Protein Loss on HPLC Columns

The answer to the following reader question about protein separations was supplied by LCGC’s “Directions in Discovery” columnist Tim Wehr.

Q: This question is about protein liquid chromatography (LC) methods for impurity quantitation. With the LC methods that we use (this includes reversed-phase, size exclusion, and ion-exchange chromatography), low recoveries at lower protein amounts loaded are obtained. For example, when 15 µg of protein is loaded on a reversed-phase HPLC column with UV absorbance detection at 215 nm, the area response obtained is near 15,000 mAU. This response is linear in the range of 80–120% for a 15-µg sample. Further, we assess the response at lower amounts to establish a dose response for (lower) ranges — that is, the amounts at which impurities would be present. We have noticed that for these methods as these lower amounts are injected, the area response obtained (recovery?) is lower than expected. What could be causing this decreased response?

Tim Wehr: Loss of protein at low concentrations on HPLC columns is a common problem. The usual cause is the presence of active sites on the packing surface which bind to proteins. These sites are saturated when injecting large amounts of protein, so the protein loss is negligible relative to the total amount of protein in the injection. However, when small amounts of protein are injected, the loss by binding to active sites can be significant. There are several ways of dealing with this problem:

•Make sure you are using a column made with high-quality type B silica. This material is of high purity, and contains very low levels of contaminating metals. Metalscan bind protein directly, or can lower the pK values of adjacent silanols so that they can participate in ion exchange interactions. Most major column manufacturers (such as Agilent and Waters) use type B silica.

•Try injecting a large amount of a “conditioning” protein before using a new column. This should be a protein other than your analyte protein. For example, many chromatographers make several injections of bovine serum albumin on a new column before using it for analytical separations of their protein samples. I work exclusively with peptides, and before I put a new column into use, I condition it with several injections of a BSA tryptic digest.

•If your analyte protein is known to have binding sites that could participate in strong interaction with the column, you can modify your chromatographic conditions to minimize binding. For example, if your protein is known to bind a metal, you might add a small amount of metal to the mobile phase. Alternatively, you might add a small amount of a chelating agent such as EDTA to the mobile phase.

In general, protein recovery is improved by operation under denaturing conditions. Denaturation is favored by operation at elevated temperature (for example, at 60 °C), use of acidic mobile phases (for example, pH 2), and addition of chaotropes (for example, urea or guanidine) to the mobile phase. Protein recovery can be compromised by inadequate equilibration of the column at initial gradient conditions before injection. Note that the use of elevated temperatures can reduce the lifetime of most silica-based columns, and a polymeric packing may be preferred. Use of chaotropic mobile phase additives particularly favors recovery of hydrophobic and high molecular weight protein.