Detecting protein-protein interactions in live cells
Applications
Watch them meet, and watch them part
splitFAST2 lights up when two proteins meet and goes dark when they separate, so you can follow an interaction, and a drug that breaks it, in living cells.
Why protein-protein interactions are not so easy to watch
Proteins meet, act and part again, often within seconds. The questions that matter are about time: when does an interaction form, how long does it last, and can a drug break it? Answering them takes a reporter that lights up when two proteins meet and goes dark when they separate, in a living cell.
Ever wished for reversible BiFC?
Classic BiFC cannot let go. Once the two halves of a GFP-family protein meet, the barrel folds and the chromophore matures: “complementation is followed by chromophore maturation, which results in irreversible complex formation” (Tebo and Gautier, Nat. Commun. 2019). Split-GFP forms “highly stable and essentially irreversible complexes” (David et al., Plant J. 2026). You see that two proteins met, never that they parted.
splitFAST is BiFC without a chromophore to mature. Its two halves hold a small fluorogen only while they are together. When the proteins part, the fluorogen leaves and the signal goes.
How splitFAST works
FAST is a 14 kDa tag that switches its fluorogen on: dark in solution, bright when bound. Split it, fuse one half to each protein, and the halves light up only when the proteins bring them together.
We recommend splitFAST2 (2024): its halves barely attract each other on their own, about seventy times less than in the 2019 version, so the signal reports your proteins, not the tag (Rakotoarison et al., ACS Chem. Biol. 2024).
When to choose splitFAST2
| Your question | Method | Why |
|---|---|---|
| When does it form, and when does it come apart? | splitFAST2 | Rapid, reversible complementation in live cells |
| Can a drug break a complex that has already formed? | splitFAST2 | An irreversible reporter cannot show a complex coming apart |
| Is a third protein in the complex? | Tripartite-split-pFAST | Signal only when all three partners assemble |
| Where is the complex, below the diffraction limit? | splitFAST2 with STED | Super-resolution of interactions in live cells |
Versus NanoBiT: both are reversible, with halves in the same affinity range (SmBiT 190 µM, BioEssays 2025; splitFAST2 about 220 µM). The difference is the signal: fluorescence shows where the complex is, even under STED, and no substrate is consumed.
Versus FRET: also reversible, but it needs controlled stoichiometry of the two partners (Bottone et al., Nat. Methods 2023) and “specialized imaging apparatuses not readily available in all laboratories” (Calì et al., Nat. Rev. Mol. Cell Biol. 2025). splitFAST reads on a standard confocal, a flow cytometer or a plate reader.
What researchers have measured with it
Drugs that break a complex
A 2026 study screened 907 ORFs with BiFC, then needed splitFAST2 for the one question BiFC cannot answer: does the drug EI-52 break the ERK1/MyD88 complex once it has formed? “The reversible nature of splitFAST2 allows assessing the potential destabilizing effect of EI-52 postcomplex formation” (J. Biol. Chem. 2026). In a preprint, splitFAST showed known inhibitors breaking the HIV-1 Nef homodimer, cell-free, on a plate reader (bioRxiv 2026, not yet peer reviewed).
Signalling and the cell cycle
MEK1/ERK2 in the cytosol, ERK2/MKP1 in the nucleus, K-Ras/Raf1 at the plasma membrane (Tebo and Gautier 2019). FAM122A with B55α, a regulatory subunit of the phosphatase PP2A, in live U-2 OS cells (Wasserman et al., Nat. Commun. 2024).
Binding specificity, by flow cytometry
13 pairs of JUN variants and bZIP partners, scored quantitatively in human cells and matched against a yeast deep mutational scan (Bendel et al., Nat. Commun. 2024).
Receptors, neurons and axons
GPR158/PLCXD2 at the plasma membrane, in HEK293T cells and primary hippocampal neurons (Verpoort et al., Dev. Cell 2025). fignl1/RAD51 in the growth cones of retinal axons, in retina explants, by TIRF (bioRxiv 2026, not yet peer reviewed).
Plants, yeast and an archaeon
CAM5/ceQORH in Arabidopsis thaliana and Nicotiana benthamiana, with the splitFAST2 generation (David et al., Plant J. 2026). Cox4/Cox5a as a readout of respiratory complex IV assembly in yeast (bioRxiv 2026, not yet peer reviewed). FdhA/FdhB, the two subunits of formate dehydrogenase, in the methanogenic archaeon Methanococcus maripaludis under anoxia (J. Bacteriol. 2022; see imaging in anaerobes).
