How to Analyze Protein-Protein Interaction: Top Lab Techniques

Introduction

Understanding protein-protein interactions (PPIs) is essential for deciphering the molecular mechanisms underlying virtually every cellular process, from signal transduction and gene expression to immune responses and metabolic regulation. Mapping these interactions provides insight into fundamental biology and accelerates drug discovery, biomarker identification, and systems biology modeling.

In this article, we provide a comprehensive overview of the most widely used experimental and computational techniques for studying protein interactions, including FRET, BiFC, proximity labeling, yeast two-hybrid (Y2H), pull-down assays, co-immunoprecipitation (Co-IP), mass spectrometry (MS), and in silico prediction methods. Each method is discussed in terms of its principle, workflow, key examples, and research applications, offering researchers a valuable reference for selecting the appropriate tools to unravel the complex interactome of their proteins of interest.

Ⅰ. Fluorescence Resonance Energy Transfer (FRET)

Principle:

FRET is a non-radiative energy transfer that occurs by dipole–dipole coupling from an excited state donor (D) fluorophore to a ground state acceptor (A; another fluorophore or a quencher) when appropriate spectral overlap and proximity requirements are satisfied.

Figure 1. Overlap of donor emission (Em.) and acceptor absorption (Abs.) spectra and the resulting spectral overlap integrand (red curve) (Algar et al., 2019^[1])

Example:

To create a FRET biosensor capable of detecting the dimerization of estrogen receptor alpha (ER α), Han et al. (2024) first used AlphaFold to analyze the proximity between the two terminals of the structure of the ER α homodimer. Han et al. (2024) then attached mNeonGreen (mNeonG) and enhanced cyan fluorescent protein (ECFP) as FRET pairs to the N-terminal of each ER α monomer. mNeonG and ECFP were chosen as acceptor and donor FPs, respectively, since their overlap integral can sufficiently induce resonance energy transfer, making them suitable for use as FRET biosensors. They introduced 10 μM of 4-hydroxytamoxifen (a compound with a high affinity for ERα), leading to notable changes in FRET efficiency 30 min after treatment.

Figure 2. Mode of action of ER αα FL FRET biosensors (Han et al., 2025^[2])

Applications:

• Use for nanoscale proximity detection with a typical range of roughly 1–10 nm (up to ~20 nm for some atypical FRET pairs). Monitor intracellular protein–protein interactions' dynamics to provide signal transduction within countless bioassays.
• Offer real-time, in situ detection without any intervention.

Ⅱ. Bimolecular Fuorescence Complementation (BiFC)

Principle:

This approach is based on complementation between two nonfluorescent fragments of the yellow fluorescent protein (YFP) when they are brought together by interactions between proteins fused to each fragment.

Figure 3. Schematic diagram representing the principle of the BiFC assay (Kerppola T. K. 2006^[3])

Example:

To identify YFP fragments that can associate to form a bimolecular fluorescent complex, Hu et al. (2002) constructed bicistronic E. coli expression vectors encoding the bZIP domains of Jun and Fos fused to N- and C-terminal fragments of EYFP (S65G, S72A, T203Y), respectively. They divided YFP into two fragments at several nonconserved amino acid residues within loops at either end of the β-barrel structure. They fused these fragments at the C-terminal ends of the bZIP domains of Fos and Jun. None of the proteins exhibited detectable fluorescence when expressed alone. The fusion proteins with the highest fluorescence intensity (bFosYC and bJunYN) were chosen for further analysis^[4].

Applications:

• Identify novel protein interactions without prior knowledge of the structural basis of the interaction.
• Screen for interactions with proteins expressed from a plasmid library using fluorescence-activated cell sorting or high-throughput transfection analysis.

Ⅲ. Proximity Labeling (PL)

Principle:

In PL (Proximity Protein Labeling), a promiscuous labeling enzyme is targeted by genetic fusion to a specific protein or subcellular compartment. The addition of a small-molecule substrate initiates covalent tagging of endogenous proteins within a few nanometers of the promiscuous enzyme. Subsequently, the biotinylated proteins are harvested using streptavidin-coated beads and identified by mass spectrometry (MS).

v TurboID: a promiscuous mutant of Escherichia coli biotin ligase

Example:

To generate new promiscuous mutants, Branton et al. (2018) used error-prone PCR to mutagenize BirAR118S, generating a library of ~107 mutants, each with an average of about two amino acid mutations relative to the template. This library was then displayed on the yeast surface as a fusion to the Aga2p mating protein. They added biotin and ATP to the yeast pool to initiate promiscuous biotinylation, followed by streptavidin-fluorophore to stain biotinylation sites on the surface of each yeast cell. FACS was used to enrich cells with a high degree of self-biotinylation over cells with low or moderate self-biotinylation. This is how they found the TurboID system.

