Researchers comprising the Renewable Protein Binder Working Group have published a study in this month's edition of Nature Methods detailing the production of more than 1,000 antibodies targeting 20 different SH2 domain proteins.
The publication describes a high-throughput workflow for generating protein affinity agents that is meant as a template for future, larger-scale efforts and could help improve and standardize antibody validation procedures, said Karen Colwill, a researcher at Toronto's Samuel Lunenfeld Research Institute and an author of the paper.
The effort grew out of a March 2008 workshop organized by members of the Structural Genomics Consortium, the Human Protein Atlas program, and the EU ProteomeBinders network program at which researchers agreed to launch a pilot study to explore methods of systematically generating renewable protein binders. Ultimately, scientists from more than a dozen laboratories around the world participated in either the antibody production or validation stages of the project.
One of the key drivers of the initiative was the SGC's ability to produce large amounts of high-quality antigens, Colwill told ProteoMonitor, noting that antigen production is typically "the first bottleneck" researchers run into in large-scale antibody work. According to the paper, using the methods described in the study, one researcher can produce milligram quantities of roughly 24 antigens per week, meaning, the authors said, that "in the near term access to antigen will not be limited."
The group chose as antigens 20 SH2 domain proteins, which are key components of cell signaling networks, binding tyrosine upon phosphorylation. These proteins were distributed to researchers in five laboratories who generated either monoclonal antibodies via hybridoma technology on a high-throughput robotics platform, or recombinant Fab or single-chain Fv binding agents using phage display.
Each laboratory performed preliminary ELISAs to screen for positive affinity agents for the 20 antigens and then passed on the top binders for further validation. Reagents were tested for their specificity via protein microarray assays done by the lab of Mathias Uhlen – head of Human Protein Atlas project – and for binding affinity via surface plasmon resonance done in the lab of University of Toronto researcher Sachdev Sidhu. The agents that passed these levels of validation then went on to immunoprecipitation, immunoblotting, and immunofluorescence assays involving endogenous proteins.
Of the three agents – monoclonals, recombinant Fabs, or recombinant single-chain Fvs – all emerged from the tests as equally effective options, Colwill said, noting that while for some antigens certain technologies worked much better than others, no broad trends emerged.
This result, she suggested, could encourage researchers to give the recombinant technology a closer look. Recombinant affinity agents "have a lot of potential, but they aren't as widely distributed commercially and aren't really used as extensively yet," she said. "I think it was good and of interest to people to see that in this side-by-side comparison both [hybridoma and recombinant] technologies are capable of producing antibodies that will recognize endogenous targets."
More important than the specific techniques used to generate the reagents, Colwill said, was the validation workflow the researchers applied on the backend.
"That's really the take-home of the paper – that these [affinity agents] can be generated on a large scale and that there should be some sort of systematic way that they are analyzed and validated," she noted.
In particular, Colwill said, performing biophysical validation steps to determine specificity and affinity prior to biologically validating the binders against endogenous proteins was key to streamlining the workflow.
"We looked at ways to arrive at a standard of validation," she said, "which means not just the biological endpoints, but to do work looking to see specificity among close family members and getting a feel for affinity, trying to figure out if there's a certain affinity cutoff [agents should meet]. Biological validation is going to be a bottleneck, so are there ways that we can quickly screen antibodies" for further validation?
"This is what we're really try to do: Set up a standard route for looking at how to best evaluate antibodies and then find what technologies seem to work best at it," she added. "As more and more data comes in, [the process] will get refined and we'll better understand what is most effective, but this is the first time it's been attempted to compare one common [validation] route for different technologies."
While there are currently more than 500,000 commercially available antibodies, the lack of well-characterized, high-quality affinity agents has long been an impediment to protein research. In addition to the efforts detailed in the Nature Methods paper, projects like Uhlen's Human Protein Atlas and the Human Proteome Organization's Proteomics Standards Initiative have sought to develop better standards for protein binding reagents.
In July 2010, PSI researchers published details in Nature Biotechnology on the Minimum Information About a Protein Affinity Reagent as part of the group's Minimum Information About a Proteomics Experiment guidelines (PM 07/16/2010).
As Sandra Orchard, senior scientific database curator at the European Bioinformatics Institute and one of the authors of MIAPAR module, told ProteoMonitor at the time, "people have wasted an awful lot of money over the years buying poor quality [affinity] reagents and coming up with misleading data because of it."
Colwill likewise noted that while companies each have their own standards for antibody validation, it can be difficult for researchers to know precisely what they are.
"Depending on where the commercial antibodies [come from] there are really different levels of validation that are done and different philosophies," she said. In some cases "it might just be 'We've done a Western and we see one band on the gel and it looks right so it's validated.'"
"This is just trying to put a little more” into the validation, Colwill said. "And when these [participating] labs go on to their next projects they're going to be looking at this as a foundation to build on."
Among the next projects the group plans to take on is the production of antibodies toward several epigenetic targets, she said.
Have topics you'd like to see covered in ProteoMonitor? Contact the editor at abonislawski [at] genomeweb [.] com.