Size-exclusion chromatographic protein refolding: fundamentals, modeling and operation

Size-exclusion chromatography (SEC) has proven its capability to refold a variety of proteins using a range of gel filtration column materials, demonstrated in the growing body of experimental evidence. However, little effort has been allocated to the development of mechanistic models describing size-exclusion chromatographic refolding reactors (SECRR). Mechanistic models are important since they provide a link between process variables like denatured and reduced protein feed concentration (C(f,D&R)), flow rate, column length, etc., and performance indicators like refolding yield (Y(N)), thereby opening the possibility for in silico design of SECRRs. A critical step, in the formulation of such models, is the selection of an adequate reaction mechanism, which provides the direct link between the separation and the refolding yield. Therefore, in this work we present a methodology using a SEC refolding reactor model, supported by a library of reaction mechanisms, to estimate a suitable reaction scheme using experimental SEC refolding data. SEC refolding data is used since it provides information about the mass distribution of monomers and aggregates after refolding, information not readily available from batch dilution refolding data alone. Additionally, this work presents (1) a systematic analysis of the reaction mechanisms considered using characteristic time analysis and Damköhler maps, revealing (a) the direct effect of a given reaction mechanism on the shape of the SEC refolding chromatogram (number of peaks and resolution) and (b) the effect that the competition between convection, refolding and aggregation is likely to have on the SEC refolding yield; (2) a comparison between the SECR reactor and the batch dilution refolding reactor based on mechanistic modeling, quantitatively showing the advantages of the former over the latter; and (3) the successful application of the modeling based strategy to study the SEC refolding data of an industrially relevant protein. In principle, the presented modeling strategy can be applied to any protein refolded using any gel filtration material, providing the proper mass balances and activity measurements are available.