NEW YORK (GenomeWeb) – Since proteomics emerged as a field some two decades ago, mass spectrometry systems have improved steadily, with each new generation of instruments offering gains in accuracy, resolution, speed, and other measures.
These gains have greatly enhanced researchers' ability to characterize proteomes, enabling, among other things, deeper coverage and higher throughput. In recent years, however, these mass spec advances have outpaced improvements in other parts of the proteomics workflow. According to one leading proteomics researcher, these other, non-mass spec components — liquid chromatography, in particular — are now the limiting factor in certain kinds of proteomic experiments.
In a commentary published last month in Cell Systems, University of Wisconsin-Madison researcher Josh Coon and his co-authors argued that improvements in peptide separation, which is currently done primarily by LC, are necessary for researchers to take full advantage of existing and future mass spec technology.
This observation stems from recent work by Coon and his colleagues that used Thermo Fisher Scientific's Orbitrap Fusion Lumos instrument, which offers a significant jump in performance compared to the previous Orbitrap machines.
As the researchers noted, in 2013, Coon's lab identified 4,000 yeast proteins in a one-hour mass spec experiment using the Orbitrap Fusion instrument, which, at the time, represented a roughly fourfold increase in speed compared to the field's previous best efforts.
More recently, he and his colleagues repeated the experiment using the newer Fusion Lumos instrument, which can collect MS/MS spectra at a rate of more than 40 Hz, double the speed of the original Fusion. However, they wrote, their results were only "minimally improved."
"We were hopeful we would get more out of the new instrument," Coon told GenomeWeb this week, "but we didn't see the benefit that you might expect to get from essentially doubling the scan rate."
Pondering why they didn't see identifications increase at the rate they had hoped for, Coon and his colleagues hit upon the limitations of their LC system.
The complexity of proteomic samples requires that they be separated in some way prior to mass spec analysis. This is commonly done using LC and, until recently, the speed with which LC systems could separate and deliver peptides to the mass spec instrument outpaced the capacity of the instrument to analyze them.
For instance, a 2011 study by Max Planck Institute researcher Matthias Mann, cited by Coon and his co-authors in their commentary, estimated that a typical LC-MS experiment produced around 100,000 peptides, and that to collect spectra on all of them, a mass spec instrument would need an MS/MS scan rate of 25 Hz.
The Fusion Lumos, Coon noted, exceeds that scan rate, suggesting that the instrument's capacity cannot be fully utilized with standard chromatography systems.
To demonstrate this point, the UW-Madison researchers built their own higher performance LC columns in house, finding in an analysis of whole yeast digests that this improved chromatography and more than doubled the number of peptide sequences they were able to identify.
The results indicate that for certain kinds of proteome profiling experiments, improvements in LC are more urgently needed than advances in mass spec instrumentation, said Evgenia Shishkova, a graduate student in Coon's lab and the first author on the paper.
"For the longest time, when people wanted better results, they have turned to investing more in [mass spec instrumentation]," she said. "But at this point, we have reached the limit of what we can get out of the [mass spec without LC improvements.]"
"There is a general understanding that chromatography is important," Shishkova said. "People have been investing in research into chromatography and improving chromatography. But I think the fact that it is now a limiting factor might have snuck up on people with this new [mass spec] instrument."
"Orbitrap systems, for example, have only in the last few years really been able to expose this problem with chromatography," Coon said. "The generations that were out four years ago, according to our analysis in the paper, were just fine using what we have. It's only recently that any system has been able to really get to the scan speed where we can now say, 'Well, double it again and it probably won't help with this particular problem that much.'"
He noted that the issue is particularly challenging for the sort of high-throughput, single-shot experiments his lab is pursuing in work like its one-hour yeast proteome effort. Many proteomics workflows use multiple dimensions of chromatography or extensive fractionation to allow researchers to go deeper into their samples, and in these cases, existing chromatography remains sufficient.
Coon, however, is interested in developing workflows using no fractionation and a single dimension of chromatography, the idea being that such methods could allow for extremely high-throughput proteomic profiling. In addition to their one-hour yeast proteome method, he and his colleagues have been refining a similar approach for profiling the considerably more complex human proteome. Currently, they are able to measure more than 8,000 proteins in under four hours in human samples.
"The ball, in a big sense, is back in the court of LC," said James Jorgenson, professor of chemistry at the University of North Carolina Chapel Hill and an LC expert who has worked with Coon's lab on high performance LC systems.
For researchers like Coon who are developing workflows using single-dimension chromatography, the solution to the LC issue essentially comes down to use of higher pressure, higher resolution systems, Jorgenson said.
"To try to squeeze out more peaks, better resolution out of a single column, there's almost no way around it," he said. "It requires longer columns, higher pressure, smaller particles, and then real attention to getting the best packing of those smaller particles you can."
"There's still plenty of room to make progress there," Jorgenson noted, though he added that it was unclear what the broader market might be for such higher pressure systems, and whether it was large enough to make it a priority for leading commercial LC firms like Waters and Agilent.
"Commercially, the pressure limit took a big jump up about 15 years ago," he said. "It had been limited to about 6,000 PSI, and now it's right around 20,000 PSI. The question is, will it take another jump eventually or just continue going up incrementally?"
Applications outside single-shot proteomics would also benefit from improved separations, Jorgenson said, but such systems would be fairly challenging to put together from a technical perspective.
"It's a pretty tough technology," he said. "You have to be able to make column connections, do injections, have really reliable pumping and valving that works at maybe 40,000, 50,000, 60,000 PSI. And it gets harder as you go up."
Jorgenson said he didn't envision commercial LC instruments going above the 60,000 PSI mark. "It would truly get difficult at that point in terms of materials and wear and tear and things like that."
He noted, though, that LC pressures in the 40,000 to 60,000 PSI range were doable and that this level of performance would likely be sufficient to take advantage of the full capacity of current cutting edge mass spec instruments.
"In our lab, we are working at 40,000, 50,000, 60,000 PSI kind of pressures with one-dimensional separations and we have been transferring that technology to other people's labs including [Coon's]," he said.
"To push further than where we are, we've got to innovate in the chromatography space," Coon said. "I don't think we really offered any solutions on how to do that innovation. But we're pointing out that, hey, this is probably the number one thing we can do right now."
On a related note, researchers investing in new mass spec systems "might want to pay careful attention to how they are doing separations," he added. "There is a broad range of quality in the separations that are being implemented out there. At least try to get yourself on the high end of what is possible right now."