Insights into the stability of a therapeutic antibody Fab fragment by molecular dynamics and its stabilization by computational design

Abstract

23Successful development of protein therapeutics depends critically on achieving stability 24 under a range of conditions, while retaining their specific mode of action. Gaining a deeper 25 understanding of the drivers of instability across different stress conditions, will potentially enable 26 the engineering of protein scaffolds that are inherently manufacturable and stable. Here, we 27 compared the structural robustness of a humanized antibody fragment (Fab) A33 using atomistic 28 molecular dynamics simulations under two different stresses of low pH and high temperature.29 RMSD calculations, structural alignments and contact analysis revealed that low pH unfolding 30 was initiated through loss of contacts at the constant domain interface (C L -C H 1), prior to C L domain 31 unfolding. By contrast, thermal unfolding began with loss of contacts in both the C L -C H 1 and 32 variable domain interface (V L -V H ), followed by domain unfolding of C L and also of V H , thus 33 revealing divergent unfolding pathways. FoldX and Rosetta both agreed that mutations at the C L -34 C H 1 interface have the greatest potential for increasing the stability of Fab A33. Additionally, 35 packing density calculations found these residues to be under-packed relative to other inter-36 domain residues. Two salt bridges were identified that possibly drive the conformational change 37 at low pH, while at high temperature, salt bridges were lost and reformed quickly, and not always 38 with the same partner, thus contributing to an overall destabilization. Sequence entropy analysis 39 of existing Fab sequences revealed considerable scope for further engineering, where certain 40 natural mutations agreed with FoldX and Rosetta predictions. Lastly, the unfolding events at the 41 two stress conditions exposed different predicted aggregation-prone regions (APR), which would 42 potentially lead to different aggregation mechanisms. Overall, our results identified the early 43 stages of unfolding and stability-limiting regions of Fab A33, which provide interesting targets for 44 future protein engineering work aimed at stabilizing to both thermal and pH-stresses 45 simultaneously. 46 3 47 Author Summary 48 49Currently, antibody-based products are the most rapidly growing class of pharmaceuticals 50 due to their high specificity towards their targets, such as biomarkers on the surface of cancer 51 cells. However, they tend to aggregate at all stages of product development, which leads to 52 decreased efficiency and could elicit an immunological response. Improvements in the stability of 53 therapeutic antibodies are generally made during the development phase, by trial and error of the 54 composition of the formulated product, which is both costly and time consuming. There is great 55 demand and potential for identifying the drivers of instability across different stress conditions, 56 early in the discovery phase, which will enable the rational engineering of protein scaffolds. This 57 work elucidated the stability-limiting regions of the ant...

Structural mechanism of Fab domain dissociation as a measure of interface stability

“…Furthermore, we found exemplary mechanisms of dissociation in the Fv and in the C H 1–C L region. In our simulations, we do not sample the unfolding of the single domains, which happens only after the loss of contacts at the interface between the V H –V L and C H 1–C L domains [ 93 ]. The Fab structures that compose our dataset dissociate in the Fv region and in the C H 1–C L at different times (SI Fig.…”

Bispecific antibodies—effects of point mutations on CH3-CH3 interface stability

“…The melting temperature is commonly used as a descriptor to quantify protein thermal stability. The complex network of protein interactions plays an important role in thermostability, and both attractive and repulsive interactions have to be considered to characterize this property ( Codina et al. , 2019 ; Folch et al.…”