Role of salt bridge formation in antigen-antibody interaction. Entropic contribution to the complex between hen egg white lysozyme and its monoclonal antibody HyHEL10.

For elucidation of the role of salt bridge formation in the antigen-antibody complex, the interaction between hen egg white lysozyme (HEL) and its monoclonal antibody HyHEL10, the structure of which has been well characterized and forms one salt bridge (Lys97 of HEL and Asp32 of HyHEL10 heavy chain variable region (VH)), was investigated. Asp32 of VH was substituted with Ala, Asn, or Glu by site-directed mutagenesis, and the interaction between HEL and the mutant fragments of the variable region of light chain was investigated by inhibition of the enzymatic activity of HEL and isothermal titration calorimetry. Inhibition assay indicated that these mutations lowered the inhibition only slightly. Thermodynamic study indicated that the negative enthalpic change in the interaction between each of the mutant variable regions of light chain and HEL was significantly increased, although the association constant was slightly decreased, suggesting that these mutations increased the entropy change upon antigen-antibody binding. These results indicate that the role of salt bridge formation in the HyHEL10-HEL interaction is to lower the entropic loss due to binding. In the mutant proteins, the numbers of residues that were perturbed structurally on binding increased, suggesting that the salt bridge suppresses excess structural movement of the antibody upon binding.

A major goal of molecular biology is to describe biological phenomena in terms of interaction between molecules. Recently, thanks to progress in structural biology, several quaternary structures of biomolecular complexes have been clarified, and now a number of biological events can be described precisely at the atomic level. Among them, DNA-repressor (1), enzyme-inhibitor (2, 3) and antigen-antibody interactions (4,5) have been most extensively investigated at the atomic level.
Antigen-antibody interactions have been studied in detail not only from an immunological but also a biochemical viewpoint as a model of protein-ligand interaction. The most striking feature of antigen-antibody interactions is the creation of strict specificity and high affinity of antibodies for their antigens. It is known that unique conformations and special locations on the surfaces of proteinaceous antigens are recognized by antibodies (6). Among the amino acids located on protein surfaces, it is believed that antibodies recognize charged residues by forming (1) hydrogen bonds and (2) salt bridges and that salt bridge plays a significant role in the interaction (7)(8)(9).
For detailed elucidation of antigen-antibody interactions, a number of approaches have been attempted, including structural analysis (10 -13) and calculation and modeling (14 -16). However, thermodynamic analysis of antigen-antibody interaction for analyzing precisely the contribution of noncovalent forces to the interaction is most suitable (17)(18)(19)(20)(21)(22), especially when combined with site-directed mutagenesis of structurally well defined ones (23)(24)(25). We have focused on the interaction of hen egg white lysozyme (HEL 1 ) with the anti-HEL monoclonal antibody Hy-HEL10 interaction as a model of proteinaceous antigen-antibody interaction (26,27). This system is advantageous for studying the mechanism of antigen-antibody interaction because the structure of the antigen-antibody complex has already been clarified by x-ray crystallographic analysis (10), and also a secretory expression system for the Fv fragment, which is composed of variable domains of immunoglobulin heavy and light chains, has been established in Escherichia coli (28,29).
The side chain of Lys97 in HEL interacts with the contact region of HyHEL10 and is recognized by HAsp32, HTyr33, and HTrp95. HAsp32 contributes to the interaction by formation of a salt bridge with Lys97 of HEL (10). In this study, for elucidation of the role of the salt bridge in the antigen-antibody interaction, HAsp32 of the Fv fragment was replaced with Ala (HD32A), which deleted a carboxyl group; Glu (HD32E), which added a methylene group; or Asn (HD32N), which added an amino group. The interactions among antigen, HEL, and the engineered HyHEL10 Fvs were examined using the inhibition of HEL enzymatic activity and isothermal titration calorimetry. The contribution of the salt bridge to the interaction will be discussed thermodynamically here.

EXPERIMENTAL PROCEDURES
Materials-All enzymes used for genetic engineering were purchased from Toyobo, Takara Shuzo, or Boehringer Mannheim. Hen egg white lysozyme was from Seikagaku Kogyo, Inc. (Tokyo, Japan). Micrococcus lysodeikticus for measuring the enzymatic activity of HEL was obtained from Sigma. HEL-Sepharose was prepared from CNBr-activated Sepharose (Pharmacia). Oligonucleotide DNA primers were synthesized by a 381-Å DNA synthesizer (Applied Biosystems). A mutagenesis kit was obtained from Bio-Rad (Tokyo, Japan). All other reagents were of biochemical research grade.
