In a webcast this week addressing the state of the protein separations field, a panel that included Ruedi Aebersold, Barry Karger, Bruno Domon, and Jean-Pierre Chervet praised multidimensional liquid chromatography and MSn, while disagreeing over the significance of emerging separations technologies such as microfluidics and high abundance protein depletion; as well as the future of 2D gels.
The webcast, a production of Reed Business Information entitled, “Evaluating Protein Separation Methods for Mass Spectrometry,” is archived here.
All the panel members agreed that no one system was yet sufficient for solving problems such as dyn-amic range, sensitivity, and throughput bottlenecks that current methods of protein separations attempt to address, but all felt that some combination of multidimensional chromatography, coupled with MS/MS analysis, was a good start. Aebersold expressed his enthusiasm for multiple rounds of MS/MS as a good solution for comprehensive sample coverage. He presented slides showing how, if one were to digest the yeast proteome, subject it to cation exchange and reverse phase chromatography plus Advion affinity minipreps, and then do MS/MS analysis, 36 to 50 percent of the peptides identified in each fraction would overlap with peptides identified in other fractions. This redundancy, Aebersold said, could be reduced by doing multiple rounds of successive MS-MS. The problem with this method is that the efficiency of MSn analysis is still generally too low to make this a desirable solution in all cases. “We need to fine tune the performance of MS with the performance of separations,” Aebersold said.
Chervet suggested that the use of linear rather than step gradients in liquid chromatography could also address peptide redundancy and reduce MS time — an issue for labs where it is not an option to spend the time and resources necessary to do multiple tandem MS rounds for every fraction. This method of performing a continuous gradient that combines the multiple dimensions of the chromatography steps without separate salt steps would also increase throughput, according to Chervet.
act was required to address this problem. The more dimensions the researcher adds up front, the more efficient the separation is and therefore the more efficient and complete the MS analysis is. At the same time, “with more and more dimensions of separation, the time element comes in because there are more samples to analyze,” Karger said. One method of increasing throughput, Karger suggested, was to decouple separations from analysis, perhaps using a MALDI rather than an ESI interface. “With MALDI, we can deposit samples by spotting or streaking, then do single stage MS analysis to survey the whole sample, or do ICAT for quantitation — then we can make decisions for which spots to do MS/MS on,” Karger said. “When using MALDI, we have the possibility of multiplexing separations, which can increase throughput.”
An issue that Chervet, Karger, and Aebersold all weighed in on was how to address the problem of dynamic range. Chervet and Karger pushed the importance of removing high abundance proteins such as albumin and IgG — which together account for about 80 percent of the proteins in human serum — up front, in order to clear the way for looking at those proteins present in lower levels. While Chervet said his group was working on affinity columns to deplete IgG and HSA, Karger favored the use of Agilent’s Multiple Affinity Removal System, which removes the six highest abundance proteins from serum samples (see PM 8-15-03). “After doing 2D LC-MS we found no albumin peptides — which means we at least eliminated a great deal,” Karger said.
In response to a question from an audience member about the dangers of losing important information during the removal process, Karger replied, “It is correct that you do take losses when you are doing a depletion experiment. The reality is that it’s a cost benefit, and we feel at this stage that the benefit outweighs the cost.” Aebersold did not agree that removing the proteins was the only solution; he once again pushed thorough fraction-ation and the proper exploitation of mass spec technology as potential alternatives. “Rarely do we actually exploit the full dynamic range of mass spec, because range is limited by sample complexity,” Aebersold said. “We need to select a subset that is of biological interest so that the proteins can be identified from a relatively simple sample mixture.”
On the subject of microfluidics and chromatography on a chip, the panel members said that the technology had not yet sufficiently addressed the following challenges: creating an efficient mass spec interface, reconciling the chip’s low volume with the mass spec’s sensitivity limitations, and dealing with surface area and ratio issues.
The 2D gel as a primary separations technique had few defenders at the panel, although Karger insisted that it was “alive and well,” and was complementary with other separations techniques. Aebersold also acknowledged that the technique continued to fill a niche, denying an audience member’s suggestion that ICAT could eventually replace gels entirely: “If you’re talking about resolving specific protein isoforms, then I think the 2D gel will for quite a long time be the best tool to address the question,” he said. Yet Aebersold challenged the notion that separating out the various isoforms was a worthwhile exercise in the end. He asked, “Once you’ve resolved albumin into 50 isoforms, is there any useful information in that?”