Protein post-translational modifications proved a primary focus at the 2012 Association of Biomolecular Resource Facilities conference held this week in Orlando, as the organization's Proteomics Standards Research Group – sPRG – and Proteomics Informatics Research Group – iPRG –both presented results from studies assessing labs' abilities to identify PTMs in complex mixtures.
The study results suggest that proteomics researchers have made considerable strides in their PTM analysis capabilities, particularly with regard to singly phosphorylated peptides. Significant challenges remain, however, as participants in both studies struggled to identify several categories of modification – sulfation and multiple phosphorylation, most notably.
For its study, the sPRG generated a standard containing 70 synthetic peptides representing modifications including acetylation, methylation, nitration, phosphorylation, and sulfation, which participants evaluated by a variety of mass spec methods.
According to Alexander Ivanov, director of the Proteomics Resource at the Harvard School of Public Health and chair of the sPRG study, the results demonstrated both the strides that the proteomics community and mass spec instrumentation have made in recent years and the room for improvement that still exists with regard to PTM analysis.
"Over the past few years, the field has improved and new [mass spec] instruments have come out on the market … [that] have allowed us to improve our sensitivity and accuracy in the detection of phosphorylated peptides," he said. "Also, we found that we could reliably look at different modifications [including] acetylation and methylation."
However, "we still suffer in finding multi-phosphorylated peptides," Ivanov said, noting that, in particular, "there is still room to improve the methods for enrichment and separation in the analysis of multi-phosphorylated peptides."
The 30 datasets returned by sPRG survey participants identified roughly 70 percent to 80 percent of the monophosphorylations, acetylations, and methylations in the sample and more than half of the PTMs present in total.
Participants fared more poorly with sulfation and multiple phosphorylation, achieving for tetraphosphorylated peptides a detection rate of just 5 percent. As Ivanov noted, though, the trendlines are encouraging. In 2003, study participants identified a pair of monophosphorylated peptides at a rate of 15 percent. By 2010 this had increased to more than 40 percent, and in this year's study, participants identified monophosphorylated peptides at a rate of between 80 and 90 percent.
The iPRG study, which tested researchers' abilities to identify PTMs within mass spec datasets, used the same peptide mix as the sPRG study but spiked it into a yeast lysate background for an additional challenge. The organizers generated the study dataset on an AB Sciex TripleTOF 5600 instrument and then distributed it to participants.
As in the sPRG study, participants struggled most with identifying sulfated and multi-phosphorylated peptides, said Robert Chalkley, associate adjunct professor at the University of California, San Francisco, and chair of the study.
"The first thing that stands out from the data is that people didn't do very well with sulfation," he said, noting that researchers often reported these modifications as phosphorylations due to their similar mass.
"There were a handful of peptides that were not identified" by any of the study's 24 participants, Chalkley said, including four tetra-phosphorylated peptides, one di-methylated peptide, and one sulfated peptide.
Overall, participants identified roughly two-thirds of the modified peptides present in the data, Chalkley noted. They were also, on average, able to correctly make site assignments for more than half of these modifications, with several making correct site assignments for almost 60 of the 70 peptides, a result he said was "impressive considering how complex the mixture was."
With regard to site assignments, mistaking sulfation for phosphorylation was again a primary issue. "The most common errors [in making site assignments] were reporting sulfo-tyrosine as phosphorylated serine/threonine" followed by "missed assignment sites for multiply-phosphorylated peptides," Chalkley said.
ABRF's Proteomics Research Group, PRG, also presented at the meeting. However, as the group is currently in the second year of a two-year longitudinal study assessing intra-lab variability in peptide LC-MS/MS analysis, it had only preliminary data to share.
The goal of the PRG effort, said Brett Phinney, manager of the Proteomics Core at the University of California, Davis, and a member of the study committee, is to investigate "if you give a sample to a core facility [for analysis] in June, will you get the same result in July? We wanted to know what the variability would be with samples run over a period of time."
To approach this question, the study leaders have sent participants purified, digested aliquots of a mix of six bovine proteins from Bruker-Michrom, asking them to run one sample per month over nine months using the same mass spec setup each time. Participants will collect data covering 45 metrics for each run, allowing for an assessment of variability within their labs over the nine months.
As an example, Phinney offered preliminary data on changes in IQ RT period – essentially the retention time period between the first eluting peptide and last elution peptide in an LC-MS run – observed by two participants in the study.
"Basically, that's your separation space for your LC-MS run … so, for example, if three months from now your retention time period is increasing, that might tell you something about [the stability] of your LC-MS system," he said.
Phinney also presented data on the types of mass spec instruments being used in the study, offering a glimpse into the relative popularity of various vendors among the participating researchers. Thermo Fisher Scientific mass specs were by far the most common, with 62 instruments – most of them Orbitraps – represented in the study. AB Sciex machines were the next most common, with 19, followed by Waters with 17, Bruker with 6, Agilent with 3, and Shimadzu with 1.
"The unique feature of the study is that we are allowing labs to determine the settings based on their own preferences and instrument type," North Carolina Central University researcher Maureen Bunger, the group chair, told ProteoMonitor in an email this week. "This will allow labs to see the variability in their own routine methods and how this may compare among the vastly different settings and instrument configurations across the study participants."
The monthly mass spec and QC measurements for the PRG study will be taken between March and November, and the results will be presented at the ABRF annual meeting next year.
Have topics you'd like to see covered in ProteoMonitor? Contact the editor at abonislawski [at] genomeweb [.] com.