The yeast and archaeon studies tagged both partners on the chromosome, at native expression level.
Two colours, far red, super-resolution, three partners
Split-greenFAST and split-redFAST follow two transient interactions at once (Tebo et al., Nat. Chem. Biol. 2021), and split-frFAST takes interaction imaging to the far red (Angew. Chem. 2020). splitFAST2 complexes, such as Bcl-xL with the BH3 domain of BAK at mitochondria, have been imaged by STED in live HeLa cells (Board and Gautier, ChemBioChem 2025). And tripartite-split-pFAST reports three-protein complexes in the cytosol, at the plasma membrane, in the nucleus and at junctions between organelles (Bottone et al., Nat. Commun. 2025).
A disease model
In a cellular model of calcific aortic valve disease, human valve interstitial cells from control donors under osteogenic treatment, the TERT/STAT5A interaction read with splitFAST was elevated from day 3 to day 7 against untreated controls, with proximity ligation and co-immunoprecipitation as orthogonal assays (Circ. Res. 2025). The authors plan to use the reporter to screen for compounds that block the interaction.
Nine of these studies come from laboratories other than the one that invented splitFAST, and six name a Twinkle Factory fluorogen in their methods.
Good to know
- Dimmer than a split fluorescent protein. An outside review says so, and adds that splitFAST “has revealed the dynamic regulation of organelle contact sites” (Gamuyao and Chang 2024).
- No head-to-head with split-GFP yet. The reversibility advantage rests on the mechanism and on each system’s own data.
- The fluorogen matters. A preprint describes a fluorogen that turns splitFAST into a dimeriser (bioRxiv 2024). Use the fluorogens the reporter was characterised with.
- Built for live cells, where reversibility counts. It also reads in liquid culture and on purified protein.
Setting up
- Reporter: splitFAST2.
- Fluorogen: TFLime (HMBR) or TFCoral (HBR-3,5DOM), at 5 to 15 µM in the studies above; TFPoppy (HPAR-3OM) for split-frFAST in the far red.
- Readouts: confocal, TIRF and STED microscopy; flow cytometry; plate readers, cell-free assays included.
- Controls: a non-interacting pair, and an orthogonal assay such as proximity ligation or co-immunoprecipitation.
Frequently asked
Can I screen compounds with it?
It has been used to test known inhibitors, in live cells and cell-free, and to ask whether a hit breaks a complex that has already formed. No compound-library screen with splitFAST has been published yet. Planning one? Ask us.
Which fluorogens do I need?
splitFAST2 was characterised with fluorogens emitting between 540 and 600 nm. See the products page, or ask us which one suits your microscope and your other labels.
Making an interaction happen instead
Two questions, two tools, one chemistry. Is this interaction happening? splitFAST2. Do you want to make it happen on purpose, then undo it? That is CATCHFIRE.
Disease models are described as reported by the cited studies. FAST, splitFAST and CATCHFIRE reagents are for research use only, not for diagnostic or therapeutic use.
Recent papers: splitFAST for protein-protein interactions
Temporal tuning of actin nucleation by the repurposed DNA repair complex RAD51/fignl1 encodes distinct axon guidance decisions
splitFAST enters developmental neuroscience. The fignl1 to RAD51 interaction is imaged live in the growth cones of retinal ganglion cell axons in retina explants, by TIRF, along F-actin filopodia, with clean negatives: a RAD51-binding-deficient fignl1 mutant and an NFAST-FRB pair give no signal in any growth cone. It is then quantified in COS-7 cells to show that PKC phosphorylation of fignl1 promotes the complex's disassembly. Obtained with TFLime
A sensitive interactome-based approach to characterize drug activity in living human cells
splitFAST chosen for what BiFC cannot do. The authors note that bimolecular fluorescence complementation is irreversible and therefore cannot report a drug destabilising a complex that has already formed, so they introduce splitFAST2 for that half of the work and read drug-sensitive ERK1 interactions in living cells. Reversibility is the stated reason for the choice.
