3D Nanotemplates for Protein Crystallisation



This thesis presents the first experimental investigations into the combined effect of nanoscale surface porosity, narrow pore size distribution and surface chemistry, on nucleation and crystallisation of proteins. Co-operative self assembly of surfactants and silica was applied to prepare 3D nanotemplates with pore diameter 3.5±1.0nm, 5.5±1.5nm, 11.0±3.0nm, 16.0±3.0nm and 22.0±5.0nm. Post synthesis functionalisation was used to graft surface with -OH, -NH2, -CF3, -C6H5, – Cl and -CH3 functional groups. The relationship between 3D nanotemplate, N2 sorption and TEM pore diameter, XPS surface composition, wettability, surface charge and protein physicochemical properties was investigated, resulting in a coherent understanding of the combined effect of nucleant surface porosity and surface chemistry on protein crystallisation. The protein systems investigated for crystallisation include lysozyme, thaumatin, trypsin, albumin, concanavalin A, catalase and ferritin; varying in molecular weight from 14kDa – 450kDa and hydrodynamic diameter ~3-20nm. Crystallisation of proteins was found to be strongly dependent on 3D nanotemplate pore size. The 3D nanotemplate with a pore diameter similar to the protein’s hydrodynamic diameter (Dh) was found to be successful in preferential crystallisation, e.g. albumin (Dh=~5nm) and catalase (Dh=~10nm) was crystallised only on 5.5±1.5nm and 11.0±3.0nm 3D nanotemplates respectively. Here, we report a direct correlation between protein hydrodynamic diameter and 3D nanotemplate pore diameter, key for controlling protein nucleation. The correlation has been experimentally validated for all protein systems investigated. The concept of preferential crystallisation was further developed to investigate the applicability of this methodology for the separation of a target protein from a protein mixture. A solution of two proteins; porcine pancreatic lipase and RNAse was selected for this separation by crystallisation. Lipase (Dh=~4.5nm) and RNAse (Dh=~1.5nm) crystals were preferentially obtained with 5.5±1.5nm and 3.5±1.0nm 3D nanotemplates respectively. Furthermore, crystallisation of lipase from a commercially available crude source, (purity ~20%) containing a mixture of lipase, amylase and protease, was also achieved on the surface of the 5.5±1.5nm 3D nanotemplate. The significance of the work here is the demonstration that only specific nucleant surfaces with narrow pore size distribution can preferentially crystallise a target protein from a protein mixture, which is not possible with nucleants of broad pore size distribution, reported to be most suitable candidate for “universal nucleant”. Crystals of four out of the seven well studied proteins (lysozyme, albumin, concanavalin A and catalase) were obtained at lower protein concentration, whilst thaumatin was crystallised at par protein concentration but 3× lower precipitant concentration, on the surface of 3D nanotemplates functionalised with -OH, -CH3 and -NH2 functional groups. Here, we report that a reduction of 50- 92% protein concentration compared to the lowest reported literature values was achieved with the use of the 3D nanotemplates. As crystallisation was achieved for all seven protein systems at protein concentrations in the range of 2-20mg/mL, this approach can be used to engineer surfaces for preferential crystallisation of a target protein directly from industrial bioreactor broths. In summary, the findings of this thesis offer a first systematic approach for controlling the nucleation and crystallisation of biological macromolecules. The use of 3D nanotemplates offers the possibility of crystallising complex proteins (e.g. enzymes, antibodies, protein complexes, DAbs, MAbs) for structural determination and also novel crystallisation routes for downstream bioseparations.