to the CPHMD methodology as well as those due to the underlying solvation model is given. The average absolute error for the computed pKa values in OMTKY3 is 1.0 pK unit. In RNase A the average absolute errors for the carboxyl and histidine residues are 1.6 and 0.6 pK units, respectively. In contrast to the previous work, the new model predicts the correct sign for all the pKa shifts, but one, in the benchmark proteins. The predictions of the tautomeric states of His12 and His48 and the conformational states of His48 and His119 are in agreement with experiment. Based on the simulations of OMTKY3 and RNase A, the current work has demonstrated the capability of the CPHMD technique in revealing pH-coupled conformational dynamics of protein side chains.     

Exploring the role of the active site cysteine in human muscle creatine kinase

All known guanidino kinases contain a conserved cysteine residue that interacts with the non-nucleophilic eta1-nitrogen of the guanidino substrate. Site-directed mutagenesis studies have shown that this cysteine is important, but not essential for activity. In human muscle creatine kinase (HMCK) this residue, Cys283, forms part of a conserved cysteine-proline-serine (CPS) motif and has a pKa about 3 pH units below that of a regular cysteine residue. Here we employ a computational approach to predict the contribution of residues in this motif to the unusually low cysteine pKa. We calculate that hydrogen bonds to the hydroxyl and to the backbone amide of Ser285 would both contribute approximately 1 pH unit, while the presence of Pro284 in the motif lowers the pKa of Cys283 by a further 1.2 pH units. Using UV difference spectroscopy the pKa of the active site cysteine in WT HMCK and in the P284A, S285A, and C283S/S285C mutants was determined experimentally. The pKa values, although consistently about 0.5 pH unit lower, were in broad agreement with those predicted. The effect of each of these mutations on the pH-rate profile was also examined. The results show conclusively that, contrary to a previous report (Wang et al. (2001) Biochemistry 40, 11698-11705), Cys283 is not responsible for the pKa of 5.4 observed in the WT V/K(creatine) pH profile. Finally we use molecular dynamics simulations to demonstrate that, in order to maintain the linear alignment necessary for associative inline transfer of a phosphoryl group, Cys283 needs to be ionized.     

Role of Tyr84 in controlling the reactivity of Cys34 of human albumin

Cys34 in domain I of the three-domain serum protein albumin is the binding site for a wide variety of biologically and clinically important small molecules, provides antioxidant activity, and constitutes the largest portion of free thiol in blood. Analysis of X-ray structures of albumin reveals that the loop containing Tyr84 occurs in multiple conformations. In structures where the loop is well defined, there appears to be an H-bond between the OH of Tyr84 and the sulfur of Cys34. We show that the reaction of 5,5'-dithiobis(2-nitrobenzoic acid) (DTNB) with Tyr84Phe mutant albumin is approximately four times faster than with the wild-type protein between pH 6 and pH 8. In contrast, the His39Leu mutant reacts with DTNB more slowly than the wild-type protein at pH < 8, but at a similar rate at pH 8. Above pH 8 there is a dramatic increase in reactivity for the Tyr84Phe mutant. We also report (1)H NMR studies of disulfide interchange reactions with cysteine. The tethering of the two loops containing Tyr84 and Cys34 not only appears to control the redox potential and accessibility of Cys34, but also triggers the transmission of information about the state of Cys34 throughout domain I, and to the domainI/II interface.     

High-accuracy prediction of protein structural class for low-similarity sequences based on predicted secondary structure

Information on the structural classes of proteins has been proven to be important in many fields of bioinformatics. Prediction of protein structural class for low-similarity sequences is a challenge problem. In this study, 11 features (including 8 re-used features and 3 newly-designed features) are rationally utilized to reflect the general contents and spatial arrangements of the secondary structural elements of a given protein sequence. To evaluate the performance of the proposed method, jackknife cross-validation tests are performed on two widely used benchmark datasets, 1189 and 25PDB with sequence similarity lower than 40% and 25%, respectively. Comparison of our results with other methods shows that our proposed method is very promising and may provide a cost-effective alternative to predict protein structural class in particular for low-similarity datasets.     

