Protein-protein interactions in Escherichia coli: Using cross-linking/mass spectrometry (XLMS) to reveal our current...

Protein-protein interactions (PPIs) are fundamental to cellular processes and often define phenotypes more accurately than protein abundance alone. Despite their importance, confidently identifying direct physical interactions remains challenging. Even in the benchmark organism Escherichia coli, our survey of the IntAct database reveals that only 8% of reported interactions are annotated as direct physical associations. Of these, 44% rely on gold-standard structural methods, while XLMS accounts for only 11%, highlighting a lack of high-confidence, scalable data for primary interactors. In this study, we employed an XLMS workflow using the MS-cleavable cross-linker DSSO and SCX-based enrichment to investigate the E. coli interactome. This targeted approach yielded 21,599 cross-linked spectrum matches (CSMs), corresponding to 2334 residue pairs. These data mapped 663 intra-protein cross-linking and 137 PPIs, 47 of which were previously unreported. Notably, XLMS identified interactions with confidence scores below 0.7 in the STRING database, demonstrating its capability to detect low-confidence or uncharacterized associations. We further illustrate the power of this technique by analyzing the ElaB-YqjD complex, where our experimental distance constraints revealed a mismatch with AlphaFold 3 predictions. These results demonstrate how XLMS can effectively bridge the gap in current PPI datasets, providing high-confidence, mechanistically informative data even in well-studied biological systems. SIGNIFICANCE: Protein-protein interactions (PPIs) are essential to cellular organization and function, and even the extensively studied Escherichia coli, the majority of interactions remain either uncharacterized or poorly supported by high-confidence experimental data. As part of this study, we conducted a systematic analysis of the IntAct database, which showed that only 8% of annotated interactions correspond to direct physical associations, with 44% supported by gold-standard structural methods and only 11% derived from XL-MS. This analysis highlights a clear gap between interaction annotations and experimentally validated physical interactions, providing the rationale for the experimental strategy adopted here. To address this gap, we applied a state-of-the-art cross-linking mass spectrometry (XL-MS) approach using the cleavable cross-linker DSSO, coupled with SCX enrichment and advanced data analysis tools, to generate a high-confidence, experimentally derived E. coli interactome. This effort uncovered 47 novel PPIs and revealed substantial discrepancies between database annotations and physical interaction evidence. Our integrative workflow not only provides direct evidence of protein interactions within native cellular environments but also yields spatial constraints that challenge and refine current structural models, as illustrated by the elaB-yqjD complex. By bridging the gap between proteomic data and structural validation, this work demonstrates how XL-MS can meaningfully expand and validate biological interaction networks.