Site-directed Mutagenesis of HAsp32-The method used for site-* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
ʈ To whom correspondence and requests for reprints should be addressed: Dept. of Biochemistry and Engineering, Graduate School of Engineering, Tohoku University, Aoba-ku, Sendai 980-77, Japan. Tel. and Fax: 81-22-217-7274. 1 The abbreviations used are: HEL, hen egg white lysozyme; HD32A, mutant Fv fragment in which Asp is substituted with Ala at position 32 of VH; Fv, fragment of variable region of immunoglobulin; VH, fragment of variable region of heavy chain; FKBP-12, FK506-binding protein.
directed mutagenesis was followed as described previously (27). Correctness of the mutation was confirmed by DNA sequencing using a Bca-BEST sequencing kit (Takara).
Expression and Purification of HAsp32-mutated Fv Fragments-Secretory expression in E. coli and the affinity chromatography purification procedures using HEL-Sepharose were performed as described previously (26).
Measurements of CD Spectra of Wild-type and Mutant HyHEL10 Fv Fragments-Measurements of CD spectra of the wild-type and mutant HyHEL10 Fv fragments (25 M) were carried out using a Jasco J-720 spectropolarimeter. Far-and near-UV CD spectra were measured using a cell with 1-and 10-mm optical path lengths, respectively, at pH 7.2 and 25°C.
Measurement of Inhibitory Activity toward HEL-Measurement of inhibition of HEL enzymatic activity was done as described previously (28). The experimental condition was described in the legend to Fig. 1.
Isothermal Titration Calorimetry-Thermodynamic parameters of the interactions between HEL and the mutant Fv fragments were determined by microtitration calorimetry using an OMEGA titration microcalorimeter (30) from MicroCal, Inc. (Northampton, MA). The experimental conditions were the same as those reported previously (27); The wild-type HEL at a concentration of 5 M in 50 mM phosphate buffer (pH 7.2) containing 200 mM NaCl in a calorimeter cell was titrated with a 125 M solution of the mutant Fv fragments in the same buffer at five different temperatures (20, 25, 30, 35 and 40°C). The ligand solution was injected 16 times in portions of 7 l during 15 s. Thermogram data were analyzed using a computer program (Origin) supplied by MicroCal, Inc. (30).

RESULTS
Expression, Purification, Overall Structure, and Antigenbinding Activities of Mutated HyHEL10 Fv Fragments-All of the mutants were obtainable in the E. coli secretory expression system and were purified by affinity chromatography using HEL-Sepharose. The yield was about 10 mg/liter of culture.
No difference in far-and near-UV CD spectra between wildtype and mutant Fv fragments was observed (data not shown), indicating that the overall structure of mutant Fv fragments is identical with that of the wild type.
As described previously (28), the HyHEL10 Fv fragment has complete inhibitory activity toward its antigen, HEL, in a Fv: HEL molar ratio of 1:1. Although the association constant of the interaction is high, a sigmoidal curve was observed (Fig. 1). This might be due to the dilution of Fv-HEL solution and the competition with the substrate in the measurements of HEL enzymatic activity. The inhibitory activities of the HAsp32-mutated Fv frag-ments were measured (Fig. 1). The inhibition profile indicated that the inhibitory activities for the D32N mutant and D32E mutant Fv fragments were slightly decreased and that for the D32A mutant Fv was significantly decreased, indicating that all mutations lowered the association.
Thermodynamic Analyses of the Interactions between Asp32 mutant Fv and HEL-To estimate the thermodynamic parameters of the interactions between the Asp32 mutant HyHEL10 Fv fragments and HEL, we performed titration calorimetry of the association between the HAsp32 mutant Fv fragments and HEL (Fig. 2). The enthalpy change (⌬H) and binding constant (K) on antigen-antibody interaction were directly obtainable from the experimental titration curve shown in Fig. 2 (30), and the heat capacity change (⌬Cp) of the interaction was estimated from the temperature dependence of the enthalpy change. The Gibbs energy change (⌬G ϭ ϪRT ln K) of binding was calculated from the binding constant, and the entropy change (T⌬S ϭ Ϫ⌬G ϩ ⌬H) on the association could also be estimated.
Thermodynamic parameters at 30°C calculated from the titration curves are represented in Table I, and the temperature dependence of the enthalpy and entropy change is shown in Figs. 3 and 4, respectively.