NADH interactions with WT- and S94A-acyl carrier protein reductase from Mycobacterium tuberculosis: an ab initio study

We present an ab initio molecular dynamics study of the complex between acyl carrier protein reductase InhA from M. tuberculosis and isonicotinic acid hydrazide-NADH. We focus on wild-type (WT) InhA and a mutant causing drug resistance (S94A) for which structural information is available (Rozwarski et al., 1998;279:98--102; Dessen et al., 1995;267:1638--1641). Our calculations suggest that the water-mediated H-bond interactions between Ser94 side chain and NADH, present in WT InhA X-ray structure, can be lost during the dynamics. This conformational change is accompanied by a structural rearrangement of Gly14. The calculated structure of WT is rather similar to the X-ray structure of the S94A mutant in terms of geometrical parameters and chemical bonding. Further evidence for the mobility of Ser94 is provided by a 1-ns-long classical molecular dynamics on the entire protein. The previously unrecognized high mobility of Ser94 can provide a rationale of the small change in free binding energies on passing from WT to S94A InhA.     

Characterization of Escherichia coli thioredoxin variants mimicking the active-sites of other thiol/disulfide oxidoreductases.

Thiol/disulfide oxidoreductases like thioredoxin, glutaredoxin, DsbA, or protein disulfide isomerase (PDI) share the thioredoxin fold and a catalytic disulfide bond with the sequence Cys-Xaa-Xaa-Cys (Xaa corresponds to any amino acid). Despite their structural similarities, the enzymes have very different redox properties, which is reflected by a 100,000-fold difference in the equilibrium constant (K(eq)) with glutathione between the most oxidizing member, DsbA, and the most reducing member, thioredoxin. Here we present a systematic study on a series of variants of thioredoxin from Escherichia coli, in which the Xaa-Xaa dipeptide was exchanged by that of glutaredoxin, PDI, and DsbA. Like the corresponding natural enzymes, all thioredoxin variants proved to be stronger oxidants than the wild-type, with the order wild-type < PDI-type < DsbA-type < glutaredoxin-type. The most oxidizing, glutaredoxin-like variant has a 420-fold decreased value of K(eq), corresponding to an increase in redox potential by 75 mV. While oxidized wild-type thioredoxin is more stable than the reduced form (delta deltaG(ox/red) = 16.9 kJ/mol), both redox forms have almost the same stability in the variants. The pH-dependence of the reactivity with the alkylating agent iodoacetamide proved to be the best method to determine the pKa value of thioredoxin's nucleophilic active-site thiol (Cys32). A pKa of 7.1 was measured for Cys32 in the reduced wild-type. All variants showed a lowered pKa of Cys32, with the lowest value of 5.9 for the glutaredoxin-like variant. A correlation of redox potential and the Cys32 pKa value could be established on a quantitative level. However, the predicted correlation between the measured delta deltaG(ox/red) values and Cys32 pKa values was only qualitative.     

H-bonding maintains the active site of type 1 copper proteins: site-directed mutagenesis of Asn38 in poplar plastocyanin

Type I Cu proteins maintain a trigonal N2S coordination group (with weak axial ligation) in both oxidation states of the Cu2+/+ ion, thereby reducing the reorganization energy for electron transfer. Requirements for maintaining this coordination group were investigated in poplar plastocyanin (Pcy) by mutation of a conserved element of the type 1 architecture, an asparagine residue (Asn38) adjacent to one of the ligating histidines. The side chain of this asparagine forms an active site clasp via two H-bonds with the residue (Ser85) adjacent to the ligating cysteine (Cys84). In addition, the main chain NH of Asn38 donates an H-bond to the thiolate ligand. We have investigated the importance of these interactions by mutating Asn38 to Gln, Thr, and Leu. The mutant proteins are capable of folding and binding Cu2+, but the blue color fades; the rate of fading increases in the order Gln < Thr < Leu. The color is not restored by ferricyanide, showing that the protein is modified irreversibly, probably by oxidation of Cys84. The more stable mutants N38Q and N38T were characterized spectroscopically. The wild-type properties are slightly perturbed for N38Q, but N38T shows remarkable similarity to another type 1 Cu protein, azurin (Azu) from Pseudomonas aeruginosa. The Cu-S(Cys) bond is longer in Azu than in Pcy, and the NH H-bond to the ligating S atom is shorter. Molecular modeling suggests a similar effect for N38T because the threonine residue shifts toward Ser85 in order to avoid a steric clash and to optimize H-bonding. These results demonstrate that H-bonding adjacent to the type 1 site stabilizes an architecture which both modulates the electronic properties of the Cu, and suppresses side reactions of the cysteine ligand.     

