The specificity of alternative methods of affinity depletion was evaluated by analyzing the bound fractions on 2-D gels at a load equivalent to that of the initial serum or plasma samples; i.e., bound fractions derived from 100 mg. These results show that the dye-based affinity resin contained the largest number of non-specifically bound proteins, since this fraction should contain only albumin and IgG (Fig. 3A). Similarly, the 2-D gel of the bound fraction from the albumin/IgG immunoaffinity spin column showed extensive nonspecific removal of additional proteins (Fig. 3B). This fraction contained nearly as many proteins as the dye assay bound fraction, suggesting that improved purification buffers are needed to minimize nonspecific binding. In contrast, all major protein spots observed in the bound fraction from the MARS HPLC column were the six targeted proteins (Fig. 3C). Furthermore, most of the moderate- and low-intensity spots were tentatively identified as proteolytic fragments of these six proteins based upon comparisons to the 2-D plasma protein map on the Swiss-Prot website (http://us.ex-pasy.org/cgi-bin/map2/def?PLASMA_HUMAN). The more specific removal of targeted proteins by the MARS HPLC column compared with the albumin/IgG immunoaffinity spin column cannot be due to clonality, since both are polyclonal antibodies. However, there may be differences in affinity of the antibodies that affects specificity and, in addition, the methods for purifying the polyclonal antibodies may differ. Another factor is the type of resin and cross-linker used and a final consideration is the purification buffers used for protein depletion. The proprietary purification buffers associated with the MARS HPLC column purification scheme have apparently been well optimized for disruption of weak nonspecific interactions. However, these buffers are apparently not necessarily universally suited for major protein depletion using immunoaffinity resins because when we used the MARS purification buffers with the prototype albumin/IgG immunoaffinity column, the binding capacity was greatly reduced and most proteins, including the majority of albumin and IgG, passed through the column. Subsequent analysis of the bound fractions on 2-D gels showed little binding of albumin and IgG, and many other nonspecific proteins adhering to the column resin (data not shown).
Fig. 3 Evaluation of bound fractions after depletion. (A) Cibacron blue/Protein G bound fraction after column stripping. While this column is meant to deplete albumin (arrow) and IgG (ellipses) only, there are a number of other low-abundance proteins comigrating with the major proteins. (B) Dual-antibody immunoaffi-nity (albumin and IgG) bound fraction displays
Fig. 3 Evaluation of bound fractions after depletion. (A) Cibacron blue/Protein G bound fraction after column stripping. While this column is meant to deplete albumin (arrow) and IgG (ellipses) only, there are a number of other low-abundance proteins comigrating with the major proteins. (B) Dual-antibody immunoaffi-nity (albumin and IgG) bound fraction displays a similar degree of nonspecific binding of proteins other than albumin and IgG as that observed with the dye-based affinity system. (C) MARS HPLC antibody column shows a much cleaner bound fraction; the six targeted proteins (arrow for albumin, enclosed in ellipses for other proteins) can be seen with few other nonspecific proteins.
Fig. 4 Effect of Top-6 depletion on detection of low-abundance protein spots. Left panel - a convenient spin column version of the MARS column was used to increase the throughput; the equivalent of 1 mg of human serum was separated on 3-10L 2-D gels and silver stained. Spots were selected that were not detected on a reference gel loaded with 100 mg of non-depleted sample. Also certain spots were chosen that were thought to be residual albumin
(2, 3) and transferrin (12, 13) to verify completeness of depletion. Right panel - the equivalent of 2 mg of human plasma depleted using the MARS HPLC column. New spots (circles with numbers) selected at this higher load compared with undepleted sample were selected. All spots selected were in-gel trypsin-digested and analyzed using a linear ITmass spectrometer.
Was this article helpful?