Protein ion exchange chromatography: Effects of solute size, adsorbent pore structure, and protein charge on solute elution

We are interested in the problem of relating protein elution times in ion exchange chromatography (IEC) to quantifiable features of protein and ion exchange resin structures. Developing a purification protocol can be a lengthy process of screening numerous resins and mobile phase operating conditions. Existing models of protein elution are correlative tools, requiring experimental data to determine lumped parameters that capture the system's ionic and three-dimensional structural interactions; these parameters can not be predicted a priori or readily extended to other proteins, experimental conditions, or ion exchange resins.; A macromolecular retention model (MRM) was developed to investigate solute size (hydrodynamic radius, Rh) and adsorbent structure (cylindrical pores with lognormal distribution of pore radii; Rlogmean, sigma). Analytical relationships were developed for size exclusion retention times (t0) and column phase ratios (phi), traditionally treated as fitting parameters. We can calculate retention model parameters for a given size solute once the adsorbent pore structure has been characterized. A key result of the modeling was the prediction of "solute size effects" that can complicate data analysis. Different sized solutes with the same solute-adsorbent interaction energies can elute with very different retention times; proteins eluting with the same retention time can have very different interaction energies.; We devised a novel experimental system to investigate protein charge and charge nonuniformity that eliminated the potential for solute size effects. This is the first study utilizing a series of structurally homologous proteins differing so dramatically in ionic character (net charge: -38 to +4; neutral dipole moment: 200--823 Debye). We measured isocratic retention times for seven acid proteases (porcine pepsin A1, porcine pepsin A2, bovine pepsin, chymosin A, chymosin B, endothiapepsin, mucoropepsin) as a function of mobile phase pH (4--5.5) and ionic strength (0--1 M sodium chloride) on three related size exclusion, cation exchange, and anion exchange resins (Tosoh Biosep HW65S, SP, SuperQ). We concluded that charge alone, whether net charge or regions of high local charge density determined via protein crystal structure analysis, was incapable of explaining the acid protease retention data; non-ionic interactions must be considered.