Potential aggregation hot spots in recombinant human keratinocyte growth factor; a computational study

Abstract

Background: Recombinant human keratinocyte growth factor is a highly aggregation-prone therapeutic protein. The high aggregation liability of rhKGF is manifested by loss of the monomeric form of the protein and accumulation of the aggregated species even at moderate temperatures. Here, we analyzed rhKGF for its vulnerability towards aggregation by detection of aggregation-prone regions (APRs) using several sequence-based computational tools including TANGO, SolubiS, ZipperDB, AGGRESCAN, Zyggregator, Camsol, PASTA, SALSA, WALTZ, SODA, Amylpred, AMYPDB, and structure-based tools including Aggrescan3D and molecular dynamics-based spatial aggregation propensity (SAP) algorithm. Results: The sequence-based prediction of APRs in rhKGF indicated that they are mainly located at positions 10-30, 40-60, 61-66, 88-120, and 130-140 which are rich in β-branched aliphatic, hydrophobic, aromatic and Glutamine/Aspargine (Q/N) residues. Mapping on the rhKGF tertiary structure revealed that most of these residues including F16-R25, I43, E45, R47-I56, F61, Y62, N66, L88-E91, E108-F110, A112, N114, T131, and H133-T140 are surface-exposed in the natively folded protein which can promote aggregation without major unfolding event or the conformational change may occur in the oligomers composed of natively folded monomers. The other regions are buried in the native state and their contribution to non-native aggregation is mediated by a preceding unfolding event in the monomeric state of the protein. The structure-based prediction of APRs using SAP tool limited the number of identified APRs to the dynamically-exposed hydrophobic residues including V12, A50, V51, L88, I89, L90, I118, L135, and I139 mediating the native-state aggregation. Conclusion: Our analysis of APRs in rhKGF identified the regions determining the intrinsic aggregation propensity in both folded (native) and unfolded state of the protein. These regions are the candidate positions for engineering the rhKGF sequence to reduce its aggregation tendency.