HIV-1 Nef Homodimerization as a Structural Mechanism for Kinase Activation and Small Molecule Inhibitor Action
Showed that mutating the HIV-1 Nef dimer interface residues Leu112 and Phe121 abolishes dimerization while preserving the Nef fold, giving monomeric complexes with Hck kinase domains. Used a splitFAST complementation assay to provide direct evidence that small-molecule inhibitors disrupt Nef homodimer formation in solution.
The fluorescence‐activating and absorption‐shifting tag (FAST), a versatile protein marker for live plant cell imaging
Established FAST, greenFAST/redFAST and splitFAST for versatile multicolor imaging in plant cells.
Chemogenetic Modulation of Luciferase Emission Color for Imaging and Sensing
Introduced LumiFAST, a chemogenetic way to tune luciferase emission color, for bioluminescence imaging and sensing.
Translational antagonism enables bet-hedging in yeast via mitochondrial bistability
Used splitFAST to report cytochrome c oxidase assembly in single yeast cells, showing that a bistable switch in mitochondrial activity splits a population into fermenting arrestors and respiring recoverers.
Super-Resolution Live-Cell Mapping of Protein-Protein Interactions Using Chemogenetic splitFAST
Mapped protein-protein interactions at super-resolution in live cells using chemogenetic splitFAST.
Synthetic protein degradation circuits using programmable cleavage and ligation by Sortase A
Used splitFAST as the readout of a peptide bond being made or broken: Sortase A ligates the two fragments to switch fluorescence on, and cleaves a full-length FAST carrying an LPETGGG motif to switch it off, driving logic-gated protein degradation in live HeLa cells.
Design and evaluation of a tripartite chemogenetic fluorescent reporter for visualizing ternary protein complexes
Engineered a tripartite (three-fragment) splitFAST that lights up only when three proteins assemble, visualizing ternary complexes in live cells.
Non-Canonical TERT Activity Initiates Osteogenesis in Calcific Aortic Valve Disease
splitFAST serves here as a screening instrument. Telomerase reverse transcriptase acts outside its canonical role to drive osteogenic reprogramming in human aortic valve interstitial cells, and the TERT to STAT5A interaction is read by fusing the two proteins to NFAST and CFAST with HBR-3,5DOM at 5 micromolar. The signal is highly elevated in calcifying cells from day three of osteogenic induction
The genetic architecture of protein interaction affinity and specificity
Used splitFAST to quantify how every mutation in the transcription factor JUN reshapes binding across a protein family, mapping interaction affinity and specificity
FAM122A ensures cell cycle interphase progression and checkpoint control by inhibiting B55α/PP2A through helical motifs
Used splitFAST complementation to characterize how FAM122A binds and inhibits the B55-alpha/PP2A phosphatase to control the cell cycle.
Improving Split Reporters of Protein–Protein Interactions through Orthology-Based Protein Engineering
Improved splitFAST's brightness and performance for detecting protein-protein interactions using orthology-based protein engineering.
Fluorescence-Activating and Absorption-Shifting Tags for Advanced Imaging and Biosensing
An Accounts review by the FAST inventor summarizing how FAST-family tags and biosensors work and their imaging and biosensing applications.
The Fluorescence-Activating and Absorption-Shifting Tag (FAST) Enables Live-Cell Fluorescence Imaging of Methanococcus maripaludis
Showed FAST enables live-cell fluorescence imaging of the methanogenic archaeon Methanococcus maripaludis under anoxic conditions.
Protein-fragment complementation assays for large-scale analysis of protein-protein interactions
A review of protein-fragment complementation assays for large-scale PPI analysis, including splitFAST.
Orthogonal fluorescent chemogenetic reporters for multicolor imaging
Introduced greenFAST and redFAST, spectrally orthogonal FAST variants, for simultaneous two-color imaging in live cells and zebrafish.
A Far-Red Emitting Fluorescent Chemogenetic Reporter for In Vivo Molecular Imaging
Introduced frFAST, a far-red FAST variant, for spectral multiplexing and deep-tissue in-vivo imaging in zebrafish and mammalian cells.
A split fluorescent reporter with rapid and reversible complementation
Introduced splitFAST, FAST split into two fragments that fluoresce only when reunited, giving a rapid, reversible readout of transient protein-protein interactions in live cells.
No publications match your filters.





