Yeast Two-Hybrid (Y2H) Assay: Principle, Workflow, and Limitations

The yeast two-hybrid (Y2H) assay detects protein-protein interactions by reconstituting a functional transcription factor inside yeast cells: if a ‘bait’ protein and a ‘prey’ protein physically interact, the reporter genes they activate together tell you so. It is a genetic, cell-based screening method, which puts it in a different category from biochemical pulldown methods like co-immunoprecipitation (Co-IP) — Y2H is built for large-scale discovery screening across many candidate partners, while Co-IP is built for confirming one specific, already-suspected interaction in its native cellular context. The two are complementary, not competing: a common workflow uses Y2H to generate interaction candidates and Co-IP (or an equivalent orthogonal method) to confirm the strongest hits.

The principle: reconstituting a transcription factor in yeast

Many eukaryotic transcription factors, including the yeast activator GAL4, are built from two separable, independently functional domains: a DNA-binding domain (DBD) that anchors the factor to a specific upstream activating sequence, and an activation domain (AD) that recruits the rest of the transcription machinery. Split apart, neither domain alone can activate transcription — the DBD has nothing to recruit, and the AD has nothing to bind to.

Y2H exploits this by fusing the DBD to a bait protein and the AD to a prey protein, each expressed from its own plasmid. If bait and prey physically interact, they bring the DBD and AD back into proximity, reconstituting a functional GAL4-like activator that switches on reporter genes placed under GAL4-responsive promoters — classically an auxotrophic marker such as HIS3 or URA3 (scored as colony growth on selective, nutrient-dropout media) and/or lacZ (scored as a colorimetric readout, colony color on an X-gal-containing plate or a quantitative β-galactosidase assay). No interaction, no reconstituted activator, no reporter activation — that absence is the negative result the assay is built around.

Workflow: from bait plasmid to reporter readout

1. Bait and prey plasmid construction

The bait open reading frame is cloned in-frame with the DBD in one vector; the prey (a single candidate ORF, or a cDNA/ORF library for unbiased screening) is cloned in-frame with the AD in a compatible vector. Standard gene cloning and construct-verification practice applies to both.

2. Testing the bait for autoactivation

Before screening, the bait construct alone (with an empty AD vector) is transformed into the reporter strain and checked for reporter activation on its own. A bait that self-activates — often because it carries an intrinsic transcriptional-activation motif, or binds DNA nonspecifically — will produce a positive readout regardless of any real prey interaction. This is the single most common source of unusable baits, and most protocols either truncate the bait to remove the offending region or drop it in favor of a different domain of the same protein.

3. Mating or co-transformation

Bait and prey plasmids are combined in a single diploid cell by one of two routes: co-transformation (both plasmids introduced into one haploid strain directly) or mating (bait in a MATa strain, prey in a MATα strain, mixed and allowed to mate into diploids). Mating-based approaches are the standard for library-scale screens because they scale far better than transforming a library directly.

4. Selective media and reporter readout

Diploids are plated on media selecting first for both plasmids (auxotrophic markers on the vectors themselves), then on media additionally selecting for reporter-gene activation (e.g., histidine- or uracil-dropout plates, often with a competitive inhibitor such as 3-amino-1,2,4-triazole added to suppress the background growth auxotrophic reporters are prone to). Surviving, reporter-positive colonies are the candidate interactions.

5. Confirming a hit

Candidate prey plasmids are rescued from surviving colonies, sequenced to identify the interacting protein, and retested by fresh co-transformation or mating with the original bait (and, critically, with an unrelated control bait) to rule out artifacts of the first pass. A genuine hit is then typically validated by an orthogonal, non-yeast method before being reported with confidence — Co-IP is the most common choice, since it tests the same two proteins in a mammalian or other native cellular context rather than in yeast.

