The distribution of hapten binding enthalpies in conventionally-raised antibody populations

Abstract

Conventionally-raised rabbit antibodies against the 2, 4-dinitrophenyl (DNP)‡ and 2, 4, 6-trinitrophenyl (TNP) groups exhibit a broad range of binding free energies Δ G° for their homologous haptens. The question of whether enthalpic or entropie factors are primarily responsible for this dispersion of binding constants first arose in connection with calorimetric studies of heats of hapten binding in these systems (Barisas et al., 1972). We consider here the examination of this question by three potentially suitable methods which avoid physical fractionation of the antibody samples into affinity classes: (1) direct calorimetry; (2) temperature dependence of the Sips heterogeneity index, as measured by equilibrium dialysis; (3) temperature-jump van't Hoff measurements. We examined DNP- and TNP-lysine binding by homologous rabbit anti-DNP and anti-TNP antibody populations, respectively. Method (1) shows the binding enthalpies ΔHof tightest-binding sites to be indistinguishable from those of the average sites. For methods (2) and (3), we define an affinity-binding enthalpy correlation parameter ƒ which is zero if all affinity groups have the same ΔHand which is one if variations in ΔG° among groups are attributable to those in ΔH. Values of ƒ greater than one indicate differences in ΔH among groups larger than the corresponding differences in ΔG°, these enthalpic differences being partially compensated by entropie factors which act to reduce the free energy differences between groups. Method (2) proved to be quite insensitive, yielding results consistent with ƒ = 0 for both the anti-DNP and anti-TNP antibody preparations. Method (3) deals with the temperature dependence dθ dT of the total extent of saturation θ of a heterogenous system. When a plot of dθ dT vs θ is skewed to the right and when its values are higher than those predicted when ƒ = θ is assumed, a positive value of ƒ is indicated. For the anti-DNP and anti-TNP systems examined, positive values of f between 0 and 1.5 were obtained, thus it can be concluded that the distribution of binding constants in these systems is at least in part caused by enthalpic heterogeneity, the inconclusive results yielded by methods (1) and (2) notwithstanding. Methods (2) and (3) may be useful in examining the distribution of binding enthalpies in heterogeneous populations of isolated receptor molecules.