Target Deconvolution

The direct physical capture of target proteins using an immobilized small molecule as bait. Chemical proteomics approaches like pull-down assays and affinity chromatography isolate binding partners from cellular lysates.

• Compound immobilization: The bioactive molecule is tethered to a solid matrix via a linker, preserving binding capacity
• Competition binding: Free compound competes with immobilized probe to confirm specificity
• Quantitative mass spectrometry: SILAC or TMT labeling distinguishes true targets from background binders
• Key limitation: Structural modification for linker attachment may alter binding pharmacology

A label-free method that detects target engagement by measuring ligand-induced thermal stabilization of proteins in intact cells or lysates. Thermal proteome profiling extends CETSA to proteome-wide target identification.

• Principle: Ligand binding increases a protein's melting temperature (Tm), conferring resistance to heat-induced denaturation
• Readout: Soluble protein fraction quantified by mass spectrometry or Western blot across temperature gradients
• Isobaric labeling: TMTpro 16-plex enables multiplexed analysis of drug-treated vs. vehicle samples
• Advantage: No chemical modification of the compound required; works in living cells

A method exploiting the reduced protease susceptibility of a target protein upon ligand binding. Limited proteolysis coupled with mass spectrometry identifies stabilized protein domains.

• Mechanism: Ligand-induced conformational changes protect the binding site from non-specific proteases like pronase or thermolysin
• Differential analysis: Protease-resistant bands in drug-treated vs. control samples are excised and identified by LC-MS/MS
• Strengths: No compound modification needed; detects both direct and indirect binding partners
• Complementary technique: Often paired with SPROX (Stability of Proteins from Rates of Oxidation) for orthogonal validation

A covalent capture strategy using photoreactive functional groups to crosslink a compound to its target upon UV irradiation. Diazirine and benzophenone moieties are common photoreactive warheads.

• Click chemistry: Alkyne or azide handles enable subsequent conjugation to biotin or fluorophores for enrichment
• In situ labeling: Crosslinking occurs in live cells before lysis, capturing transient or low-affinity interactions
• Quantitative proteomics: SILAC-based comparison of labeled proteins in presence vs. absence of excess competitor
• Resolution: Identifies specific amino acid residues at the binding interface when coupled with tandem mass spectrometry

Genome-wide perturbation screens that identify genes whose knockout or activation modulates compound sensitivity. Resistance and sensitization screens reveal the drug's target pathway.

• CRISPR knockout (GeCKO): Loss-of-function screens identify genes essential for compound activity
• CRISPR activation (CRISPRa): Gain-of-function screens pinpoint genes whose overexpression confers resistance
• CRISPR interference (CRISPRi): Tunable knockdown enables dose-response genetic interaction mapping
• DrugZ algorithm: Computational tool for identifying significantly enriched or depleted sgRNAs in drug-treated populations
• Haploinsufficiency profiling: Heterozygous deletion libraries in yeast (HIP) or human cells (HAP1) reveal direct targets through increased drug sensitivity

In silico methods that predict molecular targets by comparing a query compound against databases of known ligand-target interactions. Pharmacophore mapping and molecular docking are core techniques.

• Similarity ensemble approach (SEA): Compares chemical similarity of query molecule to ligand sets for thousands of targets using BLAST-like algorithms
• SwissTargetPrediction: Combines 2D and 3D similarity measures with a naive Bayesian classifier trained on 370,000 bioactive compounds
• Reverse docking: Docks the query compound into a panel of protein binding sites; targets ranked by docking score
• Network-based inference: Guilt-by-association methods propagate known drug-target relationships across protein-protein interaction networks
• Limitation: Performance depends on completeness of reference databases like ChEMBL and DrugBank