In a study published in this month's edition of Molecular Systems Biology, researchers at Harvard Medical School and the University of Lille compared the use of Western blotting, top-down and bottom-up mass spectrometry, and nuclear magnetic resonance spectroscopy to quantitate the distribution of phospho-forms of the protein kinase Erk in in vitro samples.
They identified two new Erk phosphorylation sites via NMR analysis and discovered discrepancies in their Western blotting data that suggest that some antibodies commonly used in kinase research may not be reliable for quantitative studies, said Jeremy Gunawardena, associate professor of systems biology at HMS and lead author on the paper.
"One of the main ideas we were trying to get across is the importance of what we call the phosphor-form distribution, the relative proportion of each of the global patterns of [phosphorylation] that might be present," he told ProteoMonitor. “It's not just looking at whether a particular site is phosphorylated. You really have to worry about the global pattern of phosphorylation."
As Gunawardena noted, there's a growing appreciation among proteomics researchers that an understanding of protein variants and isoforms is key to basic biological and clinical research. In areas like epigenetics and biomarker and cell-signaling work, scientists are seeking to better understand and quantify global patterns of protein modification.
"The pattern of phosphorylation on a region of a protein contains additional information, and in the case of these important kinases [like Erk], that extra complexity could be a new area to expand [research] on," said Thomas Graeber, professor of molecular and medical pharmacology at the University of California, Los Angeles, whose lab focuses on cancer signaling and protein phosphorylation. "Maybe the reason we don't yet know what percentage of [signaling events] this more complex control is important in is just because we haven't had the tools to measure it."
With that in mind, Gunawardena's team compared four different methods for quantitating Erk phospho-forms: Western blotting on a Li-Cor Biosciences imaging platform with phospho-specific antibodies for each Erk phosphor-form; peptide-based mass spec on a Thermo Fisher Scientific LTQ XL-Orbitrap instrument; intact protein mass spec on an Applied Biosystems QSTAR XL; and NMR on a Bruker 600 MHz Avance II spectrometer.
Top-down and bottom-up mass spec were quantitatively consistent with each other and to NMR within 10 percent. Western blotting, however, produced significantly different results, suggesting, Gunawardena said, that the antibodies typically used for detecting Erk phospho-forms might not be suitable for determining phospho-form distribution.
"We wouldn't have been surprised if we got large quantitative discrepancies between the antibodies and mass spec," he said, but, in fact, Western blotting actually gave the opposite answer from mass spec in some cases.
"Mass spec would say there was proportionally more of [a given Erk phosphor-form] in sample one compared to sample two, and when we did the same measurement with antibodies there were conditions where the antibodies would give us the reverse message," Gunawardena said.
He speculated that this might be caused by interference from other phospho-forms present in the mixture, noting that, while antibodies have excellent sensitivity for detecting the presence of various phosphor-forms, "our experience has made us very cautious about any quantitative numbers you get using them. One has to treat them with great caution."
Graeber noted, however, that while such potential antibody issues "may be very much under appreciated by the general community," it's possible that the problem that arose in Gunawardena's in vitro work might not arise when using actual biological samples.
"It may be that the form [causing the interference] is so rare that it's not a problem, or it could be that in a biological sample the amount of that form correlates well with the forms [the antibody is meant to target] so it wouldn't mess up your correlations," he said. "Those are just possibilities, of course, but it doesn't mean that the antibody becomes less useful. It just requires more care."
Ultimately, the researchers determined that a combination of peptide-based mass spec and intact protein mass spec yielded the most accurate quantification of Erk phospho-form distributions, with the top-down approach revealing the distribution of the isobaric groups and the bottom-up method offering more detailed information on the phosphorylation state of particular sites.
Even using this combination, however, distinguishing between positional isomers remains a challenge, the authors noted, adding that they're now trying new separation strategies including UPLC and ion mobility to better separate these proteins.
The results from NMR tracked closely with those obtained using mass spec, Gunawardena said. Additionally, because, unlike bottom-up mass spec, NMR is unbiased in that it analyzes all the peptides in a sample, the technique enabled the researchers to identify two additional Erk phosphorylation sites not picked up via the other methods.
NMR has long been used in structural analysis of proteins but is rare in proteomics work. It has several qualities that could make it useful, however, Gunawardena noted.
In particular, NMR allows researchers to monitor the biochemical reactions in a sample over a period of time. "You can watch to see how [post-translational] modifications dynamically change," he said. "That's something that we're doing and it's much better than with mass spec where you'd have to take samples at different times points."
Another advantage, he added, is that NMR can be used on intact cells, potentially allowing for "in vivo measurements on intact cells."
Guy Lippens, a University of Lille researcher and co-author on the paper, has been using NMR for several years with his colleague Isabelle Landrieu to study the phosphorylation of the tau protein, a potential biomarker for Alzheimer's disease.
"We'd really like to understand the phospho-code of tau in terms of physiology and pathology," he told ProteoMonitor. "What is the influence of different phosphorylation sites on its function in microtubule assembly? What is the function of phosphorylation and the assembly of the pathological fibers” linked to Alzheimer's?
He also plans to use the technique to investigate the phosphorylation patterns of proteins tied to hepatitis C, as well as a number of other post-translational modifications. Such NMR analysis is "not limited to just phosphorylation," he said. "Acetylation, methylation, glycosylation – I think it's a general method."
There are drawbacks that limit the method's usefulness, Lippens noted, including sensitivity that tops out in the micromolar range, large sample needs, long run times, and the requirement that proteins being analyzed are isotope-labeled.
There's also the fact, Gunawardena said, that most proteomics researchers are unfamiliar with NMR and therefore need to partner with an expert in the technique if they want to use it.
"Mass spectrometry has gotten to the point where it's almost just a service," he said. "You have mass spectrometry core facilities. You take your sample there, someone comes back and tells you what you're proteins are. Here, though, you have to build a relationship with some NMR researcher, [and] get them interested in your post-translational modification."
Use of NMR by proteomics researchers will likely "be limited in the next few years to sort of fairly targeted studies of proteins of interest," Gunawardena said. "But I think if that begins to show some interesting insights, it could be a spur to improving" some of the difficulties surrounding the technique.
Moving forward, the researchers plan to investigate the phosphor-form distribution of receptor tyrosine kinases like the epidermal growth factor receptors, he said, noting that the study helped establish the usefulness of a "hybrid top-down, bottom-up mass spec strategy" for doing so.
The study "has given us considerable confidence that [such a method] is accurately reporting what's there," he said. "Now we want to probe a number of different signaling pathways and proteins that are multiply phosphorylated."
Have topics you'd like to see covered in ProteoMonitor? Contact the editor at abonislawski [at] genomeweb [.] com.