After recombinant protein purification, the endotoxin level is too high. Which removal kit provides good recovery?

Your sample is eluted from the nickel column, and the imidazole concentration is critical. Many anion-exchange endotoxin removal resins show reduced LPS binding capacity under high salt or imidazole conditions, and your recombinant protein may be co-adsorbed if it carries a negative charge. It is recommended to first switch to pyrogen-free PBS or Tris buffer and reduce imidazole to below 20 mM before passing through the endotoxin column. Affinity-type ligands such as polymyxin B have slightly better imidazole tolerance, but the instructions should still be checked. Kits based on different principles also show preferences for endotoxin subtypes; for E. coli-derived LPS, either polymyxin B or histamine ligands can be considered. In terms of recovery, affinity-type resins generally show lower adsorption of conventional proteins but have low binding capacity; ion-exchange-type resins have high binding capacity but require optimization of pH and salt concentration.

When comparing suppliers, I usually ask the sales rep to provide the following: the EU binding capacity for endotoxin per mL of medium, whether they have data on recovery rates for similar recombinant proteins, compatibility with imidazole/reducing agents, and whether NaOH regeneration is possible. For in-stock items, don't just look at the product page—ask about the actual shipping location and whether a batch-specific CoA is available. Domestic manufacturers' prepacked columns often have stock, while imported ones may face customs delays; I've seen 3–7 days or longer, so base it on actual inventory. For budgeting, don't just calculate the column unit price—pyrogen-free tips, buffers, endotoxin detection reagents, and disposable columns all need to be included. For your 5 mL sample, a 1 mL column might be enough for a single run, but if two treatments are needed, compare the unit cost of buying one 5 mL column versus two 1 mL columns.

Operational details often affect recovery more than the kit itself. Columns, collection tubes, pipette tips, and buffers must all be depyrogenated; do not use ordinary EP tubes. Before passing the sample through the column, 0.1%–0.5% non-ionic detergents such as Triton X-114 or octyl glucoside may be added to help depolymerize endotoxin micelles, but kit compatibility must be confirmed. During collection, discard the column dead volume first; otherwise, dilution will affect protein concentration. After treatment, always perform protein quantification and LAL in parallel; do not look only at whether endotoxin has decreased, as the target protein may be adsorbed or precipitated. Endotoxin spike recovery is recommended to avoid low-concentration detection bias. If a single treatment does not meet the specification, replace with a new column or reduce the previous loading amount; repeatedly passing through the same column is not recommended.

5 mL and 1 mg/mL are not particularly small volumes. I recommend prioritizing column-based formats, especially 1 mL or 5 mL prepacked columns, which can be run by gravity or centrifugation. The advantages of columns are that the loading volume is matched, the flow rate is controllable, and the binding capacity is usually higher than that of magnetic beads. For samples with tens of EU/mL, they can generally be reduced to single digits, but the prerequisite is that the flow-through rate must be slow. Magnetic bead formats are more suitable for precious samples from several hundred microliters to 1 mL. For 5 mL, either split the sample into multiple runs or add a large amount of magnetic beads, which will increase the cost per run. In terms of recovery, there is no absolute answer; it mainly depends on nonspecific adsorption between the protein and the ligand. For the first run, you can take 0.5 mL and pass it through the column, measure A280 and LAL, and then decide whether to process the remaining sample.