NEW YORK(GenomeWeb) – Researchers at the Max Planck Institute of Biochemistry have devised a mass spec-based phosphoproteomics platform that they said allows for high-throughput quantification of phosphoproteomes at depths of greater than 10,000 phosphosites.
Using the method, named EasyPhos, the researchers generated time-resolved phosphoproteomics data following insulin signaling in mouse livers, identifying several novel aspects of insulin signaling.
Comparing their findings to previous efforts in which they used traditional phosphoproteomic workflows, the team reported that with the EasyPhos approach they were able to identify three times the number of phosphosites in a third of the time while using one-tenth the amount of sample.
Key to the EasyPhos workflow is its avoidance of sample fractionation and its use of buffers that don't require the sample to be desalted downstream, said Max Planck researcher Sean Humphrey, one of the method's developers.
Many conventional phosphoproteomic workflows rely on heavy fractionation of samples to achieve depth of coverage. Improvements in mass spec technology and improved specificity in phosphopeptide enrichment, however, allowed the Max Planck team to skip fractionation and instead analyze samples in a single mass spec run.
This, Humphrey told GenomeWeb, is a crucial advantage in that running whole samples let the researchers significantly up their throughput while also reducing sample handling steps that can lead to peptide loss and increased variability.
The researchers also switched to 2,2,2-trifluroethanol (TFE) for their digestion buffer, which, unlike traditional buffers, does not require subsequent desalting steps downstream. This, Humphrey noted, further streamlined the process, as desalting "involved quite labor intensive steps."
Additionally, it helps reduce peptide loss because in some cases peptides are not efficiently re-solubilized after desalting, and hydrophilic peptides in particular may not bind well to the substrates used in desalting. This latter issue is especially a problem for phosphopeptide analysis, Humphrey said, as phosphopeptides tend to be more hydrophilic than peptides in general.
Also key, he noted, is the method's highly specific phosphopeptide enrichment, which, he said, is typically greater than 95 percent. This high specificity improves depth of coverage by reducing the complexity of the proteome being analyzed, meaning that "the mass spectrometer is able to spend more of its time sequencing [the phosphopeptide] data that we actually want to measure," Humphrey said.
It also improved quantitation "because non-phosphorylated peptides are not present as a major background interference matrix when trying to quantitatively compare samples," he added.
The highly-specific enrichment and reduction in peptide losses also allows the researchers to use smaller amounts of sample for their analyses. This is potentially important for phosphoproteomic analysis of clinical samples like tumor tissue, where mass spec has sometimes required too much sample volume to make it a feasible method.
"Requiring less protein starting material certainly makes the workflow more applicable to a wider range of sample types, including real clinical samples," Humphrey said, noting that he and his colleagues have to date worked with sample sizes as low as 100 micrograms.
More generally, he added, the simplification of mass spec-based phosphoproteomics is a step toward increased clinical use.
"I see the simplification and streamlining of the methods, combined with the ability to work with smaller amounts of material as key in making these technologies amenable to the clinic, where large numbers of samples are common and reproducibility and robustness are important considerations," he said.
In their work, detailed in a study published last week in Nature Biotechnology, the Max Planck researchers applied the method to an analysis of mouse liver insulin signaling. Using a Thermo Fisher Scientific Q Exactive instrument, they quantified mouse liver proteomes after treatment with insulin at 11 different time points starting at five seconds after administration of insulin through to 10 minutes after exposure. They did at least six biological replicates for each time point and as many as 10 replicates for the earliest time point, profiling 91 different liver phosphoproteomes.
In all, they quantified 31,605 phosphopeptides and accurately identified 25,505 phosphorylation sites, finding that insulin affects roughly 10 percent of the liver phosphoproteome.
The study required three days for sample preparation and then roughly two weeks for mass spec analysis of the 91 different samples. By way of comparison, Humphrey said, doing the same experiment using conventional workflows would require around 100 days of mass spec time as well as a significant increase in the amount of time needed for sample prep.
The ability to collect phosphoproteomic profiles from large numbers of samples in relatively short periods of time opens up new opportunities for analysis, Humphrey noted. For instance, he said, the ability to look at insulin signaling in the liver with high temporal resolution provided new insights into the mechanisms underlying this process.
He cited as an example the researchers' identification of rapid phosphorylation of the protein Akt at the serine478 residue.
By taking multiple measurements at time points under 30 seconds, "we found that phosphorylation of this site appeared before the nearby activating site (Serine474) that is widely measured by antibody-based approaches," he said. "This was very surprising because it potentially adds another important dimension to the mechanisms governing the activation of this key node in the insulin-signaling pathway."
Another interesting finding enabled by the study's high temporal resolution, Humphrey noted, was the rapid phosphorylation of the Akt substrate FOX01, which, they found, is phosphorylated even more quickly than the activating residues of Akt itself.
"This is in agreement with existing models of Akt phosphorylation and its substrates, which have shown that only a fraction of total cellular Akt activity is required for maximal phosphorylation of its substrates." he said.
Given that FOX01 resides in the cell nucleus, the observation also serves as a reminder, he noted, that the "flow of information between organelles can be very rapid, and that factors other than cellular distribution play important roles in determining how quickly signals are propagated throughout the cell."