Thermodynamic Analyses Reveal Role of Water Release in Epitope Recognition by a Monoclonal Antibody against the Human Guanylyl Cyclase C Receptor*

The thermodynamics of a monoclonal antibody (mAb)-peptide interaction have been characterized by isothermal titration microcalorimetry. GCC:B10 mAb, generated against human guanylyl cyclase C, a membrane-associated receptor and a potential marker for metastatic colon cancer, recognizes the cognate peptide epitope HIPPENIFPLE and its two contiguous mimotopes, HIPPEN and ENIFPLE, specifically and reversibly. The exothermic binding reactions between 6.4 and 42 °C are driven by dominant favorable enthalpic contributions between 20 and 42 °C, with a large negative heat capacity (ΔC p ) of −421 ± 27 cal mol−1 K−1. The unfavorable negative value of entropy (ΔS b 0) at 25 °C, an unusual feature among protein-protein interactions, becomes a positive one below an inversion temperature of 20.5 °C. Enthalpy-entropy compensation due to solvent reorganization accounts for an essentially unchanged free energy of interaction (ΔΔG b 0 ≅ 0). The role of water molecules in the recognition process was tested by coupling an osmotic stress technique with isothermal titration microcalorimetry. The results provide direct and compelling evidence that GCC :B10 mAb recognizes the peptides HIPPENIFPLE, HIPPEN, and ENIFPLE differentially, with a concomitant release of variable and nonadditive numbers of water molecules (15, 7, and 3, respectively) from the vicinity of the binding site.

Immune system recognition of ligands involves the formation of a multitude of specific intermolecular interactions between the components of the host and the foreign body (1)(2)(3)(4)(5)(6)(7). Atomic level investigations on antigen-antibody complex formation reveal that, although charged and polar side chains can be critical in binding, notably important contributions originate from hydrophobic side chains just as they do in other protein-protein complexes (1, 6, 8 -11). Solution state studies of the binding of two different proteins to the same antibody, for example, have shown that, in certain cases, the free energy of binding may arise from the accumulation of many interactions of varying strength over the entire protein-protein interface (12). Thus, the perception as propounded by crystal structural results that protein-protein recognition is mediated by only a few strong interactions may not be a general feature (13). Complications also arise from the observation that a majority of direct contacts in the antigen-antibody complex may be energetically neutral (14). This highlights the need for a complete characterization of binding energetics of antigen-antibody interactions as a prelude to understanding the underlying structural basis of the molecular recognition process.
The net balance between the energy versus distance functions of the aspecific (macroscopic) repulsion that usually prevails between antigen and antibody molecules in aqueous media vis-à -vis the specific (microscopic) attraction between the epitope and paratope of the antigen and the antibody, respectively, determines their propensity for complex formation (15). Rearrangement of solvent water molecules accompanying macromolecular complexation events is believed to be largely responsible for the observed nonlinear thermodynamic phenomena (16). Water molecules serve a variety of roles and, in different positions, significantly affect the energetics of biomolecular interactions (17). The pioneering work of Parsegian et al. (18) that involved controlled alteration of water activity using the osmotic stress technique has laid the foundation for a direct address of different roles of water molecules in solution. We report here an application of a coupled osmotic-isothermal titration calorimetry (ITC) 1 approach to delineate the role of water molecules in the recognition of a cognate peptide epitope, the determinant that confers the specificity of interaction, by a mAb (16,19).
Membrane-associated receptor guanylyl cyclase C (GCC) is a potential marker for metastatic colon cancer (20 -23). GCC:B10 mAb, generated against the human GCC receptor, binds specifically to the unique, conserved epitope HIPPENIFPLE (23). The epitope contains two independent recognition motifs (mimotopes) for GCC:B10 mAb, one represented by HIPPEN and the other by ENIFPLE, an example of novel topological mimicry and a probable duplication of the cognate epitope in the native GCC receptor sequence (24). The thermodynamics of the GCC:B10 mAb-peptide binding reactions were considered in terms of the site binding constant (K b ) and changes in the free energy (⌬G b 0 ), the binding enthalpy (⌬H b 0 ), and the binding entropy (⌬S b 0 ). The binding reaction between the GCC:B10 mAb-binding site and the peptide ligand (L) is given by GCC: B10 mAb ϩ L º GCC:B10 mAb⅐L, in 10 mM phosphate buffer * This work was supported by grants from the Department of Science and Technology, Government of India (to A. S.), Indian Institute of Science (to S. S. V. and A. S.), and the Department of Biotechnology, Government of India (to S. S. V.). The matrix-assisted laser desorption ionization mass spectrometer is funded by the Department of Biotechnology, Government of India. 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. ** Research Associate of the Council of Scientific and Industrial Research of India.