On comparison of all of the mutant Fv fragments with the wild-type Fv fragment, the negative values of the enthalpy change of the interaction between each mutant Fv fragment and HEL were increased by 17-24 kJ mol Ϫ1 at 30°C. The binding constants were decreased only slightly in all mutants, resulting in a small change in Gibbs energy. The increases in the negative enthalpy change (Ϫ⌬⌬H) for the mutant Fv fragments were completely compensated by the increase in negative entropy change (ϪT⌬⌬S), which was superior to Ϫ⌬⌬H, suggesting that Asp32 contributed to the interaction by sup- pressing the decrease in entropy change (i.e. the entropy loss due to binding). From the temperature dependence of the enthalpy change shown in Fig. 3, the heat capacity change was estimated to be ϳϪ2.2 kJ mol Ϫ1 K Ϫ1 for all the mutant Fvs, which was ϳ0.8 kJ mol Ϫ1 K Ϫ1 larger in negative value than that of the wild type (Table I). DISCUSSION Thermodynamic Features for Deletion of the Salt Bridge in the HyHEL10-HEL Interaction-We discuss the result reported here on the supposition that the effect of any change of protonation upon complex formation by mutation is completely compensated by the deprotonation of buffer. Although the structure of free Fv fragment has not been determined, far-and near-UV CD spectra of free mutant Fv fragments are identical with that of wild-type Fv (data not shown), indicating that no difference in overall structure between wild-type Fv and mutants exists. This suggests that the difference in thermodynamic parameters of the interaction between mutant Fv and HEL from those of wild-type Fv does not originate from the reconstruction of the partly unfolded structure of the mutant protein in the free state.
The data reported here indicated that deletion of the salt bridge in the HyHEL10-HEL interaction by mutation led to increased changes in both negative enthalpy and negative entropy. The negative Gibbs energy was slightly decreased in the same direction as the previously reported mutants (27). These results suggest that although the interaction between HAsp32 mutant Fv and HEL is enthalpically more favorable than that for the wild-type Fv fragment, the entropic loss due to binding is also increased, indicating that deletion of the salt bridge is enthalpically favorable and entropically unfavorable. This  means that the thermodynamic role of salt bridge formation is to decrease the entropic loss due to binding. Increased changes in both negative enthalpy and negative entropy by mutagenesis of the antigen-binding site of an antibody have also been observed in the mutant antibody-antigen Salmonella serogroup B O-polysaccharide interaction (20). Substitution of HHis101, which is exposed to solvent and forms a hydrogen bond with O-4 of mannose, with Gly or Asp produces a dramatic increase of negative enthalpy change in the mutant antibody-antigen interaction. However, the unfavorable entropy change outweighs the enthalpic advantage, resulting in a 10-fold lower binding constant. In the case of the interaction between Y82F mutant FK506-binding protein (FKBP-12) and its ligand (FK506), large increases in negative enthalpy (Ϫ⌬⌬H) and negative entropy (ϪT⌬⌬S) change have been observed in comparison with the wild-type FKBP-12 ligand interaction (32). The binding constant of the interaction between Y82F mutant FKBP-12 and its ligand is slightly decreased in comparison with the wild type. A crystallographic study of the complex formed between the wild-type FKBP-12 and its ligands has indicated that two water molecules are located in the hydroxyl group of Tyr82 in unliganded FKBP-12 and that on ligand binding these water molecules are lost and the hydroxyl group forms a hydrogen bond with the ligand. From these experimental data, it has been suggested that the formation of a hydrogen bond with the ligands results in a large decrease of negative entropy change due to removal of the water molecules (32).
A structural study of a free HEL has suggested that Lys97 is exposed and hydrated (33). Although the structure of a free HyHEL10 Fv fragment has not yet been determined, HAsp32 is thought from calculation to be exposed and hydrated (34,35). Because no water molecule is observed in the HyHEL10-HEL complex (10), the solvent molecules must be removed on binding of HEL. The increase in entropy change might originate mainly from the removal of solvent molecules on formation of the HyHEL10-HEL complex, because desolvation is accompanied by a considerable entropy advantage (36).
The present data indicate that a significant increase in negative enthalpy change is observed in the interactions between the three mutants and HEL. Substitution of the Asp at site 32 of VH with Glu resulted in a large increase in negative enthalpy and entropy, although both the enthalpy and entropy changes produced by dehydration of Asp or Asn are almost the same (37). Furthermore, although the entropic change of dehydration of the side chain of Ala is quite different from that of Asn (37), the difference in the negative entropy changes of the interactions between HD32A Fv and HD32N Fv was only subtle (1 kJ mol Ϫ1 K Ϫ1 at 30°C; Table I). These results suggest that the changes in thermodynamic parameters produced by mutation cannot be understood merely by the changes in dehydration occurring upon binding. Why, then, is negative entropy change increased by mutation?
