NEW YORK (GenomeWeb) – Researchers at the Max Planck Institute of Biochemistry have developed a mass spectrometry workflow that can identify up to 10,000 proteins in a 100-minute single-shot experiment.
Presented in a paper published this week in Nature Methods, the approach, named the BoxCar acquisition method, adjusts the sampling of ions at the MS1 level during mass spec analysis to expand the instrument's dynamic range and sampling depth. According to Florian Meier, the study's first author and a graduate student in the lab of Max Planck researcher and senior author Matthias Mann, the approach significantly boosts the sensitivity and reproducibility of conventional data-dependent acquisition mass spec.
He added that it could also potentially be coupled to data-independent acquisition (DIA) workflows, enabling deeper proteome coverage there, as well.
As the authors note, shotgun proteomic experiments consist of two scans, an initial MS1 scan that measures the mass-to-charge ratio and signal strength of the peptides, and then an MS2 scan that analyzes the fragmentation pattern of the peptides, allowing researchers to match them to patterns in peptide databases and make identifications.
A major challenge for proteomics, particularly in experiments using complex samples with high dynamic ranges, is that only a small proportion of peptides in a sample are selected for analysis. This means that high-abundance peptides are overwhelmingly selected for, making identification and quantification of lower-abundance molecules challenging and highly variable across different samples.
One approach to tackling this problem is improving MS2-level analysis by developing methods capable of fragmenting and analyzing a higher proportion of the precursor ions introduced into the mass spec. This, Meier noted, has been a primary focus of mass spec method development in recent years, leading to, for instance, various data-independent acquisition approaches.
Improvements to the MS1 level of analysis have been a lesser area of focus, Meier said, but he noted that there is substantial room for improvement in this area.
In their BoxCar paper, Meier and his colleagues focused on this MS1 level, and the limitations of the ion storage device (C-trap) used in some of Thermo Fisher Scientific's Orbitrap instruments, specifically.
In these instruments, ions are generated by electrospray and then passed into the C-trap before moving into the Orbitrap for analysis. However, according to the authors, the C-trap can store only 1 million charges at a time, which, they noted, is around 1 percent of the ions generated during its fill time, meaning that around 99 percent of ions generated are never analyzed. Because high-abundance ions are overrepresented in the overall sample, they will also be overrepresented in the 1 percent of the ions that are ultimately analyzed, crowding out lower-abundance molecules.
Importantly, because a large proportion of ions filling the C-trap will be the same few high-abundance species, the m/z range captured in the trap will be heavily weighted towards the m/z of those particular high-abundance species. Given this, the Max Planck team realized that by using the upfront quadrupole of Orbitrap instruments like the Q Exactive or the Orbitrap Fusion, they could select ions from a wide variety of m/z ranges to fill the C-trap, thereby diluting the presence of the high-abundance species and allowing a larger number of low-abundance peptides into the trap.
Dividing the ion current into ten different m/z segments consisting of 100,000 charges each, the researchers found in an analysis of a HeLa cell digest that low-abundance peptides showed 30- to 60-fold improvements in signal-to-noise.
Applying the method to human plasma, the researchers found that it provided an additional order of magnitude of dynamic range, a notable boost in performance, given that plasma can exceed ten orders of dynamic range.
They also found it helped substantially with the "missing data" problem that has limited the usefulness of shotgun proteomics in work like clinical biomarker research, where reproducible quantification across large numbers of samples is key. Because shotgun experiments select ions for analysis stochastically, some ions quantified in one run will not be quantified in another run, leading to missing values that make comparisons across runs challenging. This is particularly an issue for lower-abundance peptides.
In an analysis of 10 HeLa cell digests using 45-minute mass spec runs, the researchers found the BoxCar method quantified 7,222 proteins per run, 6,216 of which were quantified in all 10 runs. An equivalent experiment using a conventional shotgun method quantified 5,050 proteins, 4,180 of them in all 10 runs.
Meier and his colleagues also developed an approach that enabled extremely deep proteome coverage by combining the BoxCar method with a pre-built peptide library, generated by extensive fractionation of the sample of interest, followed by mass spec analysis. By matching the MS1-level data generated experimentally using the BoxCar approach to the MS2-level peptide ID data in the pre-built library, they were able to identify an average of 7,775 proteins per 45-minute run. Applying this approach to a study of mouse cerebellum, they were able to identify an average of more than 10,000 proteins across five 100-minute mass spec runs, 9,270 of which they identified in all of the five replicates.
Such an approach could be particularly useful for clinical work, Meier said, noting that it allows for high-throughput analysis with great depth of coverage. He added that the Mann lab has developed streamlined approaches to generating peptide libraries that have made the time required for this step "negligible."
"It's now no longer a big deal to generate a very deep library of a sample by fractionating," he said. "If you have a lot of samples, then the time you spend on generating libraries is not a lot compared to the time you can spend on actually quantifying your samples."
"That's the workflow we have set up for clinical samples in particular," he added.
While the Nature Methods paper focuses on the application of the method to standard shotgun proteomic workflows, Meier said it is applicable to DIA approaches where the boost in MS1-level information could improve analyses. It could also help in more targeted mass spec assays where the presence of a particular MS1 signal triggers MS2 analysis, he said.
Implementation of the approach is based on software control of the instrument's quadrupole and does not require any hardware modifications, Meier said. The lab has been in contact with Thermo Fisher about possibly developing a software tool that researchers interested in the approach could download and use to run their mass specs in this mode.