Utilizing the Split-Ubiquitin Membrane Yeast Two-Hybrid System to Identify Protein-Protein Interactions of Integral Membrane Proteins

See allHide authors and affiliations

Science's STKE  15 Mar 2005:
Vol. 2005, Issue 275, pp. pl3
DOI: 10.1126/stke.2752005pl3


Various modifications of the conventional yeast two-hybrid system have played an essential role in confirming or detecting protein-protein interactions among nuclear and cytoplasmic proteins. These approaches have permitted the identification of novel interaction partners, as well as provided hints as to their function. However, membrane proteins, such as receptor tyrosine kinases, G protein–coupled receptors, membrane-bound phosphatases, and transporters, which represent important classes of signaling molecules, are difficult to study using classical protein interaction assays because of their hydrophobic nature. Here, we describe a genetic system that allows the identification of integral membrane-interacting proteins. This so-called "split-ubiquitin membrane-based yeast two-hybrid assay" involves fusing the halves of ubiquitin to two interacting proteins, at least one of which is membrane bound. Upon interaction of these two proteins, the halves of ubiquitin are brought together, and the transcription factor that is fused to a membrane protein of interest is cleaved and released. The free transcription factor then enters the nucleus and activates transcription of reporter genes. We also describe how this technology is used to screen complementary DNA libraries to identify novel binding partners of a membrane protein of interest.


As numerous model genomes have been sequenced, the elucidation of protein function is the next challenge toward the understanding of biological processes. One way to determine a protein's function is to identify its interacting partners, because proteins often work in pairs or as part of large complexes. Traditionally, researchers have used biophysical or biochemical methods, such as surface plasmon resonance, affinity chromatography, or co-immunoprecipitation, to study protein-protein interactions. More recently, yeast two-hybrid (YTH) and phage-display systems have dominated the field with clear advantages, the most attractive of which is the immediate accessibility to the protein's genetic information (1). However, because of their highly hydrophobic nature, integral membrane proteins—one of the largest classes of proteins in any cell—remain the most problematic proteins to analyze with any of the above-mentioned protein interaction methods. Together, two systematic studies of protein-protein interactions in the yeast Saccharomyces cerevisiae, based on YTH analysis (2, 3) and affinity purification linked to mass spectrometry (4, 5), have clearly shown that interactions between integral membrane proteins are underrepresented.

During recent years, several interactive proteomics techniques, such as fluorescence resonance energy transfer (FRET) (6), β-lactamase protein fragment complementation assay (7), and β-galactosidase complementation assay (8), have allowed the monitoring of membrane protein interactions in real time. However, none of these techniques has thus far been used to identify membrane protein interactions in a screening procedure.

In this article, we describe the split-ubiquitin membrane yeast two-hybrid system (MbYTH) as an easy, efficient, and sensitive genetic method to monitor and detect membrane protein interactions in vivo (9-11).

The MbYTH was designed to overcome the limitation of the conventional YTH, which is based on reconstitution of transcription factor (TF) activity that must occur in the nucleus (12). The MbYTH is based on the observation that ubiquitin can be experimentally separated into two moieties that functionally reconstitute when present in close proximity to one another (13). Reconstitution of split ubiquitin has also been exploited as a method to study the interactions of membrane proteins fused to the ubiquitin moieties (9). Ubiquitin is an evolutionarily conserved 76–amino-acid protein that serves as a tag for proteins targeted for degradation by the 26S proteasome. The presence of ubiquitin is recognized by ubiquitin-specific proteases (UBPs) located in the nucleus and cytoplasm of all eukaryotic cells. In the MbYTH, a membrane protein of interest, the so-called "bait," is fused to the C-terminal half of ubiquitin (Cub), along with an artificial TF that consists of the bacterial LexA-DNA binding domain and the Herpes simplex VP16 transactivator protein. The putative interacting proteins, called the "prey," are either membrane-bound or cytosolic, and are fused to the N-terminal half of ubiquitin (Nub). Because of their high affinity, the two halves of ubiquitin spontaneously reconstitute and are recognized by the UBPs. Hence, to prevent spontaneous association, an isoleucine (I) to glycine (G) exchange at position 13 of the Nub moiety has been introduced (NubG) (13). Only upon interaction of the two proteins does the reconstitution of ubiquitin (Cub + NubG) occur; this so-called split ubiquitin is then recognized by abundant UBPs, resulting in the cleavage of the TF. The released TF then enters the nucleus and activates transcription of the reporter genes, which can be monitored by growth on selective plates or by colorimetric assays. The reporter strains used in our system have the HIS3, lacZ, and/or ADE2 reporter genes (Fig. 1).

