Lotus Technical Data

Preliminary Data. BMB has been developing the Lotus platform with the goal of identifying and eliminating host proteins that interfere with standard purification processes. We have focused on ion exchange chromatography because of its widespread use in the biopharmaceutical industry and its limited cost. Using this chromatography technique, we sought to identify the top contaminants in E. coli overexpressing various recombinant proteins including afGFPuv, CD2, CD3, CD2a and a proprietary protein from the Mayo Clinic and evaluate the impact of removing these proteins on capture and purity.

Identified top contaminants. We initially cataloged E. coli’s separatome—the collection of proteins that bind to the purification column. To do so, we subjected clarified E. coli lysates in the absence of target protein to a diethyl-aminoethyl (DEAE) anion-exchange column, eluted protein fractions at 12 different NaCl concentrations, and subjected each fraction to LC-MS/MS based on established techniques [18]. The resulting dataset revealed the constituents and their relative abundance under each elution condition independent of an overexpressed target. We then formulated an importance score (IS) for each constituent based on the following equation:


where i is the protein constituent, j is the elution window, b1 is a chromatography-specific adjustment factor, yc,j is the concentration of the elution solvent, ymax is the maximum solvent concentration, hi,j is the concentration of protein i in the elution window, hj,total is the total protein concentration in the window, hi,total is the concentration of protein i across all windows, MW is the molecular weight of protein i and α is 1 when the column is saturated. The importance score provides a ranking based on the adsorption strength, adsorption specificity, adsorption abundance, protein size, and extent of column saturation. The summation can be adjusted to account for the windows in which the therapeutic protein will be bound. The resulting rankings indicate which proteins represent the greater barriers to maximizing column binding and the purity of the eluted biotherapeutic. Using this strategy, we identified the top-ranked proteins based on loading in 10 mM NaCl (Top 12 in Table 1), which are target independent.

Table 1


Substantially improved protein capture in multi-deletion strain. We next deleted six of the top-ranked non-essential genes (rfaD, usg, rraA, cutA, nagD, speA) from E. coli MG1655 (bolded genes in Table 1). Note that this strain is often used in biomanufacturing settings over BL21(DE3) because the latter requires induction of T7 polymerase with the expensive inducer IPTG. In comparison to the parent MG1655 strain, the deletion strain showed a 16% decrease in total host cell protein adsorbed when the DEAE column is loaded at 100 mM NaCl, and, despite these deletions, only partially diminished growth (details in Project 1). We next tested the extent of initial capture following overexpression of afGFPuv in the deletion strain and in MG1655. After applying equal amounts of afGFPuv, equal concentrations of afGFPuv, and equal amounts of total protein, we consistently observed capture efficiencies that were 37 – 38% higher for the deletion strain than for MG1655 (details in Project 2).

Project 1: Growth Studies without an Expressed Target

Project 2: Growth Studies with an Expressed Target

What is a Separatome?