Figure 4. Yeast-display-based selection scheme of TurboID system (Branon et al., 2018^[5])

Applications:

• Study dynamic protein-protein interaction networks.
• Enable proteomic analysis of specific cell types, organelles, and subcellular structures.

v APEX2: an engineered soybean ascorbate peroxidase

Example:

APEX2 is genetically targeted to a cellular organelle or protein complex of interest. Then, live cells are treated for 1 minute with H2O2 in the presence of biotin-phenol. APEX2 catalyzes the one-electron oxidation of biotin-phenol to generate a very short-lived biotin phenoxyl radical. This radical covalently tags endogenous proteins proximal to APEX2, allowing their subsequent enrichment using streptavidin beads and identification by mass spectrometry.

Figure 5. Yeast-display-based selection scheme of APEX2 system (Lam et al., 2015^[6])

Applications:

• Intracellular specific protein imaging by EM.
• Spatially-resolved proteomic mapping.

Ⅳ. Yeast Two-Hybrid (Y2H)

Principle:

The system relied on the fact that the DNA-binding domain (BD) and transcriptional activation domain (AD) of the S. cerevisiae transcription factor GAL4 can be separated. When separated, neither can drive transcription from Gal4-responsive promoters on its own. However, a functional transcription factor can be reconstituted and drive transcription when a protein fused to a BD interacts with a protein fused to an AD. In turn, reporter gene activity allows one to infer a direct interaction between the proteins fused to the AD and BD.

Figure 6. Schematic of the Y2H system (Galletta et al., 2015^[7])

Example:

To investigate the interaction between rice disease resistance protein OsRLR1 and WRKY transcription factors. Du et al.(2021) constructed AD/BD fusion vectors containing the CC domain of OsRLR1 (OsRLR1CC) and truncated forms of five rice WRKY proteins (OsWRKY13/19/47/68/76). Screening in yeast AH109 revealed that only OsWRKY19 could grow on QDO/AbA selective media and activate reporter genes, indicating a specific interaction^[8].

Applications:

• Classical Y2H is exclusively applicable to nuclear-localized proteins, whereas membrane proteins, secretory proteins, or organelle-resident proteins rely on variant techniques such as the split-ubiquitin system.
• Genome-wide mapping of protein-protein interactions to construct cellular signaling networks, and can be performed directly in living cells.

Ⅴ. GST Pull-Down Assay

Principle:

The pull-down assay is a powerful in vitro technique for studying protein-protein interactions. It immobilizes a "bait" protein (often tagged with GST, His, or other affinity tags) onto beads coated with specific ligands. When incubated with a lysate containing potential "prey" proteins, binding partners are captured through direct physical interaction. After washing away non-specific binders, the retained proteins are eluted and analyzed by SDS-PAGE or mass spectrometry.

Example: Shown in Figure 7.

Figure 7. Schematic representation of the Pull-Down procedure (Xu et al., 2020^[9])

Applications:

• Validate the direct interaction between two known proteins in vitro.

Ⅵ. Co-immunoprecipitation (Co-IP)

Principle:

Co-IP is a widely used in vivo technique for detecting physical protein interactions. The method relies on antibodies specific to a target ("bait") protein to capture it along with its binding partners ("prey") from cell lysates.

Example: Shown in Figure 8.

Figure 8. Schematic representation of the Co-immunoprecipitation procedure (Xu et al., 2020^[9])

Applications:

These reflect protein-protein interactions occurring in living cells or organisms. However, we are unable to determine whether the observed bait-prey association represents a direct physical contact or requires intermediary proteins for complex formation.

Ⅶ. Mass Spectrometry (MS)

Principle:

Many strategies have been developed to enrich proteins of interest and their binding partners. Enriched proteins with their binding partners are submitted to mass spectrometry for protein identification. After data filtration with bioinformatics tools, the protein‒protein interaction network or interactome can be established with lists of high-confidence interaction proteins (HCIPs).

