Skip to main content
Premium Trial:

Request an Annual Quote

PROPER-Seq Uses NGS to ID Undiscovered Protein-Protein Interactions


NEW YORK – A new method for profiling protein-protein interactions (PPIs) using next-generation sequencing as a readout has the potential to reveal thousands of previously undiscovered combinations in human cells.

The method, called PROPER-seq (protein-protein interaction sequencing), tags proteins with RNA barcodes that form chimeric sequences when brought into proximity with each other.

"PROPER-seq scans approximately 15,000 times 15,000 protein pairs," said Sheng Zhong, a professor at the University of California, San Diego whose lab developed the method.

In a study of  human embryonic kidney cells, T lymphocytes, and endothelial cells researchers led by Zhong found more than 200,000 new PPIs. Among these, 1,365 were supported by published co-immunoprecipitation studies and 2,480 were supported by affinity purification mass spectrometry data. Another 17,638 had been computationally predicted but not experimentally validated. And about 100 overlap known human synthetic lethal gene pairs. The researchers published their results last month in Molecular Cell.

The method draws on a decade of experience probing molecular interactions in cells with DNA barcodes. "The idea is simple, but systematically labeling tens of thousands of proteins at once turns out to have some challenges," Zhong said. A key is to validate that the protein display step that forms conjugated mRNA-protein chimera is successful for nearly all the genes. 

Already, he is seeing interest for PROPER-seq both from other researchers as well as at least one company looking to license the technology for a research-use assay kit.

The path to PROPER-seq began about eight years ago, Zhong said, with the idea that PPIs could be analyzed in a similar way to what his lab was doing with RNA-RNA and RNA-DNA interactions.

A technique called SMART-display, similar to mRNA display methods developed by Harvard University Professor and Nobel Laureate Jack Szostak, results in cells marked by mRNA-protein fusions. Then, to capture the interactions of those proteins, the method applies chemistry that converts the exposed mRNA to cDNA, uses proximity ligation to link barcodes of interacting proteins, and then collects the chimeric DNA for sequencing.

The result is a so-called "many-to-many" approach to analyzing PPIs, as opposed to "one-to-one" or "one-to-many" approaches. Proper-seq also joins other NGS-based methods, such as synthetic agglutination (SynAg) of yeast. That method, described in a 2017 study, is being commercialized by A-Alpha Bio.

The potential to discover large numbers of previously unknown interactions is a major draw, according to Tao Zhang, a postdoc in Jazz Dickinson's plant biology lab at UCSD.

"The really interesting thing for me is [PROPER-seq] can reveal many unknown PPIs," he said. "Many unknown mechanisms under many different biological processes would be found based on this technology." He has used yeast two-hybrid assays for PPI studies before, but "it takes a long time to set up, and much labor."

Costs could also be lower, compared to other methods, such as mass spec-based ones. "PROPER-seq's cost is at a similar scale of RNA-seq experiments," Zhong said. He estimated that sample prep costs for an experiment, which would not include the cost for the cell lines or sequencing, were approximately $2,000. Antibody-based methods for PPI studies cost several hundreds of dollars for each antibody used, he said, making them cost-prohibitive for a similar scale, while a synthesized gene library for all human genes could cost on the order of $100,000, he said.

He recommended 250 million to 500 million read pairs per experiment for sensitive detection. "However, the strongest interactions can be detected at a depth of  about 30 million read pairs," he said, adding that sequencing 800 million paired-end reads costs approximately $3,300. 

The method does not test the affinity of protein pairs, Zhong noted. There is also a threshold for the overall number of chimeric read pairs one can set that influences the final number of associations. Setting a higher threshold results in a smaller number of PPIs reported, but results in better reproducibility. "The more stringent the threshold, the more overlap there is" between studies, he said.

The paper also did not present head-to-head comparisons with other methods, including on precision and recall. But PROPER-seq's ability to identify biologically relevant interactions is evident by its overlaps with computationally predicted PPIs and synthetic lethal gene pairs," Zhong suggested.

Plant studies are just one of the possible applications of this method. "Protein-protein interactions are important for every biological process," Zhong said. "People can use this in their cell type or tissue of interest."