J. Am. Chem. Soc. 136, 11860–11863 (2014).

Protein–protein interactions regulate many important cellular processes. However, their often transient nature means that identifying which proteins interact with which partners is far from trivial. One approach to identifying protein partners is to start with a known protein as a 'bait' protein, which is then used to covalently label its partner protein(s) — the sought-after, so-called 'prey' proteins. Expression of a bait protein within a cell enables it to form biologically relevant protein–protein interactions with its appropriate protein partners — which can then be labelled. Subsequent enrichment allows the prey proteins to be identified using techniques such as protein mass spectrometry. However, inefficient labelling can lead to important interactions being missed, and the labelling of protein pairs that interact non-specifically can lead to false partners being identified.

Credit: © 2014 ACS

A team led by Peng Chen at Peking University, China, have now developed a tool for the crucial labelling step based on an unnatural amino acid. This amino acid, which can be genetically encoded into the bait protein, contains a selenium atom at the γ position and a diazirine group at the end of the side-chain. After the protein–protein interaction forms, the diazirine group can be irradiated with light, which induces crosslinking between the proteins, irrespective of their identity. Subsequently, the unnatural amino acid can be cleaved next to the selenium atom using H2O2, which separates the bait and prey proteins, but cleverly the prey protein is left with a selenic acid label. This can then be tagged with a fluorescent dye or with biotin, enabling enrichment and subsequent identification. Splitting the prey proteins from the bait simplifies the separation of the mixture of prey proteins, and also their subsequent identification.

To demonstrate the reliability and effectiveness of their approach the team profiled the binding partners of an Escherichia coli acid chaperone HdeA, under conditions of acid stress. Their analysis identified a number of previously known binding partners — showing that this method can reliably label prey proteins — and three new potential binding partners.