EXPERIMENTAL PROCEDURES
Materials-All reagents were of analytical or ultrapure grade from Sigma. Deionized Milli-Q water was used throughout. The peptide HIP-PENIFPLE was a product of Peptidogenic Research and Co. (Livermore, CA), whereas the peptides HIPPEN and ENIFPLE were synthesized by standard Fmoc (N-(9-fluorenyl)methoxycarbonyl) chemistry. All peptides were purified by reverse-phase high pressure liquid chromatography on a Shim-PAK Prep-ODS(H) preparative column and quantitated by amino acid analysis as described earlier (24). The purity and the mass of the peptides were ascertained by matrix-assisted laser desorption ionization mass spectrometry. The monoclonal antibody GCC:B10 was generated as described earlier (23). The concentration of GCC:B10 mAb was determined spectrophotometrically in a quartz cuvette of 1-cm path length at 280 nm using a molar extinction coefficient of ⑀ ϭ 210,000 at pH 7.2 and expressed in terms of a dimer of M r ϭ 150,000 (25).
Preparation and Analysis of Solutions-GCC:B10 mAb solution was prepared in PBS, dialyzed overnight in a large volume of the same buffer, and centrifuged to remove any insoluble matter. The peptide solutions were prepared by weight in the dialysate to minimize differences between the protein buffer solution and the ligand buffer solution in the ITC measurements. For osmotic stress studies, the mAb solution was dialyzed extensively against glycerol and ethylene glycol solutions in the above buffer. Peptide solutions were then prepared with the final dialysate. Care was taken that the neutral solute osmolalities used were not significantly different from their ordinary molal concentrations (16,26).
ITC Measurements and Analyses-The titration calorimetric measurements were performed with a Microcal TM Omega titration calorimeter (19). Samples carefully scrutinized for precipitate after the titration revealed none either in the presence or absence of the osmolytes. The quantity c ϭ K b ⅐[GCC:B10 mAb] 0 , where [GCC:B10 mAb] 0 is the initial macromolecular GCC:B10 mAb concentration and K b is the association constant, was in the range of 2 Ͻ c Ͻ 200 as required for ITC studies (19). This corresponds to a binding regime that is best suited for the most precise measurements of the binding stoichiometry, K b , and ⌬H b 0 simultaneously in a single ITC experiment. The total concentration of GCC:B10 mAb used was from 0.02 to 0.04 mM, whereas the total concentration of peptide ligand taken was from 0.8 to 1.2 mM. The titration of ligand solution in this concentration range with the buffer solution alone gave negligible values for the heat of dilution both in the presence and absence of osmolytes at 293.2 K. Nonetheless, for every experiment, the heat of dilution of the ligand was measured and subtracted from the runs conducted with the mAb. The time duration between the injections was at least 3 min to allow the peak to return to the base line, and the number of additions of the peptide titrant was fixed such that the area below the peak was reduced by at least an order of magnitude or until antibody-binding sites were saturated Ͼ95%. All measurements were made at a constant stirrer speed of 395 rpm. Since GCC:B10 mAb contains two binding sites per molecule, an identical-site model utilizing a concentration of the protein dimer was the simplest binding model found to provide the best fit to the ITC data using the Origin TM program (19,27). Values for ⌬S b 0 were obtained from the basic equation of thermodynamics, where n ϭ number of moles, T is the absolute temperature, and R ϭ 8.315 J mol Ϫ1 K Ϫ1 .

RESULTS
The results of a typical titration calorimetry measurement, which consisted of adding 15-l aliquots of 0.8 mM HIPPENIPLE solution to 0.03 mM GCC:B10 mAb solution in PBS at 283.2 K, together with the nonlinear least-squares fit of the data are shown in Fig. 1. The results exhibit a monotonic decrease in the exothermic heat of binding with successive injections until saturation is achieved. As shown by the solid curve in   (Table I). A plot of the change in enthalpy of binding of HIP-PENIFPLE to GCC:B10 mAb as a function of temperature yields a ⌬C p of Ϫ421 Ϯ 27 cal mol Ϫ1 K Ϫ1 (Fig. 3). The temperature dependence of entropy is shown in Fig. 4. The unfavorable negative value of ⌬S b 0 at 25°C becomes a positive one below an inversion temperature of 20.5°C, an unusual feature among protein-protein systems (28). There is almost no temperature dependence of the binding free energy (i.e. ⌬⌬G b 0 Х 0) within the temperature range examined (Fig. 5). This occurs because the contributions from the enthalpic and entropic components are reciprocal (Figs. 3 and 4); hence, they balance each other out in such a fashion that the changes are relatively equal in magnitude, but opposing one another.