Known limitations

Y2H’s failure modes run in both directions, and both matter for how much weight to put on a result:

• False positives. Self-activating baits (see above), prey proteins that are ‘sticky’ and activate reporters with many unrelated baits, and interactions that only occur because bait and prey are both forced into the yeast nucleus at high, non-physiological concentrations all inflate the hit list. This is why unconfirmed Y2H hits are treated as candidates, not conclusions.
• False negatives from missing post-translational modifications. Because the interaction has to happen inside a yeast cell, any interaction that depends on a post-translational modification yeast doesn’t perform on the bait or prey — certain phosphorylation events, glycosylation patterns, or other modifications specific to higher-eukaryote signaling — will not be detected, even if the interaction is real in the organism of interest.
• Membrane protein incompatibility. The classical Y2H system requires bait and prey to enter the yeast nucleus to reconstitute the transcription factor, which is incompatible with most integral membrane proteins. Variants exist that relocate the readout (e.g., split-ubiquitin systems for membrane proteins), but the standard GAL4 Y2H format is effectively limited to soluble, cytoplasmic/nuclear proteins.

Y2H vs. Co-IP: discovery screening vs. targeted confirmation

The two methods answer different questions and are usually most useful in sequence rather than as substitutes for each other:

• Y2H is unbiased-discovery-shaped: a single bait can be screened against a whole cDNA/ORF library in one experiment, making it the practical choice for building an initial interaction map or finding novel partners for a protein of interest. It can also pick up weak or transient interactions that a biochemical pulldown might lose during lysis and wash steps, because binding happens continuously in living cells rather than in a single lysate snapshot.
• Co-IP is targeted-confirmation-shaped: it tests whether two specific, already-suspected proteins come down together from a lysate under conditions much closer to the protein’s native cellular environment, including native post-translational modifications and native complex partners the yeast nucleus can’t reproduce. It doesn’t scale to library screening, but it’s the stronger evidence for a specific, already-hypothesized interaction.

A common design pairs the two: use Y2H (or a related library-scale method) to generate candidate interactions, then confirm the strongest candidates by Co-IP or another orthogonal approach before publishing or building further work on them. For interactions found by either method that need proteome-wide context, see STRING for interaction-network databases, and mass spectrometry proteomics sample prep for unbiased interactor identification by MS rather than a single-antibody or single-prey approach.

Frequently asked questions

What’s the difference between yeast two-hybrid and Co-IP?

Y2H is a genetic, cell-based method that detects an interaction indirectly, through reconstituted transcription-factor activity and reporter-gene readout in living yeast; it is well suited to screening one bait against many candidate partners. Co-IP is a biochemical method that detects an interaction directly, by physically pulling a protein complex out of a lysate; it is well suited to confirming one specific, already-suspected interaction in a more native cellular context.

Can Y2H detect weak or transient interactions?

Often yes, and this is one of its practical advantages over lysate-based pulldowns: because the bait and prey are continuously co-expressed in living cells rather than sampled once in a lysate, Y2H can catch interactions that are too weak or too transient to survive a Co-IP’s lysis and wash steps. The tradeoff is that this same sensitivity contributes to its higher false-positive rate.

What causes false positives in a yeast two-hybrid screen?

The most common causes are a self-activating bait (tested for and excluded before screening), a ‘sticky’ prey that activates reporters with many unrelated baits, and interactions that only occur because both proteins are forced into unusually close proximity in the yeast nucleus at non-physiological concentrations. This is why Y2H hits are treated as candidates requiring orthogonal confirmation, not as standalone conclusions.

Can Y2H test membrane protein interactions?

Not with the standard GAL4-based format, because the assay requires bait and prey to enter the yeast nucleus, which most integral membrane proteins can’t do while remaining functional. Specialized variants (such as split-ubiquitin membrane Y2H systems) exist specifically to work around this, but they are a different assay design from the classical nuclear GAL4 system described here.

What is a Y2H library screen versus pairwise testing?

Pairwise testing checks one specific bait against one specific prey to confirm a hypothesized interaction. A library screen instead mates or transforms one bait against a whole cDNA/ORF prey library at once, so that every candidate interactor in the library gets tested in parallel — this is the large-scale interactome-discovery use case Y2H is best known for, and the main reason it remains in use alongside newer proteomics methods.