Fig. 1.

Outline of the membrane YTH system. (A) A membrane bait protein of interest is fused to Cub followed by the artificial transcription factor LexA-VP16 (blue), while another prey membrane (or cytoplasmic) protein is fused to the NubG domain (red). If the bait and the prey do not interact, there is no reconstitution of ubiquitin and no UBP-mediated cleavage of the transcription factor occurs; this results in HIS3/ADE2 and LacZ yeast. (B) On interaction of the bait and prey proteins, ubiquitin reconstitution occurs, leading to proteolytic cleavage by UBPs and the subsequent release of the transcription factor. This factor enters the nucleus and activates reporter genes by binding to the Lex A operator sites (lexAops) within their promoters. This results in HIS3+/ADE2+ and lacZ+ yeast cells (in case of a three-reporter strain, for example, THY.AP4).

Here, we detail the methodology to detect membrane protein interactions in vivo using the MbYTH, which is well suited for both small-scale and library-screening approaches. The Protocol provides comprehensive instructions for the construction of bait and prey gene fusion plasmids, followed by steps involved in screening with prey cDNA libraries and confirmation of putative interactors (Fig. 2).

Fig. 2.

A typical flow chart of a MbYTH prey cDNA library screen using bait protein of interest.


0.2-ml PCR tubes [Axygen]

1.5-ml tubes [Greiner Bio-One]

50-, 15-, and 12-ml polypropylene screw-cap tubes (Greiner Bio-One)

145- and 94-mm diameter plates (Greiner Bio-One)

Bacto agar (Difco)

Bacto peptone (Difco)

Bacto yeast extract (Difco)

Cuvettes, 10 mm × 10 mm × 45 mm, plastic (Greiner Bio-One)

Filter (syringe), 0.2 μm cellulose acetate [Schleicher & Schuell]

Glass beads (diameter 0.4 to 0.6 mm) (Sigma-Aldrich)

Inoculation loops (Greiner Bio-One)

NitroBind pure nitrocellulose 0.22-μ membrane (Osmonics, #EP2HY00010)


NucleoSpin plasmid prep kit (Macherey-Nagel)

QIAquick Gel Extraction Kit (Qiagen)

X-ray film (Typon Imaging, Burgdorf, Switzerland)

Yeast nitrogen base without amino acids (Difco)




37.5:1 Acrylamide/bis-acrylamide [Interchim]

5-Bromo-4-chloro-3-indolyl-β-𝒹-galactoside (X-Gal) [Alexis]

Adenine (Sigma-Aldrich)

Agar powder

Agarose, DNA electrophoresis grade

Amino acids (Sigma-Aldrich)


Bio-Rad Protein Assay Solution

Bromophenol blue

Deoxyribonucleoside 5′-triphosphates set, 100 mM solutions (dATP, dCTP, dGTP, dTTP) [Amersham, #27-2035-01]

1,4-dithio-DL-threitol (DTT)

Drop-out mix

Note: Drop-out mix is available from multiple commercial sources (for example, Qbiogene or Clontech).

DyNAzyme EXT DNA polymerase (Finnzymes, Espoo, Finland)

Ethanol (70% and 100%)

Ethidium bromide

Ethylenediamine tetraacetic acid disodium salt (EDTA)




Hydrochloric acid, concentrated (11.6 N)



Lithium acetate (LiOAc)

N,N-dimethylformamide (NNDMF)

N,N,N′,N′-tetramethylethylenediamine (TEMED)


Mg2+-free PCR buffer (Sigma-Aldrich)

Milk powder

Polyethylene glycol 4000 (PEG 4000)

Protease inhibitors [Complete Protease Inhibitor Cocktail Tablets, Roche Diagnostics]

Single-stranded herring sperm DNA (BD Biosciences)


Sodium chloride

Sodium dodecyl sulfate (SDS)

Sulfosalicylic acid

Taq DNA polymerase (Sigma-Aldrich)

Triton X-100

Tris base


Uptilight HRP blotting chemiluminescent substrate (Interchim, #UP99619A)

Xylene cyanole FF

Yeast lytic enzyme (lyticase) [ICN Biomedicals]