Long-term molecular dynamics simulation of copper plastocyanin in water

A long molecular dynamics simulation (1.1 ns) of fully hydrated plastocyanin has been performed and analysed to relate protein dynamics to structural elements and functional properties. The solvated structure is described in detail by the analysis of H-bond network. During all the simulation, the crystal H-bond network is maintained in the beta-sheet regions, while several H-bonds are broken or formed on the external surface of the protein. To evaluate whether such changes could be due to conformational rearrangements or to solvent competition, we have examined the average number of H-bonds between protein atoms and water molecules, and the root mean square deviations from crystal structure as a function of protein residues. Protein mobility and flexibility have been examined by positional and dihedral angle rms fluctuations. Finally, cross-correlation maps have revealed the existence of correlated motions among residues connected by hydrogen bonds.     

The Interaction of the Chemotherapeutic Drug Chlorambucil with Human Glutathione Transferase A1-1: Kinetic and Structural Analysis

Glutathione transferases (GSTs) are enzymes that contribute to cellular detoxification by catalysing the nucleophilic attack of glutathione (GSH) on the electrophilic centre of a number of xenobiotic compounds, including several chemotherapeutic drugs. In the present work we investigated the interaction of the chemotherapeutic drug chlorambucil (CBL) with human GSTA1-1 (hGSTA1-1) using kinetic analysis, protein crystallography and molecular dynamics. In the presence of GSH, CBL behaves as an efficient substrate for hGSTA1-1. The rate-limiting step of the catalytic reaction between CBL and GSH is viscosity-dependent and kinetic data suggest that product release is rate-limiting. The crystal structure of the hGSTA1-1/CBL-GSH complex was solved at 2.1 Å resolution by molecular replacement. CBL is bound at the H-site attached to the thiol group of GSH, is partially ordered and exposed to the solvent, making specific interactions with the enzyme. Molecular dynamics simulations based on the crystal structure indicated high mobility of the CBL moiety and stabilization of the C-terminal helix due to the presence of the adduct. In the absence of GSH, CBL is shown to be an alkylating irreversible inhibitor for hGSTA1-1. Inactivation of the enzyme by CBL followed a biphasic pseudo-first-order saturation kinetics with approximately 1 mol of CBL per mol of dimeric enzyme being incorporated. Structural analysis suggested that the modifying residue is Cys112 which is located at the entrance of the H-site. The results are indicative of a structural communication between the subunits on the basis of mutually exclusive modification of Cys112, indicating that the two enzyme active sites are presumably coordinated.     

Conformation, stability, and active-site cysteine titrations of Escherichia coli D26A thioredoxin probed by Raman spectroscopy.

The active-site cysteines (Cys 32 and Cys 35) of Escherichia coli thioredoxin are oxidized to a disulfide bridge when the protein mediates substrate reduction. In reduced thioredoxin, Cys 32 and Cys 35 are characterized by abnormally low pKa values. A conserved side chain, Asp 26, which is sterically accessible to the active site, is also essential to oxidoreductase activity. pKa values governing cysteine thiol-thiolate equilibria in the mutant thioredoxin, D26A, have been determined by direct Raman spectrophotometric measurement of sulfhydryl ionizations. The results indicate that, in D26A thioredoxin, both sulfhydryls titrate with apparent pKa values of 7.5+/-0.2, close to values measured previously for wild-type thioredoxin. Sulfhydryl Raman markers of D26A and wild-type thioredoxin also exhibit similar band shapes, consistent with minimal differences in respective cysteine side-chain conformations and sulfhydryl interactions. The results imply that neither the Cys 32 nor Cys 35 SH donor is hydrogen bonded directly to Asp 26 in the wild-type protein. Additionally, the thioredoxin main-chain conformation is largely conserved with D26A mutation. Conversely, the mutation perturbs Raman bands diagnostic of tryptophan (Trp 28 and Trp 31) orientations and leads to differences in their pH dependencies, implying local conformational differences near the active site. We conclude that, although the carboxyl side chain of Asp 26 neither interacts directly with active-site cysteines nor is responsible for their abnormally low pKa values, the aspartate side chain may play a role in determining the conformation of the enzyme active site.     

