Affinity purification of viral protein having heterogeneous quaternary structure: Modeling the impact of soluble aggregates on chromatographic performance

Prokaryote-expressed polyomavirus structural protein VP1 with an N-terminal glutathione-S-transferase tag (GST-VP1) self-assembles into pentamer structures that further organize into soluble aggregates of variable size (3.4 × 10²–1.8 × 10⁴ kDa) D.I. Lipin, L.H.L. Lua, A.P.J. Middelberg, J. Chromatogr. A 1190 (2008) 204]. The adsorption mechanism for the full range of GST-VP1 soluble aggregates was described assuming a dual-component model T.Y. Gu, G.J. Tsai, G.T. Tsao, AICHE J. 37 (1991) 1333], with components differentiated by size, and hence pore accessibility, rather than by protein identity. GST-VP1 protein was separated into two component groups: aggregates small enough to access resin pores (LMW: 3.4 × 10²–1.4 × 10³ kDa) and aggregates excluded from the resin pores (HMW: 9.0 × 10²–1.8 × 10⁴ kDa). LMW aggregates bound to resin at a higher saturation concentration (29.7 g L⁻¹) than HMW aggregates (13.3 g L⁻¹), while the rate of adsorption of HMW aggregates was an order of magnitude higher than for LMW aggregates. The model was used to predict both batch and packed bed adsorption of GST-VP1 protein in solutions with known concentrations of HMW and LMW aggregates to Glutathione Sepharose HP resin. Asymmetrical flow field flow fractionation with UV absorbance was utilized in conjunction with adsorption experimentation to show that binding of HMW aggregates to the resin was strong enough to withstand model-predicted displacement by LMW aggregates. High pore concentrations of LMW aggregates were also found to significantly inhibit the diffusion rate of further protein in the resin pores. Additional downstream processing experimentation showed that enzymatic cleavage of LMW aggregates to remove GST tags yields more un-aggregated VP1 pentamers than enzymatic cleavage of HMW aggregates. This model can be used to enhance the chromatographic capture of GST-VP1, and suggests an approach for modeling chromatographic purification of proteins that have a range of quaternary structures, including soluble aggregates.