High-throughput protein target mapping enables accelerated bioactivity discovery for ToxCast and PFAS compounds

Abstract

Chemical pollution is a global threat to human health, yet the toxicity mechanism of most contaminants remains unknown. Here, we applied an ultrahigh-throughput affinity-selection mass spectrometry (AS-MS) platform to systematically identify protein targets of prioritized chemical contaminants. After benchmarking the platform, we screened 50 human proteins against 481 prioritized chemicals, including 446 ToxCast chemicals and 35 per- and polyfluoroalkyl substances (PFAS). Among 24,050 interactions assessed, we discovered 35 novel interactions involving 14 proteins, with fatty acid-binding proteins (FABPs) emerging as the most ligandable protein family. Given this, we selected FABPs for further validation, which revealed a distinct PFAS binding pattern: legacy PFAS selectively bound to FABP1, whereas replacement compounds, PFECAs, unexpectedly interacted with all FABPs. X-ray crystallography further revealed that the ether group enhances molecular flexibility of alternative PFAS, to accommodate the binding pockets of FABPs. Our findings demonstrate that AS-MS is a robust platform for the discovery of novel protein targets beyond the scope of the ToxCast program and highlight the broader protein-binding spectrum of alternative PFAS as potential regrettable substitutes.

Fig. 1 |. Benchmarking the AS-MS platform for scalable protein target discovery.

(a) Workflow of AS-MS platform. (b) Benchmarking the AS-MS platform using a double-blind experiment. The AS-MS platform successfully identified a known ligand (blue dot, K_D = 60 nM) of WD repeat-containing protein 5 (WDR5) from a library of ~16,000 compounds. (c) Detection rate of compounds according to their dissociation constants (K_D). (d) Top 10 compound categories in the 1,114-chemical library based on the Chemical Products Categories (CPCat) database^,. (e) The fraction of MS activity compounds in the 1,114-chemical library. (f) A subset of 85 bioactive compounds from the library exhibited AC₅₀ < 10 μM against nuclear receptors.

Fig. 2 |. Mapping interactions between 481 selected chemicals and 50 human proteins.

(a) Overview of 50 selected human proteins. (b) Scatter plot of hits interacting with 50 proteins. The gray line represents a fold change cutoff of 10. Red dots indicate hits with a fold change greater than 10 and a p-value less than 0.05. The fold change for each compound-protein interaction was calculated by dividing the average LC-MS peak intensity of the target protein by the average signal across all other proteins on the plate and p-values were determined using Student’s t-tests. (c) Chemical space of the 1,114 compounds clustered based on their molecular fingerprints and visualized using uniform manifold approximation and projection (UMAP). Detected ligands are highlighted in different colors according to their target proteins, with chemical structures labeled on the plot. (d) Comparison of the top 10 compound categories in the 481 selected compounds (left column) with the categories of ligands that showed protein interactions in the AS-MS screening (right column).

Fig. 3 |. Distinct and novel interactions between ToxCast/PFAS chemicals and FABPs.

(a) Overview of the binding affinity range (Log₁₀ K_D, μM) for ligands interacting with seven FABPs measured by fluorescence displacement assay. Each blue dot represents a single ligand. (b) Principal component analysis (PCA) of ligand binding affinities (Log₁₀ K_D, μM) across seven FABPs. Fluorescence displacement assays were used to determine binding affinities, with 1,8-ANS as the probe for FABP1, 2, 3, 4, 5, and 7, and bisANS for FABP6, while oleic acid (OA) K_D values for the seven FABPs were obtained from the literature^,,. (c) Solvent-accessible volumes (light gray) of FABPs. FABP1 (PDB: 3STM) is shown as a representative structure for the seven FABPs. For FABP isoforms with multiple available crystal structures, mean and standard deviation values were calculated to determine solvent-accessible volume variations. (d) Distribution of known ligands (blue dots) and newly discovered ligands (purple dots) for FABP1 within the chemical space of the 1,040 ToxCast compounds and 74 PFAS compounds. Dose-response results from the fluorescence displacement assay validation are shown, along with the measured dissociation constants (K_D). DTXCID5027304 (CAS, 135080–03-4; top left panel) was tested with six concentrations (3.12, 6.25, 12.5, 25, 50 100 μM); FK-739 (CAS, 136042–19-8; bottom left panel) was tested with nine concentrations (0.12, 0.62, 3.12, 6.24, 12.5, 25, 50, 100, 200 μM); PFNA (CAS, 375–95-1; top right panel) was tested with seven concentrations (1.78, 3.12, 6.25, 12.5, 25, 50 100 μM); PFO3TDA (CAS, 330562–41-9; bottom right panel) was tested with six concentrations (0.37, 1.11, 3.33, 10, 30, 50 μM).

Fig. 4 |. Selective interactions between alternative PFAS and FABPs.

(a) Chemical structures of representative legacy PFAS and PFECAs (alternative PFAS). (b) The relative fluorescence signals were measured when different legacy PFAS (100 μM) were added to compete with the FABP bound fluorescence probes; 1,8-ANS were used for FABP 1, 2, 3, 4, 5, and FABP7, and bisANS was used for FABP6. Dose-response curves showing the competition between increasing compound concentrations and fluorescent probes for binding to the seven FABPs. The three compounds are displayed as follows: PFO3TDA(c) and HFPO-TeA(d). PFO3TDA (CAS, 330562–41-9) was tested with six concentrations (0.37, 1.11, 3.33, 10, 30, 50 μM) across seven FABPs; HFPO-TeA (CAS, 65294-16-8) was tested with six concentrations (3.12, 6.25, 12.5, 25, 50 100 μM) across seven FABPs. (e) PFO3TDA co-crystalized with FABP7 (PDB: 9NIU). The carboxylate head group of PFO3TDA formed two hydrogen bonds with residues arginine 127 and tyrosine 129. (f) The Gibbs free energy difference between ‘trans’ and ‘gauche’ conformations for PFOA, PFO3TDA (CF2-X-X-CF2), and PFO3TDA (O-CF2-CF2-X). (g) The diagram showing the interactions of endogenous ligands (oleic acid, retinoid acid and bile acid), legacy PFAS, and alternative PFAS (long-chain PFECAs) with FABPs and their corresponding peroxisome proliferator-activated receptors (PPARs), distributed across different body tissues^,–.