The dependence of the thermodynamic parameters for the binding of the three peptides, HIPPENIFPLE, HIPPEN, and ENIFPLE, to GCC:B10 mAb on the osmolal concentration of a representative neutral solute used, viz. glycerol, is shown in Fig. 6. For each solute, binding free energies show a linear dependence on solute osmolal concentration, consistent with an exclusion of these solutes from the water surrounding the associating peptide and antibody surfaces. The positive sign of the slope defines the net release of water molecules by the GCC:B10 mAb-peptide complex during the process of binding. The magnitude of the effect of osmotic stress depends both on the solute osmolal concentration, which is equivalent to the bulk water chemical potential, and on the difference in the number of solute-excluding water molecules associated with the GCC:B10 mAb-peptide complex and the number associated with the free GCC:B10 mAb and peptide molecules. The slope of log K a versus ln a w , where a w is the water activity, is 2.303 ⌬n w . This value, ⌬n w , gives the change in the number of soluteexcluding water molecules coupled to the binding process (16,18). Since ln a w ϭ Ϫ[solute] osmolal /55.56, where [solute] osmolal is the solute osmolal concentration and 55.56 is the moles of water in 1 kg, the slope of the lines shown in Fig. 6  Our results demonstrate that GCC:B10 mAb recognizes the peptides HIPPENIFPLE, HIPPEN, and ENIFPLE differentially, with a concomitant release of variable numbers of water molecules, viz. 15, 7, and 3, respectively, from the vicinity of the binding site (Tables II and III). Control experiments in which aliquots of glycerol or ethylene glycol were injected into GCC: B10 mAb or peptide solutions did not result in any measurable heats of binding, ruling out the possibility of specific interactions between the stressing solute and GCC:B10 mAb or peptides, affirming the neutrality of the chosen osmolytes. DISCUSSION Formation of specific protein-protein complex is mediated by a few productive interactions that dominate the energetics of association, whereby the residues at the periphery make only a minor contribution to the binding energy (12). Among the relatively few protein-protein interaction systems for which thermodynamic data are available, there appears to be a group of complexes characterized by a large favorable enthalpy of binding that is partly offset by an unfavorable entropy, and most antigen-antibody interactions studied so far seem to belong to this group (28 -35). There is a contrasting group of protein complexes that are stabilized by a large positive entropy of binding, partly compensated by an unfavorable enthalpic term as found among protease-protease inhibitor complexes (36 -39). Within the limited data set available, there are only a few cases wherein both the enthalpic and entropic contributions are favorable (40 -43). For the interaction of the HIPPENIFPLE   peptide with GCC:B10 mAb at physiological temperature (37°C) as well as at room temperature (25°C), the binding reaction is driven strongly by the enthalpic component, i.e. the binding enthalpy is negative and contributes favorably to the binding free energy (Table I). At these temperatures, the binding entropy is negative and hence contributes unfavorably to the binding energy. Interestingly, there is a strong temperature dependence of the binding entropy; and therefore, below the inversion temperature of 20.5°C, it changes sign and becomes a positive one (Fig. 4). Hence, below 20.5°C and up to 2.3°C, the inversion temperature of the enthalpy of binding of the HIPPENIFPLE peptide to GCC:B10 mAb, there are favorable contributions to the free energy from both enthalpic and entropic components. Below 2.3°C, the binding enthalpy changes sign (Fig. 3) and becomes positive, and the reaction is driven only by favorable entropic contributions. The lowest temperature at which we could study the binding of peptides to GCC:B10 mAb was 6.4°C, below which the heats of reaction are too small to be measured accurately (19). Despite the limited number of antibody residues that can be used for complexation, several modes of molecular recognition abound in protein-protein interactions (6, 12, 14, 44, 45). Mari-uzza and co-workers (12) have defined two categories of functional epitopes, viz. ones in which ligand binding is mediated by a small subset of contact residues and the others in which the free energy of binding arises from many productive interactions distributed over the entire mAb-peptide interface. The structures of antigen-antibody complexes show that, although the interfaces in protein-protein interaction are large, only a small number of contact side chains are important for binding (1,6,8,9). Bridging water molecules are seen between polar or charged residues in regions of contact interface that are less important (13,46). The correlations are much better when one considers only the burial of well packed hydrophobic side chains (11). Both polar and nonpolar residues have been shown to play a prominent role in the interaction of human growth hormone with its cognate monoclonal antibody (47), whereas growth hormone-receptor interactions suffer substantial reductions in affinity when hydrophobic amino acids are replaced (13).
