Protein purification is more than monkey business

One client of ours – a small biotech company with five employees – was experiencing extremely low yields of their recombinant protein during purification. Upwards of 25–30 repetitive purification steps were needed to collect enough mass of the desired product. They were unable to diagnose the origin of problem through the use of traditional protein characterization methods such as western blot or size exclusion chromatography.

At SOLVE, through the use of asymmetric flow field-flow fractionation (AF4), we were able to pinpoint that the properties such as size and molecular weight of the protein were dramatically different in two different buffer systems (see Figure). In their desired buffer system (red in Figuer), the recombinant protein was found to be aggregated. It possessed a broad and extremely high molar mass distribution and an average radius of 140 nm. On the contrary, in 6 M urea the protein was found to be mostly in the desired monomeric state. With this information the company was able refine their purification protocol and realize increased protein yields of 10-15 fold. This allowed them to meet their timeframe for submitting for primate based efficacy studies. Follow up revealed that they experienced successful efficacy studies and were applying for additional funding. An AF4 centric approach was chosen for this study because it allowed the properties of the protein to be investigated in desired solution environments while the gentle nature of the separation prevented disruption of the protein aggregate structures, which were previously undetectable by other means.

The potential instability of proteins in solution could lead to a variety of undesirable outcomes when designing biological based therapeutics. One possibility is that the protein remains in solution but has unexpected properties and behavior which can adversely affect downstream processing and manufacturing. Comprehensive characterization strategies can help to overcome these challenges.

Fractogram from AF4 showing the molar mass distribution of the sample (time is elution time from the AF4 channel). The protein’s physicochemical properties were different in 6M urea (blue) compared with the company’s desired buffer (red). About 70 % of the protein was monomeric in 6 M urea (average molar mass for monomer: 80 kDa and aggregate: 510 kDa), while polydisperse large size networks were observed in the desired buffer (average molar mass: 22 200 kDa, average radius of gyration: 140 nm).