NEW YORK (GenomeWeb) – A team led by researchers at the Swiss Federal Institute of Technology Zurich has developed a selected-reaction monitoring mass spec assay for measuring a panel of "sentinel" proteins – proteins that can serve as markers of the activation state of a particular biological process.
Detailed in a study published last week in Nature Methods, the assay, which was developed in yeast, could help researchers study in high throughput the effect of a wide variety of perturbations across a broad range of biological pathways, said Paola Picotti, an ETH Zurich researcher and senior author on the paper.
While the desirability of such an assay might seem fairly obvious, and the SRM mass spec technology used to build it is widespread within proteomics, to Picotti's knowledge no such assay previously existed, she told ProteoMonitor, adding that she has "been wondering why we haven't been scooped" by another lab.
One likely reason, she suggested, was the amount of work required to put together the list of sentinel proteins.
"You need biological knowledge to do something like this because you need to put together a list of markers that are very well characterized by experimentation," she said. "And so you need to put together a mass spec person and a biologist to decide on which are the best markers."
Indeed, this was the bottleneck for her team's efforts, Picotti said, noting that identifying the best sentinel protein candidates took them more than two years.
Ultimately, she and her colleagues identified 570 potential sentinels based on biological data gathered through literature searches, antibody databases, protein databases, and discussions with biologists. They also identified a set of 30 additional sentinel candidates through a computational prediction method they developed. The method evaluated the likelihood of a protein being a sentinel based on characteristics including its degree of functional characterization and its specificity for particular pathways.
They then developed SRM assays to 157 sentinel proteins and 152 sentinel phosphopeptides, running these in two 30-minute mass spec runs. In total, the hour-long assay provided information on the activity of 188 different biological processes in yeast.
Applying the assay to yeast cells subjected to eight different perturbations — osmotic stress, osmotic stress adaptation, rapamycin treatment, amino acid and nitrogen starvation, entrance into stationary phase, 30 and 60 minutes of heat shock, and recovery from heat shock — the researchers managed to quantify 202 sentinels, letting them monitor a wide range of pathways across the eight different conditions.
In addition to SRM, the researchers used the Swath data independent acquisition approach to collect quantitative data on the target proteins.
Less sensitive than traditional SRM, Swath was able to measure only 66 percent of the sentinel proteins covered by the SRM assay. However, Picotti noted, because Swath collects quantitative information on all proteins it detects in a sample – as opposed to SRM, which measures only targeted proteins – the technique could prove useful for re-analysis of pathway data.
For instance, she said, "if you see [using SRM] that a given pathway is activated under a given condition, you would then have to do another experiment to look deeper into that pathway."
On the other hand, because Swath collects data on everything it sees, "if you have the Swath data, you could mine that data again and look at the behavior of more components," she said.
The researchers performed their Swath analysis on an AB Sciex TripleTOF 5600 instrument. AB Sciex recently released an updated version of Swath for use on its new TripleTOF 6600 mass spec, which, Picotti said, would likely provide better coverage than she and her team were able to achieve using the original Swath method.
Of the 202 sentinel proteins the researchers quantified across the eight conditions, 74 percent of them exhibited the expected response. Interestingly, the computationally predicted sentinels performed essentially the same, with 73 percent matching the expected response.
Regarding those proteins that didn't behave as expected, Picotti said a variety of factors could underlie the results.
"One problem we might be having is pathway crosstalk – that one sentinel is a marker not only for the activity of one pathway but responds also to another pathway and that this was simply not known based on the previous literature," she said.
Another possibility might simply be technical problems, she noted. "Maybe the coordinates of the SRM assay were not as specific as they needed to be."
The best way to study these outlier proteins is to expand the assay to include multiple proteins per pathway, Picotti said, noting that the researchers are currently in the process of such an effort.
"If you have three sentinels [for a pathway] and two agree and one doesn't, then at least you know that something is wrong there," she said. "So we are trying to increase the number of sentinels right now."
They are also working to transfer the yeast assays to human studies.
"Many of the pathways we captured with the sentinels are conserved [between yeast and human], and even some of the sentinels are conserved," she said. "So in principle, the homologs of some of the sentinels we used for yeast could be moved to mammalian systems."
One challenge to this effort, though, is assay sensitivity, Picotti said, noting that the greater complexity of the human proteome will mean that some assays that worked in yeast will probably not be sensitive enough for work in humans.
Beyond basic development of the assay, the ETH Zurich team is also applying it to various research interests – in particular, Picotti said, investigations of how cells respond to intracellular protein aggregation.
"We are using it to study the mechanism of action of a set of known modulators of protein aggregation toxicity," she said.