NEW YORK – University of Chicago researchers have developed a method using mass spectrometry-based thermal profiling to identify protein post-translational modifications that are likely to be functionally important.
Described in a study published this week in Nature Methods, the approach uses thermal profiling to identify PTMs that influence protein stability, which is indicative of functional significance. According to Raymond Moellering, an assistant professor of chemistry at the University of Chicago and senior author on the study, the technique could help researchers better prioritize which of the vast number of protein PTMs to investigate more thoroughly.
Protein PTMs like phosphorylation are key to cellular function, contributing to processes like protein signaling and localization. However, Moellering noted, given the tens of thousands of protein modification sites, it is impossible to follow up with in-depth functional studies of all or most modifications.
For instance, Moellering said, his lab studies protein phosphorylation, "and the ability to detect them [using mass spec] is great. But it gives us a list of hundreds or thousands or maybe tens of thousands of sites… and there's really no way to prioritize or predict [which might be functional]."
"I think it's widely accepted now that a large percentage of the modifications observed throughout the proteome… are present at very low levels and they may just be kind of background modifications that may not serve a given function," he said.
He added that this may be true even in the case of a well-studied PTM like phosphorylation, which is key to protein signaling.
"The enzymes aren't going to be perfect," he said. "They recognize relatively small motifs, and there may be sites where [modifications] are coming on and off but may not be having a big impact on the protein itself."
To get at this question of functionality, he and his colleagues looked to thermal profiling, the basic notion being that bound ligands are known to change the thermal stability of proteins and PTMs are essentially just small ligands.
"So we hypothesized that PTMs that had a differentially high effect on the stability of a protein might be functionally important," Moellering said.
To generate thermal profiles of protein phosphorylation, the researchers heated samples of live HEK293T cells to temperatures ranging from 37° C to 67° C. They then lysed the cells and centrifugated the samples to remove proteins that had been denatured by heating. They followed that with protein digestion, phosphopeptide enrichment, TMT labeling, and mass spec analysis.
TMT labeling allowed for multiplexed quantitation of samples across the temperature range, with less heat-stable molecules present in lower quantities as the temperature applied increased. Thermal profiles were generated for phosphopeptides and for corresponding unmodified proteins by observing these temperature-linked quantity changes. A difference in thermal profile between a phosphopeptide and its corresponding unmodified protein indicated that the modification changed the molecule's thermal stability and might therefore be functional.
Using the approach, Moellering's team was able to compare the thermal profiles of 2,883 phosphopeptides and unmodified proteins, finding in 719 of these comparisons a significant difference between the two.
Among the phosphosites exhibiting an influence on protein thermal profiles were well-characterized functional modifications on proteins including CDK1, STAT3, and STAT1. The technique also identified a known interaction between the protein eukaryotic translation initiation factor 4E binding protein 1 (4EBP1) and eukaryotic translation initiation factor 4E (EIF4E) based on the lower stability of phosphorylated 4EBP1, which prevents the two molecules from binding.
This latter observation points to a future direction for the work, Moellering said, noting that while he and his colleagues were initially focused on the idea that a PTM would change a protein's thermal profile by changing its structure, "as we start digging through the data, we think that a lot of the shifts we are seeing are resulting from altered intermolecular interactions, changes in protein-protein interactions, for instance, or protein-metabolite binding."
This suggests the possibility of combining the approach with other forms of omics data, he said. "You could imagine simultaneously tracking the metabolome and the modified proteome as a way of elucidating larger-order interaction networks."
It might be similarly applied to looking at the impact of drug treatments or other cellular perturbations, he said.
Mass spec-based thermal profiling has previously been used for drug research including proteome-wide drug screening, wherein researchers looked for changes in protein stability to identify proteins binding to an agent of interest.
Moellering said the method could also be useful for exploring the functional impact of genetic mutations, allowing researchers to look at how changes in peptide sequence, as opposed to PTMs, impact a protein's thermal stability.
The researchers used MS2 level TMT labeling in the study, which faces challenges of precursor interference that limits the method's quantitative accuracy. Moellering said that using newly developed MS3 TMT methods would likely allow researchers to further improve their proteome coverage.
Moellering said that his lab is using the technique to study the ability of metabolic networks to impact protein interaction networks.
"We are using this pretty extensively to look at how the dysregulated metabolic conditions found in cancer and inflammatory diseases alter the kind of subcomplexes and spatial distribution of metabolic proteins in the cell," he said.
He added that his team is also using the technique to look at modifications beyond phosphorylation.
"We started with phosphorylation because it's the most abundant and well characterized modification, but we have used it and are using it for other modification sites," he said.