Comprehensive Interactome Mapping of the DNA Repair Scaffold SLX4 Using Proximity Labeling and Affinity Purification

Abstract

The DNA repair scaffold SLX4 has pivotal roles in cellular processes that maintain genome stability, most notably homologous recombination. Germline mutations in SLX4 are associated with Fanconi anemia, a disease characterized by chromosome instability and cancer susceptibility. The role of mammalian SLX4 in homologous recombination depends critically on binding and activating structure-selective endonucleases, namely SLX1, MUS81-EME1, and XPF-ERCC1. Increasing evidence indicates that cells rely on distinct SLX4-dependent complexes to remove DNA lesions in specific regions of the genome. Despite our understanding of SLX4 as a scaffold for DNA repair proteins, a detailed repertoire of SLX4 interactors has never been reported. Here, we provide a comprehensive map of the human SLX4 interactome using proximity-dependent biotin identification (BioID) and affinity purification coupled to mass spectrometry (AP-MS). We identified 221 unique high-confidence interactors, of which the vast majority represent novel SLX4-binding proteins. Network analysis of these hits revealed pathways with known involvement of SLX4, such as DNA repair, and several emerging pathways of interest, including RNA metabolism and chromatin remodeling. In summary, the comprehensive SLX4 interactome we report here provides a deeper understanding of how SLX4 functions in DNA repair while revealing new cellular processes that may involve SLX4.

Keywords: AP-MS; BioID; DNA repair; SLX4; genome stability; proteomics.

Figure 1

Nuclear localization and functionality of BioID2-FLAG-SLX4 and SLX4-FLAG-BioID2 in human cells. (A) Domain architecture of the human SLX4 scaffold highlighting known interaction partners and their binding regions. Abbreviations for protein domains: UBZ4, ubiquitin-binding zinc finger type 4; MLR, MUS312-MEI9 interaction-like region; BTB, Broad-Complex, Tramtrack and Bric a Brac; TBM, TRF2-binding motif; SIM, SUMO-interacting motif; SAP, SAF-A/B, Acinus and PIAS; CCD, conserved C-terminal domain. (B–E) Confocal immunofluorescence microscopy images of untransfected Flp-In T-Rex HEK293 cells (B), or cells stably expressing BioID2-FLAG-eGFP-NLS (C), BioID2-FLAG-SLX4 (D), or SLX4-FLAG-BioID2 (E). Cultures were left untreated (top panels) or incubated with tetracycline for 25 h to induce construct expression (middle panels) and biotin for 8 h to induce the biotinylation of proximal proteins (bottom panels). Cells were fixed and processed for immunofluorescence using anti-FLAG (red) and streptavidin-conjugated Alexa488 (green) antibodies. Dashed lines demark nuclear boundaries, as determined using DAPI staining. Scale bar represents 50 μM. (F) Western blot analysis of Flp-In T-Rex HEK293 cells treated with tetracycline for 25 h to induce expression of the indicated construct. Cells containing BioID2-FLAG-eGFP-NLS were induced with 0.001 μg/mL tetracycline, whereas BioID2-FLAG-SLX4 and SLX4-FLAG-BioID2 were induced with 1 μg/mL tetracycline. FLAG-tagged proteins were immunoprecipitated from 2.5 mg of extract using α-FLAG beads and analyzed by western blotting for the indicated proteins (left). In parallel, 20 μg cell extract was analyzed by western blotting for the indicated proteins (right). β-tubulin was used as a loading control. Abbreviations: UB = unbound fraction, B = bound fraction on α-FLAG beads.

Figure 2

BioID and AP-MS reveal known and novel SLX4-binding partners. (A) Comparison of the total number of high-confidence preys that appeared in at least two biological replicates of BioID (left), AP-MS (right), or both (middle) with a Bayesian false discovery rate (BFDR) of ≤1%. A detailed summary of the overlapping candidates categorized by protein function is shown below the Venn diagrams. (B,C) Approximation of previously identified SLX4 interactors versus novel protein partners detected by BioID (B) or AP-MS (C). (D) Venus-FLAG-SLX4 co-immunoprecipitates SMC1A, CDC73, and PGAM5 from Flp-In T-REx U2OS cells. Cells were treated with 1 μg/mL tetracycline for 48 h to induce expression of the indicated construct. Venus-FLAG-tagged proteins were immunoprecipitated from 3 mg of extract using GFP-TRAP beads and analyzed by western blotting for the indicated proteins (left). In parallel, 20 μg cell extract was analyzed by western blotting for the indicated proteins (right).

Figure 3

Gene ontology (GO) analysis of high-confidence SLX4 interacting proteins detected by BioID and AP-MS. (A) Cellular component GO terms associated with SLX4 protein hits captured by BioID (red) or AP-MS (blue). Terms were selected from Revigo after applying a small list cutoff (0.5) to minimize redundancy.P-values as listed by Princeton Generic GO Term finder. GO terms related to protein complexes were removed for clarity. Graphics were created with BioRender.com. (B) Bubble plot displaying biological function GO terms associated with SLX4-binding proteins captured by BioID (red) or AP-MS (blue). Terms were selected from Revigo after applying a medium list cutoff (0.7).P-values as listed by Princeton Generic GO Term finder. Fold enrichment represents the number of genes associated with a specific GO function in each proteomics dataset compared to the number of genes related to that function in the human genome.

Figure 4

Network maps of SLX4-binding proteins identified by BioID and AP-MS. (A,B) High-confidence SLX4 interactors detected in at least two biological replicates of BioID (A) or AP-MS (B). Proteins (nodes) were manually clustered in Cytoscape based on biological function. Spectral count is a measure of the number of times a specific protein is detected by mass spectrometry. Unique peptide count reflects the number of peptides detected by mass spectrometry that correspond to one protein. Values were averaged over 2 or 3 biological replicates and are represented according to the scale bars found to the right of each network.

Figure 5

Compilation of a comprehensive SLX4 interaction network using proximity labeling and affinity purification proteomics. SLX4 protein interaction network based on preys identified in at least two biological replicates of BioID and AP-MS. Proteins (nodes) were clustered using Cytoscape ClusterONE, which uses the STRING database to group nodes into functional complexes; the confidence score cutoff was set to 0.70 (high-confidence). Line (edge) thickness corresponds to STRING scores, with thicker edges representing higher confidence interactions between nodes. Shapes and colors correspond to the STRING classification of each protein, where filled circles represent high significance nodes and filled rectangles represent nodes with multiple clusters (overlap). Gray circles and diamonds represent the least significant nodes and outliers (unclustered interactions), respectively. White circles represent nodes that did not pass the confidence cutoff score. The SLX4 node is outlined in red.

Figure 6

Functional landscape of the SLX4 interaction network. Candidate SLX4-binding partners identified in at least two biological replicates of BioID and AP-MS were grouped based by function. Shapes indicate the method(s) used to identify high-confidence preys, as defined in the legend (i.e., BioID, AP-MS, or both). Known SLX4 interactors, as defined in Table 1, are bolded and italicized. Biological processes and molecular functions previously associated with SLX4 are outlined with thick lines. Graphics were created with BioRender.com.