Figure 9. Strategies for studying interactomes with MS (Zhen Chen and Junjie Chen, 2021^[10])

Procedures:

• Sample Preparation & Protein Complex Enrichment
  • Affinity Purification: Use antibodies, epitope tags, or Tandem Affinity Purification (TAP) to capture target proteins and their interaction partners. Examples include Immunoprecipitation (IP) or affinity chromatography.
  • Proximity Labeling: Techniques like TurboID or APEX2 use engineered enzymes to label proximal interacting proteins in live cells, followed by biotin-streptavidin purification.
  • Chemical Crosslinking: Stabilize transient/weak interactions using crosslinkers that covalently link spatially close amino acid residues.
• Separation & Purification of Protein Complexes
• Chromatography: Size Exclusion Chromatography or Ion Exchange Chromatography to isolate complexes from free proteins.
• Density Gradient Ultracentrifugation: Separate complexes by molecular weight for large-scale interactome studies.
• Protein Digestion & Peptide Preparation
  • Denaturation: Use reducing agents and alkylating agents to break disulfide bonds.
  • Enzymatic Digestion: Trypsin cleaves proteins into peptides (5-20 amino acids) for mass spectrometry compatibility.
  • Peptide Separation: Reverse-Phase Liquid Chromatography or Capillary Electrophoresis to reduce sample complexity.
• Mass Spectrometry Analysis
  • MS1 (Primary Mass Spectrometry): Measure peptide mass-to-charge ratio (m/z) for precursor ion selection.
  • MS/MS (Tandem Mass Spectrometry): Fragment peptides via Collision-Induced Dissociation or Electron Transfer Dissociation to generate sequence-specific spectra.
• Data Analysis & Validation
  • Database Search: Match spectra to theoretical databases using tools like MaxQuant or Mascot for peptide/protein identification.
  • Interaction Network Construction: Filter High-Confidence Interaction Partners using tools like STRING to generate interactome maps.
  • Validation: Confirm interactions via Western blot, RNA Immunoprecipitation, or Fluorescence Co-localization.

Applications:

• Discover Novel Protein Complexes, and achieve large-scale identification of interaction partners in endogenous protein networks.
• Identify which protein regions mediate binding and how phosphorylation, ubiquitination, or acetylation regulates binding.

Ⅷ. Computational Prediction

The main idea behind predicting protein interactions with computers is to use bioinformatics and machine learning. These tools infer potential interaction likelihood by analyzing protein features such as sequence, structure, evolutionary relationships, function, or network properties. Computational approaches overcome experimental limitations, enabling rapid genome-wide screening, narrowing down candidate interaction pairs, and guiding site-directed mutagenesis and binding site analysis. Here are some computational methods for protein prediction:

• Protein Structure and Interaction Prediction: SWISS-MODEL^[11] and AlphaFold3^[12] generates high-accuracy structural models. AlphaFold-Multimer^[13] and ZDOCK^[14] predict interaction interfaces for protein-protein complexes.
• Functional Annotation and Classification: Infer function by aligning sequences to known functional proteins by BLAST. Predict enzymatic activity, subcellular localization, etc., from sequence/structural features by DeepGO^[15].
• Dynamic Behavior Simulation: Simulate conformational changes and binding dynamics by using GROMACS^[16] or AMBER^[17].
• Functional Network and Pathway Analysis: Visualize networks and identify hub proteins by using Cytoscape^[18]. Link proteins to biological pathways(e.g., KEGG Mapper^[19] and DAVID).

Applications:

• Predict drug targets and binding sites.
• Evaluate structural/functional effects of disease mutations.
• Design artificial proteins or optimize enzyme activity.

Comparison of The Different Protein-Protein Interaction (PPI) Methods

Conclusion

Protein interaction mapping is a cornerstone of molecular and cellular biology, enabling researchers to uncover interaction networks that govern cellular physiology and pathophysiology. From real-time live-cell imaging techniques like FRET and BiFC to large-scale proteomics approaches such as proximity labeling and MS-based interactome profiling, each method offers unique advantages depending on the experimental context and resolution required. As computational tools like AlphaFold-Multimer and network-based analytics continue to evolve, they are reshaping how we predict and interpret PPIs at the structural and systems levels. By integrating diverse methodologies, researchers can achieve a more complete, dynamic, and quantitative understanding of protein networks across biological systems.