The formation of the specific complex between the HIP-PENIFPLE peptide and GCC:B10 mAb requires the presence of at least a single proline residue (24). This feature is shared by the two half-peptides HIPPEN and ENIFPLE, which retain reactivity with GCC:B10 mAb. Since randomized peptide sequences with amino acid composition identical to the native peptide sequence do not show any reactivity with GCC:B10 mAb, HIPPEN and ENIFPLE represent two independent mimotopic recognition motifs present in the same epitope, viz. HIPPENIFPLE (24). The presence of a large number of nonpolar amino acids in the epitope sequence such as Ile, Phe, and Leu suggests that water molecules could be released from the interacting interface when the HIPPENIFPLE peptide binds to GCC:B10 mAb, consistent with the establishment of a hydrophobic contact, which is substantiated by the observation of a large negative value of ⌬C p ϭ Ϫ421 Ϯ 27 cal mol Ϫ1 K Ϫ1 . The enthalpically driven reaction for the binding of the peptides to GCC:B10 mAb suggests that hydrogen-bonding interactions contribute favorably to the net binding free energy, whereas the large negative ⌬C p suggests that hydrophobic forces are also involved in driving the binding process. These two results, seemingly at variance with each other, are consistent with similar observations in other antigen-antibody interaction systems (28 -35). For example, apart from the existence of hydrogen bonding in the complex, binding of anti-p24 (human im- munodeficiency virus) mAb to the antigen-derived epitope peptide GATPQDLNTnL (n ϭ norleucine) and the epitopehomologous peptide GATPEDLNQKLAGN occurs through hydrophobic and van der Waals contacts observed predominantly with the aromatic side chains of complementarity determining regions of both variable light and variable dark chains (6). An analysis of 21 protein complexes, including antigen-antibody, protease-inhibitor, and other heterologous associations, has shown that both hydrophobic and hydrogen-bonding interactions are important for stability to protein association (48).
Thus, it appears that nature would have evolved antibodies to exploit a unique theme for recognizing peptide and protein antigens, namely a combined usage of hydrogen-bonding and hydrophobic interactions as energetic signatures, as opposed to predominantly hydrophobic association for other protein-protein interactions.
Enthalpy-entropy compensation has been consistently observed in antigen recognition by antibodies (28 -35). Enthalpyentropy compensation is associated with solvent reorganization accompanying protein-ligand interactions (16,49,50) and also appears to be a general feature of weak intermolecular reactions (51). It has recently been demonstrated that water reorganization plays a direct role in enthalpy-entropy compensation at least in a lectin-sugar system (16). The thermodynamics of binding of peptides to GCC:B10 mAb show compensatory changes in ⌬H b 0 and T⌬S b 0 (Fig. 7). When a group of analogous species interact by the same mechanism, then a linear relationship between enthalpy and entropy can be expected, with the slope exactly equal to unity being a result of complete compensation (16,49). Our results also show a linear relationship between ⌬H b 0 and T⌬S b 0 , with a slope of 0.91 (correlation coefficient ϭ 0.98), a good indication for a single interaction mechanism for the specific recognition of the epitopic and mimotopic peptides by GCC:B10 mAb, namely solvent reorganization (Fig.  7). This is substantiated by the evidence that the slope of the enthalpy-entropy compensation plot is less than unity, suggesting that the free energy of binding is more sensitive to changes in entropy as seen in other antigen-antibody interaction systems (28 -35). This validates the proposal that enthalpy-entropy compensation be considered as a diagnostic of a true osmotic effect (16).
