Affinity capture of polyribosomes followed by RNAseq (ACAPseq), a discovery platform for protein-protein interactions.

Xi Peng,Philip M Smallwood,Francesco Emiliani,Mark F Sabbagh,Amir Rattner,Jeremy Nathans,Hong Lei

doi:10.7554/elife.40982

Abstract

Defining protein-protein interactions (PPIs) is central to the biological sciences. Here, we present a novel platform - Affinity Capture of Polyribosomes followed by RNA sequencing (ACAPseq) - for identifying PPIs. ACAPseq harnesses the power of massively parallel RNA sequencing (RNAseq) to quantify the enrichment of polyribosomes based on the affinity of their associated nascent polypeptides for an immobilized protein 'bait'. This method was developed and tested using neonatal mouse brain polyribosomes and a variety of extracellular domains as baits. Of 92 baits tested, 25 identified one or more binding partners that appear to be biologically relevant; additional candidate partners remain to be validated. ACAPseq can detect binding to targets that are present at less than 1 part in 100,000 in the starting polyribosome preparation. One of the observed PPIs was analyzed in detail, revealing the mode of homophilic binding for Protocadherin-9 (PCDH9), a non-clustered Protocadherin family member.

Highlights

Current platforms for large scale unbiased identification of protein-protein interactions (PPIs) utilize a variety of methods, each of which has strengths and limitations (Gonzalez, 2012; Legrain and Rain, 2014)
In developing ACAPseq, we focused on mammalian protein domains that reside in the extracellular space, because (1) many of these proteins have biologically important interactions, (2) many extracellular binding interactions involve domains that are likely to fold independently, (3) most single-pass transmembrane proteins, including many cell surface receptors, consist of an amino-terminal extracellular domain (ECD) and a carboxy-terminal intracellular domain (ICD), the optimal arrangement of the ECD for a method that relies on binding in the context of a nascent polypeptide, and (4) most extracellular domains contain one or more disulfide bonds and, in general, these domains do not fold properly in the reducing environment of the cytoplasm and are not amenable to yeast or mammalian two-hybrid approaches
Protein baits were produced as secreted human IgG1 Fc fusion proteins following transient transfection of mammalian (HEK/293T) cells, immobilized on the surface of Protein-G-coated magnetic beads, and used to capture polyribosomes prepared from neonatal mouse brain, a tissue that expresses a large and diverse set of genes (Figure 1A)

Summary

Introduction

Current platforms for large scale unbiased identification of PPIs utilize a variety of methods, each of which has strengths and limitations (Gonzalez, 2012; Legrain and Rain, 2014). Mass spectrometry of purified protein complexes provides an unbiased approach to identifying PPIs, but it is limited to relatively high affinity interactions as well as by sample quantity and purity (Gavin et al, 2011; Frei et al, 2012). Identifying PPIs based on a bioassay has high specificity and biological relevance but requires some knowledge of protein function (e.g. Lin et al, 2008; Zhang et al, 2014). Proximity labeling assays, such as BioID, tag nearby proteins but only a subset of these is likely to interact directly with the protein of interest (Fernandez-Suarez et al, 2008; Roux et al, 2012; Schopp et al, 2017)

Methods

Results

Conclusion