SALT LAKE CITY — The Proteomics Standards Research group of the Association of Biomolecular Resource Facilities has begun a two-year project to study the human plasma proteome and eventually develop a standard.
The group, which unveiled its plans during a session at ABRF’s annual conference held here this week, said the goal of the study is to develop a standard, such as a peptide sequence, that could be applied to the entire human population, or a large part of it.
Every year, the various ABRF research groups conduct studies to gauge how researchers do their work, gather information about problems they may encounter when carrying out their research, and provide suggestions for doing their research more effectively.
Other proteomic studies presented this week outlined by research groups looked at methods for determining qualitative differences in proteins; assessed the quality of protein bioinformatics; and assessed methods of identifying recombinant proteins in E. coli.
Last year, the group, known as sPRG, developed phosphoprotein standards and evaluated how core facilities identify proteins and their phosphorylation sites in a simple sample [See PM 04/05/07].
The plasma proteome study, which is set to be announced this summer, is still in development, so sPRG members declined to speak on the record about it.
Complex, Yet Ubiquitous
Human plasma is a notoriously complex fluid with thousands of proteins and a dynamic range of at least 10 orders of magnitude. Despite this, it is a body fluid of choice for many proteomics experiments because it is readily obtainable and “well preserved for prolonged periods of storage,” the sPRG said in a poster presented during the conference.
It reflects “a unique signature of the physiological state of all tissues” of a person and is considered a “source of protein biomarkers comprehensibly sampling the phenotype of the human body.”
While various mass spec-based approaches have been developed to analyze plasma, efforts to create a successful proteomic assay have come up short, sPRG said in its poster.
“Availability of standards with which to quantify plasma proteins, spanning a broad dynamic range, would facilitate methods development, optimization, and validation for reliable plasma analysis,” sPRG said.
For now, the group said it plans to base its study on depleted plasma spiked with stable-isotope-labeled peptides of up to 50 human proteins covering five orders of magnitude.
The group used Sigma-Aldrich’s ProteoPrep20 columns and Aqua peptides to prepare its proteins, which were chosen based on criteria such as peptide stability and synthesis affinity. In the latter, for instance, the peptide length had to have at least eight but no more than 20 amino acid residues. The group also considered heterogeneity and detectability.
During the first year of the study, which began last year, the group sought to develop criteria for selecting candidate proteins and peptides; explore plasma preparation and methods of complete trypsin digestion; explore mass-spec platforms to use in the study; and assemble sequences to synthesize as stable-isotope labeled standards.
As it proceeds, sPRG will try to synthesize and quality-control peptide standards; create a “publicity” domain to recruit and collect feedback from participants; have group members test the standard in development using alternative platforms; develop guidelines for conducting the study; distribute the standard; and collect data from study participants and evaluate their ability to quantify plasma proteins, while analyzing the performance of the methods, instruments, and analyses.
At the ABRF session announcing the study, members of the audience appeared to appreciate the effort but expressed concern that it might be too ambitious. In particular, some said that having to carry out a study on 50 proteins could be too much of a task for most core labs.
One audience member said, “I’m not sure I’d have the stamina to do all 50 proteins.”
sPRG members, for their part, conceded that they will continue to develop the specifics of the study to ensure that participation will include novice researchers as well as expert ones. One committee member said that it may have been “too aggressive” in its reach and “maybe we need to take a step back.”
All the RAGE
Meanwhile, another research group, the Proteomics Research Group, was studying the methods used by researchers to determine qualitative differences in protein samples.
Nathan Yates, a researcher at Merck and the chairman of sPRG, said during his presentation that the group, PRG, also wanted to highlight the type of information obtained through different approaches and to share best practices with fellow ABRF members.
In the study, 57 participants were asked to analyze two samples of related human proteins expressed in E. coli and affinity-purified, labeled Sample A and Sample B. Specifically, the extracellular region of the receptor for advanced glycation end products, RAGE, “designated as sRAGE for soluble RAGE … and two truncated variants (VC1 and V, tandem and single domain fragments, respectively) were overexpressed in [E.coli] strain, Origami-B (DE3),” according to a poster describing the study.
Participants were asked to identify the major proteins in each sample, determine any differences in protein composition between the two samples sets, and report the sequence coverage obtained from the major proteins in each sample.
“E. coli is really the workhorse of recombinant protein production.”
During the design of the study, Yates said, there was lively debate within the group about how much information about the samples should be provided to participants.
“Part of the challenge … was to start off without a complete picture” of what they were analyzing, Yates said during his presentation.
Of the 57 participants, eight were able to report the correct termini for sRAGE, VC1, and V, “attesting to the feasibility of the study,” PRG said in the poster. “Numerous participants identified the full-length form of sRAGE present in the two samples, “but were not able to distinguish the truncated forms that were present in Sample B.”