A release of interfering water molecules that are present in the binding interface would facilitate binding of antigens to an antibody-combining site. On the other hand, bound water molecules are capable of acting as a molecular glue by altering the surface complementarity and, at appropriate sites, involve themselves in hydrogen bonding between the ligand and the antibody (52). In addition to this, water molecules that are present several hydration shells away are capable of dynamically influencing the binding of ligands to macromolecules (53). An elucidation of the precise nature of the reorganization of solvent, which could involve either the uptake or the release of water molecules or the restructuring of the existing ones within and around the binding pocket, would be an important advance in our understanding of antigen recognition by antibodies. Hence, to resolve as to which of these potential role(s) of water molecules predominate during the binding reaction and thereby fashion out the parameters contributing to the association of antigens by antibodies, it is imperative that one addresses the issue in the solution state itself. Glycerol and ethylene glycol are solutes of different chemical nature, yet their effect on GCC:B10 mAb-peptide interactions reported here is similar (Table III). Circular dichroism spectroscopy measurements of the peptides HIPPENIFPLE, HIPPEN, and ENIFPLE in the far-UV region (250 to 200 nm) reveal that their structure remains unperturbed in PBS in the presence of up to 5 M glycerol or ethylene glycol, which is more than twice the highest concentration used in ITC studies. Hence, it is unlikely that an osmolyte like glycerol would have altered the structure of these peptides in the conditions used in our experiments, thus ruling out any significant contributions emanating therefrom.   These osmolytes impinge upon the water activity of the solution because of their propensity to form hydrogen bonds with water molecules. It is well known that glycerol and ethylene glycol are preferentially excluded from the tiny spaces and from the envelope around the macromolecular surface and do not interact with proteins or alter their conformation significantly (16,18,(53)(54)(55)(56)(57).
It is noteworthy that the role and extent of conformational changes occurring during the molecular recognition event of antigen binding to antibody would contribute to the net thermodynamics of the reaction (1,6,8). Calorimetric studies have shown that the enthalpies determined for the binding of antigen and whole antibodies or their Fab or Fv fragments are essentially the same, implying that major conformational changes do not occur at least in the constant domains of the antibody that affect the thermodynamics of the antigen binding reaction (30). The extent of the conformational adaptation appears to depend on the size-dependent binding surface (58). The interface adaptor hypothesis put forward by Colman (58) proposed that when the size of the antigen interface approaches the V L -V H interdomain interface, then V L -V H rearrangements might accompany and play a role in antigen-antibody complementarity. It has been demonstrated crystallographically that the binding of many peptides to the same binding site of anti-p24 (human immunodeficiency virus type 1) murine monoclonal antibody occurs without the need of elaborate induced-fit mechanisms (6). Binding can occur simply by using different contact patterns. Besides minor changes in the binding site, ligand association is often connected with a displacement of the two variable regions V L and V H , whereas rotation of the constant and variable Fab domains with respect to each other does not seem to be correlated with binding (1). For an explanation of our results from osmotic experiments, it is difficult to evoke any other common property of the two starkly different neutral osmolytes than an osmotic mechanism of solute action. Such reasoning is substantiated by the linearity of the plots of the logarithm of binding constants versus neutral solutes' osmolality for the binding of peptides to GCC:B10 mAb. This clearly indicates a linkage to a well defined and constant difference in the number of water molecules involved in ligand binding (16,18). Since the slopes are nearly the same for two chemically different classes of neutral osmolytes, it appears reasonable that the observed changes be rationalized in terms of a true osmotic effect, thus distinguishing it from other potential effects (16,18).
The net free energy change for the interaction of two hydrated species coming into contact, GCC:B10 mAb and the corresponding peptide molecule, is the difference between the free energy of the complex of GCC:B10 mAb and the peptide vis-à -vis those between GCC:B10 mAb and water as well as peptide and water. The measurement of these changes in numbers of water molecules for binding of peptides to GCC:B10 mAb is directly connected to the energetics of these hydrationdehydration reactions. The interaction of the tested peptides with GCC:B10 mAb is enthalpically driven and accompanied by the release of differential numbers of water molecules (Tables   II and III). The osmotic release of water molecules from the binding site causes a concomitant negative increase (favorable) in the free energy of binding (Fig. 6). It therefore follows that peptide recognition by GCC:B10 mAb involves the removal of bound water molecules from the vicinity of the binding sites, a result that contrasts with the observation of bound water molecules mediating the association of mAb D1.3 with hen egg white lysozyme (52). The results presented herein represent a direct and unequivocal demonstration of the role of the release of water molecules in the recognition of peptide ligand by a monoclonal antibody in the solution state. Our results present another facet of antigen recognition by mAbs and emphasize that water molecules could play more diverse roles in mediating antigen recognition by antibodies than previously believed.