Entropy Changes Resulting from Conformational Change on Binding of HyHEL10 Fv with HEL-According to the considerations of Murphy et al. (38), the total ⌬S of binding is given as where ⌬S solv is the change in entropy derived from solvent release upon binding, ⌬S conf is the change in entropy resulting from conformational changes due to formation of the antigenantibody complex, and ⌬S crat is the cratic entropy change. ⌬S solv is given by (T s ϭ 386 K; T s is the temperature at which the desolvation entropy change is considered to be zero), and ⌬S crat can be considered a constant value (Ϫ0.033 kJ mol Ϫ1 K Ϫ1 (39,40). From their experimental data, they have suggested that the overall conformational entropy loss is, to a large extent, compensated by the desolvation entropy. On the basis of this viewpoint, the present data can be analyzed as follows. In the interaction between HEL and the wild-type HyHEL10 Fv fragment, ⌬Cp was estimated to be Ϫ1.42 kJ mol Ϫ1 K Ϫ1 (Table I)  where ⌬S HE is the entropy change from hydrophobic interaction, and ⌬S RT is that from rotational and translational change. Where ⌬S ϭ 0 at a particular temperature (T s ), is obtained. ⌬S HE (T s ) is related to the heat capacity change, and ⌬S RT is considered to be constant (50 cal mol Ϫ1 K Ϫ1 ) (41). If the entropy change due to conformational changes is considered to be equal for each residue, the entropy change is calculated to be 5.6 cal mol Ϫ1 K Ϫ1 , and thus division of ⌬S others by Ϫ5.6 yields the number of residues involved in the conformational changes: R ϭ ⌬S others /Ϫ5.6 (Eq. 5) We have found that in the HyHEL10-HEL interaction, the folding in the antibody HyHEL10 is induced by binding with HEL (26). In this study, the numbers of residues (R) involved in the folding induced by the association between the mutant Fv fragments and HEL were estimated according to the above proposal (Table III). The results indicated that R values were increased by substitution with Ala, Asn, or Glu at position 32 of VH, suggesting that local conformational changes for the interaction between mutant Fv and HEL are increased by the mutation if the mutant structures of VH are not affected by the substitutions. Thus, it can be concluded that a large proportion of the increased negative entropy change induced by mutation arises from increased structural change, suggesting that the salt bridge suppresses excess local conformational change upon binding.
Conclusion-We previously analyzed thermodynamically the HyHEL10-HEL interaction using mutant Fv fragments (Tyr residues were substituted with Phe) (27). The hydroxyl groups of the Tyr residues at sites 33, 50, and 58 form hydrogen bonds with the antigen, HEL. The changes in the thermodynamic parameters induced by removal of a hydroxyl group from each of the Tyr residues at positions 33 and 50 in HyHEL10 VH are as follows; 1) the negative values of both the enthalpy and entropy change are decreased by substitution; 2) the R number is decreased. These data indicate that hydrogen bonds created by complementary association of the antibody with the antigen are one of the major contributors to the gain in binding enthalpy. However, the data reported here indicate that the salt bridge makes a different thermodynamic contribution to the interaction.
In conclusion, the removal of the salt bridge produced an increase in the negative entropy (ϪT⌬⌬S) of the interaction, and ϪT⌬⌬S was almost compensated by an increase in negative enthalpy (Ϫ⌬⌬H), resulting in a small decrease in negative binding Gibbs energy (Ϫ⌬G). The entropic contribution of salt bridge removal arose not only from the desolvation effect but also from the structural effect (i.e. excess conformational change). The salt bridge may be formed to decrease this entropic loss and to suppress the extra complementary association in the HyHEL10-HEL interaction. a Temperature where entropic contribution to the interaction is zero. Values were estimated from the plot in Fig. 4.
b Calculated from the equation ⌬S HE ͑T s ͒ ϭ 1.35 ⌬Cp ln ͑T s /386͒ c In the previous papers, we calculated the R value using Ϫ5.6 e.u. (Equation 5 in the text) and Ϫ50 e.u. (Equation 4 in the text) as shown by Spolar and Record (41). However, it seems a unit "e.u." is confused in the paper of Spolar and Record (41). "e.u." should be replaced by "cal mol Ϫ1 K Ϫ1 " (1 e.u. ϭ 1.9872 cal mol Ϫ1 K Ϫ1 ). Thus, here we calculated the R value using Ϫ5.6 cal mol Ϫ1 K Ϫ1 in Equation 5 and Ϫ50 cal mol Ϫ1 K Ϫ1 in Equation 4.