NetCTLpan: pan-specific MHC class I pathway epitope predictions

Reliable predictions of immunogenic peptides are essential in rational vaccine design and can minimize the experimental effort needed to identify epitopes. In this work, we describe a pan-specific major histocompatibility complex (MHC) class I epitope predictor, NetCTLpan. The method integrates predictions of proteasomal cleavage, transporter associated with antigen processing (TAP) transport efficiency, and MHC class I binding affinity into a MHC class I pathway likelihood score and is an improved and extended version of NetCTL. The NetCTLpan method performs predictions for all MHC class I molecules with known protein sequence and allows predictions for 8-, 9-, 10-, and 11-mer peptides. In order to meet the need for a low false positive rate, the method is optimized to achieve high specificity. The method was trained and validated on large datasets of experimentally identified MHC class I ligands and cytotoxic T lymphocyte (CTL) epitopes. It has been reported that MHC molecules are differentially dependent on TAP transport and proteasomal cleavage. Here, we did not find any consistent signs of such MHC dependencies, and the NetCTLpan method is implemented with fixed weights for proteasomal cleavage and TAP transport for all MHC molecules. The predictive performance of the NetCTLpan method was shown to outperform other state-of-the-art CTL epitope prediction methods. Our results further confirm the importance of using full-type human leukocyte antigen restriction information when identifying MHC class I epitopes. Using the NetCTLpan method, the experimental effort to identify 90% of new epitopes can be reduced by 15% and 40%, respectively, when compared to the NetMHCpan and NetCTL methods. The method and benchmark datasets are available at .     

Molecular basis of intramolecular electron transfer in proteins during radical-mediated oxidations: computer simulation studies in model tyrosine-cysteine peptides in solution

Experimental studies in hemeproteins and model Tyr/Cys-containing peptides exposed to oxidizing and nitrating species suggest that intramolecular electron transfer (IET) between tyrosyl radicals (Tyr-O(·)) and Cys residues controls oxidative modification yields. The molecular basis of this IET process is not sufficiently understood with structural atomic detail. Herein, we analyzed using molecular dynamics and quantum mechanics-based computational calculations, mechanistic possibilities for the radical transfer reaction in Tyr/Cys-containing peptides in solution and correlated them with existing experimental data. Our results support that Tyr-O(·) to Cys radical transfer is mediated by an acid/base equilibrium that involves deprotonation of Cys to form the thiolate, followed by a likely rate-limiting transfer process to yield cysteinyl radical and a Tyr phenolate; proton uptake by Tyr completes the reaction. Both, the pKa values of the Tyr phenol and Cys thiol groups and the energetic and kinetics of the reversible IET are revealed as key physico-chemical factors. The proposed mechanism constitutes a case of sequential, acid/base equilibrium-dependent and solvent-mediated, proton-coupled electron transfer and explains the dependency of oxidative yields in Tyr/Cys peptides as a function of the number of alanine spacers. These findings contribute to explain oxidative modifications in proteins that contain sequence and/or spatially close Tyr-Cys residues.     

Cy‐preds: An algorithm and a web service for the analysis and prediction of cysteine reactivity

Cysteine (Cys) is a critically important amino acid, serving a variety of functions within proteins including structural roles, catalysis, and regulation of function through post‐translational modifications. Predicting which Cys residues are likely to be reactive is a very sought after feature. Few methods are currently available for the task, either based on evaluation of physicochemical features (e.g., pKa and exposure) or based on similarity with known instances. In this study, we developed an algorithm (named HAL‐Cy) which blends previous work with novel implementations to identify reactive Cys from nonreactive. HAL‐Cy present two major components: (i) an energy based part, rooted on the evaluation of H‐bond network contributions and (ii) a knowledge based part, composed of different profiling approaches (including a newly developed weighting matrix for sequence profiling). In our evaluations, HAL‐Cy provided significantly improved performances, as tested in comparisons with existing approaches. We implemented our algorithm in a web service (Cy‐preds), the ultimate product of our work; we provided it with a variety of additional features, tools, and options: Cy‐preds is capable of performing fully automated calculations for a thorough analysis of Cys reactivity in proteins, ranging from reactivity predictions (e.g., with HAL‐Cy) to functional characterization. We believe it represents an original, effective, and very useful addition to the current array of tools available to scientists involved in redox biology, Cys biochemistry, and structural bioinformatics. Proteins 2016; 84:278–291. © 2015 Wiley Periodicals, Inc.     

Effect of pH on the semiquinone radical Q(A)- in CN-treated photosystem II: study by hyperfine sublevel correlation spectroscopy

The semiquinone radical Q(A)- has been studied by electron spin echo envelope modulation (ESEEM) spectroscopy in Photosystem II membranes treated with CN- at various pH values. Two protein 14N nuclei (N(I) and N(II)) were found to be magnetically coupled with the Q(A)- spin. N(I) is assigned to an amide nitrogen from the protein backbone while N(II) is assigned to the amino nitrogen, N(epsilon), of an imidazole. Above pH 8.5 only the N(I) coupling is present while both N(I) and N(II) couplings are present at lower pH values. These results are interpreted in terms of a model based on the structure of the bacterial reaction center and involving two determining factors. First, the non-heme iron, when present, is ligated to the imidazole that H-bonds to one of the Q(A)- carbonyls. This physical attachment of the imidazole to the iron limits the strength of the H-bond to Q(A)-. Second, a pH-dependent group on the protein controls the strength of the H-bonds to Q(A)-. The pKa of this group is around pH 7.5 in CN(-)-treated PSII.     