From our data, even in the absence of crystal structural information, it is clear that the molecular mimicry by the two mimotopes is not energetically equivalent (Tables I and II). This highlights the existence of either a set of different contact patterns at the same binding site or, equally likely, different energetic contributions from the same set of contact residues. Starkly contrasting levels of correlation between buried surface areas and free energy of association exist in literature (44, 46, 52, 59 -61). Occasionally, it is also found that amino acid residues in the peptide that are not involved in directly contacting the antibody-binding pocket do not tolerate substitution. It has been observed that a key residue (Arg 9 ) in the binding site of the peptide GLYEWGGARITNTD stabilizes the overall peptide conformation by internal hydrogen bonding and shows no contact with the anti-p24 (human immunodeficiency virus type 1) murine mAb-binding pocket (6). A detailed evaluation of these possibilities must await results from further structural studies in progress.
The large difference in the heat capacity for the binding of HIPPEN (⌬C p ϭ Ϫ541 Ϯ 25 cal mol Ϫ1 K Ϫ1 ) and ENIFPLE (⌬C p ϭ Ϫ37 Ϯ 12 cal mol Ϫ1 K Ϫ1 ) half-sites, which are topological mimics of each other, suggests corresponding differences in energetics and mode of binding. This is substantiated by the strikingly different dependence of the thermodynamic parameters ⌬H b 0 and ⌬S b 0 for the binding of HIPPEN and ENIFPLE to GCC:B10 mAb (Figs. 3 and 4). In addition, it is supported by the variation in the number of water molecules released during the binding of these two peptides to GCC:B10 mAb (Table III). The two half-peptides competing to bind to the same site of the antibody would elevate the free fractions of both half-peptides. An examination of the attribution of the free energies of binding of peptides to GCC:B10 mAb reveals that there is a nonequivalence of the binding free energies of the mimotopes as compared with that of the epitope. Following Jencks' (62) suggestion for the attribution of free energies, when the binding of the HIPPEN, ENIFPLE, or HIPPENIFPLE peptide has the same loss of entropy corresponding to ⌬G s , the free energy for the binding of the HIPPENIFPLE epitope to GCC:B10 mAb will be more favorable than the sum of the free energy contributions from the HIPPEN and ENIFPLE peptide mimotopes by an amount corresponding to this entropy loss. The loss of entropy on combining HIPPEN and ENIFPLE by a covalent bond to form HIPPENIFPLE can be as much as Ϫ40 or Ϫ32 cal mol Ϫ1 K Ϫ1 for molar and mole fraction standard states, respectively, and some unpredictable fraction of this difference will appear in the binding of HIPPENIFPLE compared with that of HIPPEN and ENIFPLE. The entropy barrier for binding (⌬G s ) is given by the difference between the intrinsic binding energies, ⌬G HIPPEN i ϩ ⌬G ENIFPLE i , and the observed binding energy, ⌬G HIPPENIFPLE 0 . An empirical way of dealing with this problem is provided by use of the ⌬G s term, according to Equations 2 and 3, as follows. The ⌬G s barrier must be overcome for the binding of the HIPPEN or ENIFPLE peptide as well as for the HIP-PENIFPLE peptide to GCC:B10 mAb. The intrinsic binding energies are additive and, once the entropy barrier has been overcome, can be expressed as increases in the observed binding energies as follows (Equation 4).  (Table IV) suggests that the two mimotopes are recognized by GCC:B10 mAb in an energetically differential fashion. Hence, the intrinsic binding energy of the HIPPEN mimotope contributes two-thirds of the intrinsic binding energy of the HIPPENIFPLE epitope, whereas the ENIFPLE mimotope contributes the remaining one-third to it. These results and those from osmotic experiments together exclude the possibility that the binding of these two topological mimics occurs solely through a competitive mode of recognition of the mimotopes and suggest that GCC: B10 mAb utilizes an accommodative mode of recognition of the two half-peptides. Taken together, these results demonstrate that the ability of GCC:B10 mAb to differentiate between the HIPPEN and ENIFPLE peptides arises from the fine differences in the underlying thermodynamics of their interaction with GCC:B10 mAb. This suggests that the native GCC receptor would have evolved sophisticated mechanisms to diversify the recognition mode and hence the selective emergence of differences in the recognition of the two half-sites.