Routine approaches such as 1D and 2D PAGE and MALDI-TOF MS were used by the majority of participants, and several groups used multiple enzymes to increase sequence coverage. None of the participants used electron-transfer and electron capture-dissociation methods for the study, however.
“We were hoping to see results for [both techniques] but didn’t,” Yates said.
Participants said that issues they encountered while carrying out the study included the presence of contaminants, such as keratin, and sample degradation. Also, a few submissions reported post-translational modifications not known to occur in E. coli expression systems.
PRG concluded that the results suggest that “definitive determination of protein sequence and identification of unexpected sites of truncation” can be achieved by mass spec-based methods.
“In many cases, the combination of two complementary approaches (e.g., 1D SDS PAGE to resolve protein components followed by LC-MS/MS for sequence information) gave a higher success rate than use of a single experimental approach,” according to the poster.
Assessing Protein IDs
Another ABRF research group, the Proteome Informatics Research group, or iPRG, presented preliminary results of a study seeking to assess the current state of the protein bioinformatics field, with a particular emphasis on the quality and consistency of protein identifications based on MS experiments.
There are a number of peptide identification and protein-inference software tools available, and users can select varying degrees of stringency for their protein lists, which has raised a number of questions around the quality and consistency of the reported results.
Sean Seymour, senior staff scientist at Applied Biosystems and outgoing chair of iPRG, said that the primary goal of the study was to determine the current state of the field with regard to reporting protein lists. A secondary goal, he said, was establishing a “benchmark” for assessing the quality of protein inference for a reference data set.
The study used a common data set — a mouse liver differential-expression study performed on a 3200 QTrap — and a single protein reference database that included 53,900 sequences with 74 decoy contaminants. The experimental data comprised 42,000 spectra. Seymour said the research group decided to provide a common MS analysis result rather than a sample in order to rule out any possible variations from the data-acquisition process.
After announcing the study in November, the group received 44 requests for the data set, but only 18 respondents submitted results, Seymour said. The iPRG members also submitted their own analysis, he said, bringing the total number of responses to 24. Nevertheless, he noted, the number of participants was not significant enough to draw any clear conclusions on the current state of proteome informatics.
As a result, he said that the group decided to keep the study open in hopes of getting more submissions. Those interested in participating can send a request for the data to [email protected].
Despite the limited response, Seymour said that the study did lead to some interesting findings. For one thing, he said that even when respondents used the same peptide-identification and protein-inference software, they still reported very different protein lists. Overall, he said, there was a “significant difference in the number of proteins” reported by the different groups, but he noted that the real significance of this is still unclear due to the low number of submissions.
One good sign was that there appeared to be very few protein-inference errors and “no gross inflation” of the number of proteins reported. However, he noted that this may not be representative of the broader proteomics community and conceded that the results may be slightly biased toward those members of the community who are very comfortable with protein-inference tools.
On the down side, Seymour said that three of the responding groups did not properly report ambiguity in accession numbers. He explained that because peptides could map to more than one protein, researchers should report all possible protein accession numbers for each peptide. However, he said that three respondents reported no accession ambiguity at all, which is problematic.
E. coli, My Old Friend
And finally, the Protein Expression Research Group, known as PERG, presented preliminary results from its ongoing study on recombinant expression in E. coli. The goal of the study, which will end March 31, is to determine the variability of results with expression and purification of a common His-tagged protein.
While acknowledging that the study may not have been the most original — a whole library of studies have looked at protein expression in E. coli — John Hawes, assistant professor of chemistry and biochemistry at the University of Miami, Ohio, said that it still bears further study because “E. coli is really the workhorse of recombinant protein production” and will continue to be for large-scale production of many proteins. “So there’s still a lot to be said for optimization and production.”
For the study, 22 participants were given samples of plasmid DNA; BSA standard; Nicotinamide adenine dinucleotide as a co-factor for enzyme assay; and a complete set of instructions and guidelines. They were then asked to transform the cell line of their choice; grow cultures and induce expression; purify the protein; use SDS-PAGE to analyze the protein assay and enzymatic activity; and perform an analysis by a platform of their choice.
Because only six of the 22 participants who received samples had returned their data, drawing any widespread conclusions would be premature, Hawes said. Nonetheless, the participants successfully obtained a recombinant protein, though the yield purity of the protein varied among different participants. In addition, all respondents detected protein activity, but specific activity varied “greatly” and was often low, Hawes said.
The methods chosen by the small pool varied, but similar patterns of contaminants were observed for different methods.
When the study ends PERG plans to submit its findings for publication and hopes to suggest which methods worked best in carrying out the study’s goals, Hawes said.
— Bernadette Toner contributed to this article.