References

1. ALGAR W R, HILDEBRANDT N, VOGEL S S, et al. FRET as a biomolecular research tool — understanding its potential while avoiding pitfalls. Nature Methods, 2019,16(9): 815-829. DOI: 10.1038/s41592-019-0530-8
2. HAN K, SUH J S, CHOI G, et al. Novel FRET‐Based Biosensors for Real‐Time Monitoring of Estrogen Receptor Dimerization and Translocation Dynamics in Living Cells. Advanced Science, 2024,10(1). DOI: 10.1002/advs.202406907
3. KERPPOLA T K. Design and implementation of bimolecular fluorescence complementation (BiFC) assays for the visualization of protein interactions in living cells. Nat Protoc, 2006,1(3): 1278-1286. https://doi.org/10.1038/nprot.2006.201
4. HU C, CHINENOV Y, KERPPOLA T K. Visualization of interactions among bZIP and Rel family proteins in living cells using bimolecular fluorescence complementation. Mol Cell, 2002,9(4): 789-798. DOI: 10.1016/s1097-2765(02)00496-3
5. BRANON T C, BOSCH J A, SANCHEZ A D, et al. Efficient proximity labeling in living cells and organisms with TurboID. Nature Biotechnology, 2018,36(9): 880-887. DOI: 10.1038/nbt.4201
6. LAM S S, MARTELL J D, KAMER K J, et al. Directed evolution of APEX2 for electron microscopy and proximity labeling. Nature Methods, 2015,12(1): 51-54. https://doi.org/10.1038/nmeth.3179
7. RUSAN M NASSER B J G. A yeast two-hybrid approach for probing protein-protein interactions at the centrosome. Methods Cell Biol., 2015,129: 251-277. DOI: 10.1016/bs.mcb.2015.03.012
8. DU D, ZHANG C, XING Y, et al. The CC‐NB‐LRR OsRLR1 mediates rice disease resistance through interaction with OsWRKY19. Plant Biotechnology Journal, 2021,19(5): 1052-1064. DOI: 10.1111/pbi.13530
9. Chongyi Xu. Pull-down and Co-immunoprecipitation Assays of Interacting Proteins in Plants. Chinese Bulletin of Botany, 2020, 55(1): 62-68.
10. CHEN Z, CHEN J. Mass spectrometry-based protein-protein interaction techniques and their applications in studies of DNA damage repair. Journal of Zhejiang University-SCIENCE B, 2021,22(1): 1-20. doi: 10.1631/jzus.B2000356
11. WATERHOUSE A, BERTONI M, BIENERT S, et al. SWISS-MODEL: homology modelling of protein structures and complexes. Nucleic Acids Research, 2018,46(W1): W296-W303.
12. ABRAMSON J, ADLER J, DUNGER J, et al. Accurate structure prediction of biomolecular interactions with AlphaFold 3. Nature, 2024,630(8016): 493-500. https://doi.org/10.1038/s41586-024-07487-w
13. OMIDI A, MØLLER M H, MALHIS N, et al. AlphaFold-Multimer accurately captures interactions and dynamics of intrinsically disordered protein regions. Proceedings of the National Academy of Sciences, 2024,121(44). DOI: 10.1073/pnas.2406407121
14. CHEN R, LI L, WENG Z. ZDOCK: an initial-stage protein-docking algorithm. Proteins, 2003,52(1): 80-87. DOI: 10.1002/prot.10389
15. KULMANOV M, KHAN M A, HOEHNDORF R, et al. DeepGO: predicting protein functions from sequence and interactions using a deep ontology-aware classifier. Bioinformatics, 2018,34(4): 660-668. DOI: 10.1093/bioinformatics/btx624
16. ABRAHAM M J, MURTOLA T, SCHULZ R, et al. GROMACS: High performance molecular simulations through multi-level parallelism from laptops to supercomputers. SoftwareX, 2015,1-2: 19-25. https://doi.org/10.1016/j.softx.2015.06.001
17. SALOMON FERRER R, CASE D A, WALKER R C. An overview of the Amber biomolecular simulation package. WIREs Computational Molecular Science, 2013,3(2): 198-210. https://doi.org/10.1002/wcms.1121
18. SHANNON P, MARKIEL A, OZIER O, et al. Cytoscape: a software environment for integrated models of biomolecular interaction networks. Genome Res, 2003,13(11): 2498-2504. DOI: 10.1101/gr.1239303
19. KANEHISA M, SATO Y. KEGG Mapper for inferring cellular functions from protein sequences. Protein Science, 2020,29(1): 28-35. DOI: 10.1002/pro.3711