Structural basis of human erythrocyte glucose transporter function: pH effects on intrinsic fluorescence.

The effects of pH on the intrinsic fluorescence of purified human erythrocyte glucose transporter (HEGT) were studied to deduce the structure and the ligand-induced dynamics of this protein. D-Glucose increases tryptophan fluorescence of HEGT at a 320-nm peak with a concomitant reduction in a 350-nm peak, suggesting that glucose shifts a tryptophan residue from a polar to a nonpolar environment. Cytochalasin B or forskolin, on the other hand, only produces a reduction at the 350-nm peak. The pH titration of the intrinsic fluorescence of HEGT revealed that at least two tryptophan residues are quenched, one with a pKa of 5.5, the other with a pKa of 8.2, indicating involvement of histidine and cysteine protonation, respectively. D-Glucose abolishes both of these quenchings. Cytochalasin B or forskolin, on the other hand, abolishes the histidine quenching but not the cysteine quenching and induces a new pH quenching with a pKa of about 4, implicating involvement of a carboxyl group. These results, together with the known primary structure and the transmembrane disposition of this protein, predict the dynamic interactions between Trp388 and His337, Trp412 and Cys347, and Trp412 and Glu380, depending on liganded state of HEGT, and suggest the importance of the transmembrane helices 9, 10, and 11 in transport function.     

Variability in the pKa of histidine side-chains correlates with burial within proteins

Acidic pKas of histidines buried within the protein interior are frequently rationalized on the contradictory basis of either polar interactions within the protein or the effects of a hydrophobic environment. To examine these relationships, we surveyed the buried surface area, depth of burial, polar interactions, and crystallographic temperature factors of histidines of known pKa. It has been found that buried environments of histidines do not always result in acidic pKas. Instead, the variability of histidine pKas increases for residues where the majority of the side-chain is buried. Because buried histidines are always found in mixed polar/apolar environments, multiple environmental contributions to pKa values must be considered. However, the quantitative relationships between heterogeneous environments and pKa values are not immediately apparent from the available data.     

The effect of chemical modification of Saccharomyces cerevisiae on electrophoretic mobility, cell-wall structure and amphotericin B uptake

Saccharomyces cerevisiae NCYC 239 suspended in solutions of NaCl showed two distinct plateaus in plots of electrophoretic mobility vs. pH, corresponding to pKa values of approx. 2 and 5. This is in contrast to cells suspended in buffer where only a single pKa (4) can be determined. Modification of cells with KI/I2 or nitrous acid led to altered electrophoretic mobility, indicating the presence of sulphydryl and amino groups, respectively, in the yeast cell surface, whereas uranyl nitrate modification had little effect, suggesting phosphate groups to be absent. Electron micrographs showed visible effects of KI/I2 and nitrous acid modification on cell membrane structure, and in these modified cells amphotericin B uptake was rapid. It is suggested that diffusion through the cell wall is the rate-limiting step for amphotericin B uptake. An activation energy of 20 kJ X mol-1 was determined for uptake of amphotericin B by unmodified cells.     

Electrostatic interaction between anions bound to site I and the retinal Schiff base of halorhodopsin

The influence of different anions on the deprotonation of the retinal Schiff base of halorhodopsin in the dark was investigated. We find that a large number of anions cause a significant increase of the pKa of the Schiff base, an effect attributed to binding to "site I" on the protein. The concentration dependencies of the spectroscopic shifts associated with the changes of the pKa yielded dissociation constants (and thus binding energies) for the anions, which were related to the Stokes radii. The data fit the predictions of electrostatic interaction between the anions and the positive charge associated with site I, if the latter is located within a few angstroms from the surface of the protein. The specificity of site I toward various anions is quantitatively explained by the differences in the change of Born energy upon transfer of the anions from water to the binding site. The changes in the deprotonation energy of the Schiff base upon the binding of anions, delta delta Gdeprot, could be calculated from the delta pKa at infinite anion concentration. Unexpectedly, the delta delta Gdeprot values were remarkably close to the energies of binding to site I. Thus, site I and the Schiff base are strongly electrostatically coupled, either because of close proximity or because of the possibility of allosteric energy transfer between them.