The interaction between the cohesin and dockerin modules serves to attach cellulolytic enzymes (carrying dockerins) to non-catalytic scaffoldin units (carrying multiple cohesins) in cellulosome, a multienzyme plant cell-wall degrading complex. This interaction is species-specific, for example, the enzyme-borne dockerin from Clostridium thermocellum bacteria binds to scaffoldin cohesins from the same bacteria but not to cohesins from Clostridium cellulolyticum and vice versa. We studied the role of interface residues, contributing either to affinity or specificity, by mutating these residues on the cohesin counterpart from C. thermocellum. The high affinity of the cognate interactions makes it difficult to evaluate the effect of these mutations by common methods used for measuring protein-protein interactions, especially when subtle discrimination between the mutants is needed. We described in this article an approach based on indirect enzyme-linked immunosorbent assay (ELISA) that is able to detect differences in binding between the various cohesin mutants, whereas surface plasmon resonance and standard ELISA failed to distinguish between high-affinity interactions. To be able to calculate changes in energy of binding (ΔΔG) and dissociation constants (K(d) ) of mutants relative to wild type, a pre-equilibrium step was added to the standard indirect ELISA procedure. Thus, the cohesin-dockerin interaction under investigation occurs in solution rather than between soluble and immobilized proteins. Unbound dockerins are then detected through their interaction with immobilized cohesins. Because our method allows us to assess the effect of mutations on particularly tenacious protein-protein interactions much more accurately than do other prevalent methods used to measure binding affinity, we therefore suggest this approach as a method of choice for comparing relative binding in high-affinity interactions. Copyright © 2012 John Wiley & Sons, Ltd.