A team of researchers led by scientists at the University of Zurich and the Swiss Federal Institute of Technology has developed a targeted proteomics approach for validating predicted microRNA targets based on single-reaction monitoring mass spectrometry.
The approach, which was detailed in a paper published in the September edition of Nature Methods, offers researchers a streamlined method for investigating the effects of miRNAs on protein expression.
It also demonstrates the potentially broad utility of SRM-MS databases like the Institute for Systems Biology's and ETH Zurich's SRMAtlas, Ruedi Aebersold, an ETH Zurich professor, co-developer of the SRMAtlas, and one of the paper's authors, told ProteoMonitor.
Traditionally, miRNA targets have been detected using transcript measurements like expression arrays. More recently, scientists have also become interested in measuring miRNA effects at the protein level, with researchers like David Bartel at the Massachusetts Institute of Technology's Whitehead Institute and Matthias Selbach at the Max Delbrück Centre in Berlin applying shotgun proteomics techniques to screen for miRNA targets.
Aebersold's team sought to apply proteomics in a more targeted way, using computational methods to first predict miRNA targets, and then developing SRM-MS assays to quantify each of the proteins predicted as a target.
"The issue [with a shotgun approach] is that you end up with a lot of proteins that change up or down and it may not have anything to do with [miRNA]. The depth of analysis is limited unless one expends an enormous amount of effort to sequence all of the proteins," he said. "What we attempted to do was focus the mass-spectrometric analysis only on proteins that are highly likely to be relevant."
The researchers identified these proteins by combining the results of five different miRNA target prediction algorithms; experimental data like microarray analysis and RNA interference screens; and literature searches. The final list comprised 861 candidate genes, from which they selected 181 proteins whose abundance in let-7(n2853) C. elegans mutants and wild-type C. elegans larvae they compared using SRM-MS. Of these 181 proteins, 29 showed significant differences in expression, marking them as let-7 miRNA targets.
Compared to computational approaches, which typically predict several hundred potential target mRNAs per miRNA, the relatively low number of proteins showing changes in expression "indicates that the prediction tools massively overpredict the number of targets," Aebersold said. "They're basically sequence alignment tools, which are, of course, quite robust, but not every sequence the [miRNA] aligns and binds to is an actual target."
"What we are proposing here is that by using the prediction tools, we can go very quickly to a hypothesis [regarding potential miRNA targets], and then hypothesis testing with a mass spectrometer, which can be very fast and very sensitive," he said.
The paper's larger point, Aebersold suggested, is to highlight SRM-MS as a technology useful for targeted experiments across a variety of research areas.
"This paper is really one in a whole series of this [SRM-MS] targeting technology that we are developing," he said. "It's one of the first significant applications of this technology. The whole idea is that we have tools now where we can very rapidly generate these targeted assays by selected-reaction monitoring. And so we can use biological knowledge or computational tools to generate hypotheses consisting of a few hundred proteins and then very quickly do the measurements with high precision and reliability."
While the Nature Methods paper focused on miRNA targets in C. elegans, resources like the SRMAtlas could make similar work even easier in humans, Aebersold noted.
The SRMAtlas, an initial draft of which was presented at last month's Human Proteome Organization's annual meeting in Sydney, is a comprehensive set of SRM-MS assays covering the human proteome. The map comprises more than 170,000 SRM assays – one each for at least five proteotypic peptides for each of the 20,300 human genes currently annotated as protein-encoding. The goal over the next year is to expand the number of available assays to between 250,000 and 300,000 (PM 09/24/2010).
The aim of the work is to build a set of standardized SRM assays that researchers can use to more easily investigate proteins of interest. Building SRM assays from scratch is a time-consuming process in which researchers must identify the peptides on which to base assays for a given protein; determine the best transition to explore using mass spec; and develop and optimize methods for performing the separation and the assay.
The ISB/ETH SRMAtlas essentially takes care of these steps for researchers, allowing them to detect and quantify proteins important to their research without going through the process of developing mass spec assays for each protein. The miRNA work, Aebersold said, is one example of the sort of research the database is intended to facilitate.
"This is exactly the idea of the [SRMAtlas]," he said. "It enables and supports targeted mass spectrometry experiments in any field of research where proteins are relevant. It could be clinical, it could be systems biology, whatever the case may be."
The miRNA study is "quite typical" of the projects for which the SRMAtlas should be useful, Aebersold noted. "There are tools that make predictions; those predictions are known to be not particularly good; and so one would like to test those predictions and for that this sort of targeting method is a powerful approach. So it's a nice example."
Given that miRNA researchers are typically "very well versed in nucleic acid measurements but not so much in proteins," scientists interested in adopting the technique will likely have to either "learn [SRM-MS] or hook up with someone who knows how to do it," Aebersold said. He added, though, that he expects tools like the SRMAtlas to increase use of SRM-MS across a variety of fields.
"We really hope that the availability of these assays will facilitate the dissemination of the use of proteomic technology in many different laboratories," he said. "The mass spectrometry used [in SRM-MS] is quite simple, and the whole method setup is rather simple once the assays have been developed. So it's much more of a turnkey system than a high-end discovery or shotgun [mass spec] instrument."