A novel technique maps antibodies’ epitope specificities

According to one count, some 5 million commercial antibodies exist, used to target, detect or destroy in contexts from research labs to patients' bodies. Raised in goat, rat, mouse, rabbit, llama and even chicken against fragments of a menagerie of proteomes, these antibodies bind with high affinity to their mostly protein targets.

Eucalyp/The Noun Project

Determining the portion of a protein surface to which an antibody binds is called epitope mapping. The gold standard approach is to determine the structure of an antibody-target complex. However, structural work can be difficult and may take months per antibody. Researchers also use peptide display technologies, presenting a library of protein fragments on a microarray or the surface of E. coli and finding out which ones the antibody binds to through microarray scanning or flow cytometry.

In a recent article in the journal Molecular & Cellular Proteomics, researchers in the labs of Xiaodong Zhao, Hua Li and Sheng-ce Tao at Shanghai Jiao Tong University introduced a new, faster method for epitope mapping. Using the new technique, they write, "One technician is able to map the linear epitopes of more than 200 antibodies in one month at an affordable cost."

The approach starts with a random phage display library, a collection of 10⁹ peptides with random sequences, each 12 amino acids in length, expressed on the surfaces of phages. (Although other researchers previously have used phage display for epitope mapping, they mostly have presented variants on a known target protein rather than random peptides.)

"The beauty of phage display is the protein sequence and the coding sequence are linked together," Tao said. Therefore, after biochemically extracting phages that bind to an antibody of interest, the team quickly could identify these phages through next-generation sequencing. By comparing all of the phage peptides that bound to each antibody, they detected common protein motifs. This let them establish a consensus binding sequence for each antibody — without any preconceptions about what it would be.

Sometimes the epitope they determined matched the protein or peptide against which the antibody was raised. In other cases, the consensus epitope sequence was not found in the protein an antibody was raised against. According to Tao, this makes sense when you think about antibodies and their targets binding in three dimensions.

"How antibodies are generated is not always based on the linear sequence," Tao said. "Sometimes it's based on conformational structure."

Because of the nooks and crannies folded into any protein's structure, amino acids that are not adjacent in its linear structure often are found next to each other on its surface. While the technique is quite useful for finding linear epitopes, the authors said, it cannot detect conformational epitopes; they hope to find new computational approaches that will bridge the gap.

Still, the research has clear scientific and commercial potential. Since the paper was published, Tao said, the team has improved the speed and efficiency of their technique. The lab has struck up a partnership with a reagent supply company to map the epitopes of 10,000 of its antibodies and anticipates using the resulting data set to learn more about the relationships among antibody, antigen and epitope. They're also considering other applications, such as identifying potential off-target binders that could cause unexpected side effects in antibody-based drugs.