Process design and scale-up of counter-current chromatography for the fractionation and recovery of polyketide antibiotics

Regulatory constraints on products of the modern biotechnology industry require final dosage forms to exhibit high and consistent levels of purity. At the same time, financial constraints on process development require cost-effective purification methods to be found. Counter-current chromatography (CCC) is a liquid-liquid chromatographic technique, in which solutes are fractionated based on their selective partitioning between two immiscible liquid phases. The absence of a solid support, as in conventional HPLC, overcomes problems of irreversible adsorption and pore diffusion, and ensures that CCC is a low pressure operation. Although an established analytical scale technique, widespread use of CCC has been hampered by the lack of generic and robust method development strategies and well-engineered industrial scale machines. This project has examined the operation and scale-up of novel J-type CCC devices for application as a generic and scalable high-resolution purification technique. The fractionation of the chemical pharmaceutical erythromycin A (EA) from its structurally similar analogues was used as a test system since this provides a difficult and realistic separation challenge, Initial research addressed the need for a generic method development strategy to increase the speed of CCC phase system selection and identification of the optimal run mode. For the purification of EA, a broad polarity quaternary phase system consisting of hexane/ethyl acetate/methanol/water was identified. Results from a matrix of simple equilibrium partition experiments were used to identify a suitable solvent system for use firstly in gradient elution mode. Based on these results, an optimised reverse phase isocratic separation was then selected which enabled the separation of EA from all of its closely related biosynthetic analogues. Subsequent optimisation studies, using a model erythromycin mixture, addressed the impact of solute loading and mobile phase flow rate on EA purity, column efficiency and throughput in a laboratory scale J-type CCC device. Under optimal conditions (8 mL.min-1; 0.6 g solute loading) a maximum throughput of 0.097 kg.day-1 could be achieved, with an EA purity and yield of 97% (w/w) and 100% (w/w) respectively. Further research focused on the feasibility of using CCC for the purification of EA from real Saccharopolyspora erythraea fermentation broths. Studies here examined the degree of pre-purification required prior to CCC separation. These used feeds consisting of either clarified broth or solvent extracts having undergone either forward or back extraction processes to determine the degree of impurity removal required to ensure a reproducible elution profile of EA. Further studies using a back extracted feed stream examined the effects of CCC mobile phase flow and solute loading on the attainable EA purity and yield. The results in all cases demonstrated a high attainable EA purity (>97% w/w). The results for both model and real systems were subsequently scaled-up using a novel, pilot scale CCC machine. From an understanding of the phase system hydrodynamics, a predictive scale-up model of the separation was established, to describe how solute fractionation at the pilot scale varied with changes in operating variables, such as feed type, mobile phase flow rate and solute loading. Linear scale-up was successfully demonstrated with both model and real erythromycin feed streams, based on the parameters taken from a single laboratory scale CCC chromatogram. Scale-up predictions were accurate to within 5-13% (model system) and 6-10% (real system) depending on the actual operating conditions. Finally, this research explored the successful application of 'Fractionation Diagram theory' as a graphical tool to allow quantification of the trade-off between product purity and solute yield in CCC separations. Combined with a new generation of robust industrial scale machines currently under construction, this work has demonstrated the potential of CCC as a generic and flexible high-resolution separation technique for the modern biotechnology industry.