NEW YORK (GenomeWeb) – Researchers from the Max Planck Institute of Biochemistry, University of Copenhagen, and liquid chromatography firm Evosep Biosystems have developed a mass spec-based proteomics workflow capable of running more than 200 samples per day.
Described in a paper published this month Molecular & Cellular Proteomics, the approach uses the efficiencies of Evosep's LC technology to enable high-throughput proteomic experiments, said Philipp Geyer, a postdoctoral researcher in the group of Max Planck professor Matthias Mann. Mann is an indirect investor in Evosep.
The Evosep system, named the Evosep One, uses pre-formed LC gradients to address the various difficulties inherent in LC separations for proteomics experiments. As the MCP authors note, the high complexity of proteomic samples requires high levels of LC separation upfront of mass spec analysis, and this, in turn, has forced researchers to use narrower, longer, and higher-pressure LC setups, all of which create challenges in terms of system robustness, reproducibility, and throughput.
This has proved a particular limitation for clinical and pre-clinical proteomic applications, where LC-MS systems need to run large numbers of samples in a routine environment and, ideally, with high levels of reproducibility across different sites.
The Evosep One uses specialized solid phase extraction tips the firm calls Evotips, through which researchers elute sample peptides at low pressure and high flow rates, forming a gradient that is then captured in a capillary loop. This pre-formed gradient can then be run on a nanoflow LC column at high pressure and the separated sample injected into the mass spec.
The ability to preform the LC gradient at low pressure and high flow volume eliminates much of the robustness, reproducibility, and throughput challenges around nanoLC, the authors wrote, while preserving the sensitivity and performance provided by nanoLC.
In addition to its work with the Max Planck and Copenhagen researchers, Odense, Denmark-based Evosep has in recent months announced collaborations with Bruker and Thermo Fisher Scientific to integrated their systems for high-throughput proteomic applications.
In the MCP study, the researchers tested the robustness of the Evosep One system by running 1,500 consecutive samples, finding that the performance of the system remained consistent across this sample set.
They also combined the system with an automated sample prep platform that loaded prepared samples into the Evotips in a 96-well format and analyzed these samples at a rate of 60 per day. Measurements across the 96 samples were highly reproducibly, with a median Pearson correlation coefficient of .98.
To demonstrate the potential throughput of the system, the researchers used 5.6 minutes LC gradients to analyze a proteomic standard kit containing 48 different proteins, 44 of which they quantified across all 200 samples they tested. The remaining four went unidentified in all runs, which led the authors to surmise they "were likely missing from the kit."
The 5.6 minutes gradient represents a 200 samples per-day capacity and could prove useful for high-throughput analysis of low-complexity samples like those used to identify single proteins from gel bands or a small set of proteins from pulldown experiments, the authors noted.
Upping the LC run time to 21 minutes (a rate equivalent to 60 samples per day), the researchers analyzed a HeLa digest split into 43 fractions, identifying roughly 2,700 proteins per fraction and 9,918 proteins total.
These experiments were done using data-dependent acquisition mass spec, but Geyer said he and his colleagues have recently become more interested in data-independent acquisition approaches, which he suggested could prove better suited for proteomic experiments using very short LC gradients than DDA methods.
DIA mass spec selects broad m/z windows and fragments all precursors in that window, which allows the instrument to collect MS/MS spectra on all the ions in a sample. This is in contrast to DDA methods which stochastically select ions for fragmentation.
Because in DIA the instrument is looking at the same peptides in every sample, quantitation is typically more reproducible across samples than in DDA, which has led proponents to promote it for work like protein biomarker research where reproducible quantitation of the same proteins across many samples is key.
Geyer said the fact that DIA fragments all the precursors in a given m/z window also makes it well-suited to experiments with short LC gradients, given that such experiments produce large amounts of MS/MS spectra that can overwhelm the sequencing speed of an instrument run in DDA mode.
"With really short gradients you're getting a lot of MS2 spectra… and with DDA you're limited by the sequencing speeds," he said. "You pick one peptide, you fragment it, you pick the next peptide, you fragment it."
"But with DIA, you always fragment a whole window, so you can have multiple fragment spectra in this one window, and so this is a big benefit [for analyses using short LC gradients]," he added.
Geyer noted that Mann's lab has not traditionally used DIA approaches extensively but added that he feels that recent improvements in mass spec performance have increased the usefulness of DIA techniques to where he plans in his group with the larger Mann lab to use it for the majority of his projects. He said he is running his DIA experiments on a Thermo Fisher Q Exactive HF-X instrument and using Biognosys's Spectronaut software for data analysis.
In the MCP paper, Geyer and his colleagues used DIA and the Evosep One system running 21 minutes gradients to quantify more than 5,000 proteins in HeLa digests. Looking across five samples, they quantified 4,904 proteins in all five runs, 3,286 of which had coefficients of variation of less than 20 percent.
Geyer said that his particular interests focused on using the Evosep-DIA combination for high-throughput analysis of plasma samples. Due to its higher complexity compared to cell digests, plasma is a more challenging sample source, and, Geyer said, the 21-minute gradient experiment that quantified 5,000 proteins in HeLa digests can measure around 300 proteins in undepleted human plasma.
While this isn't particularly deep coverage, these 300 proteins include around 60 US Food and Drug Administration-approved protein biomarkers, Geyer said, noting that this suggests potential clinical applications.
He added that even 300 proteins measured with high-throughput like 60 samples a day could be useful in enabling larger-scale protein biomarker studies.