Goat secondary antibody to mouse, horseradish peroxidase-conjugated (Sigma-Aldrich)

Goat secondary antibody to rabbit, horseradish peroxidase-conjugated (Sigma-Aldrich)

Monoclonal mouse antibody to hemagglutinin (HA) (Covance)

Polyclonal mouse antibody to LexA (Santa Cruz Biotechnology)

Polyclonal rabbit antibody to VP16 (Sigma-Aldrich)

Bacteria: E. coli

DH5α [F-,ϕ80dlacZΔM15, endA1, recA1, hsdR17(rk,mk+), supE44, thi1, gyrA96,relA1, Δ(lacZYAargF)U169, λ-]

Note: This strain of E. coli is available from several commercial suppliers (for example, Invitrogen).


Oligonucleotide primers for amplification: 5′ 60-mer and 3′ 60-mer

Note: For details on the design of these PCR primers, which will be specific to a given gene of interest, see "Primer design for type I transmembrane proteins" and Figure 3 .

Fig. 3.

Sequences of the bait and prey primers.

Oligonucleotide primers for verification: 5′ 20-mer and 3′ 20-mer

Note: These primers are used for the verification of proper recombination events by colony PCR and for sequencing. They are designed such that they are specific to the vector sequence about 40 bp upstream or downstream of the ORF of interest, respectively.


pCCW Ste2


pDL2 xN Ste2

Note: Plasmids are available from Dualsystems Biotech,

Yeast: Saccharomyces cerevisiae

L40 [MATa HIS3Δ200 trp1-901 leu2-3, 112 ade2LYS2:(lexAop)4-HIS3URA3::(lexAop)8-lacZ GAL4] (15)

THY.AP4 [MATa leu2-3,112 ura3-52 trp1-289 lexA::HIS3 lexA::ADE2 lexA::lacZ] (16)

NMY32 [MATa his3delta200 trp1-901 leu2-3,112 ade2 LYS2::(lexAop)4-HIS3 URA3::(lexAop)8-lacZ (lexAop)8-ADE2 GAL4)] [Dualsystems Biotech]


5417R microcentrifuge (Eppendorf)

Easyject electroporator (Equibio, Kent, UK)

GeneAmp PCR system 2700 (Applied Biosystems)

Genesys 10 UV scanning spectrophotometer (Thermo Spectronic)

Heat block (Techne)

Incubator (Heraeus)

Lab shaker (Heidolph)

Labtherm shaker (Kühner)

MiniSpin centrifuge 1.5 ml (Eppendorf)

Optimax x-ray film processor (Typon Imaging)

Power supply (Bio-Rad)

Robocycler Gradient 96 (Stratagene)

Rotanta 460RS tabletop centrifuge (Hettich)

Sorvall Centrifuge RC5C

Vortexer (Scientific Industries)

White light/UV transilluminator (UVP)


Recipe 1: 50-μl PCR Reactions
Reagent Amount
Plasmid (5 pg/μl)2 μl
10 mM 5′ 60-mer2 μl
10 mM 3′ 60-mer2 μl
10 mM dNTP mix1 μl
10× DyNAzyme EXT PCR buffer5 μl
1 U/μl DyNAzyme EXT DNA polymerase1 μl
ddH2O37 μl
Mix in a microcentrifuge tube and store on ice until use.
Recipe 2: 10× Loading Buffer for Agarose Gels
Reagent Amount
Bromophenol blue25 mg
Xylene cyanole FF25 mg
87% Glycerol3.4 ml
ddH2O6.6 ml
Mix components and store at room temperature.
Recipe 3: YPAD (Plates and Liquid Medium)
Reagent Amount
Bacto yeast extract10 g
Bacto peptone20 g
Glucose20 g
Adenine40 mg
Add 1000 ml of ddH2O and dissolve the medium in a bottle with a magnetic stir bar. Autoclave at 121°C for 15 min.
For plates, add 20 g of Bacto agar to the bottle of medium and autoclave at 121°C for 15 min. Pour the plates while the agar is still warm (~50°C).
Note: 2× YPAD medium used in the library transformation protocols is simply YPAD medium with double amounts of all components, with the exception of adenine, which is added in the same amount as in 1× YPAD.
Recipe 4: 1 M Tris-HCl Buffers
Dissolve 121.1 g of Tris base in 800 ml of ddH2O, with heating if necessary. Adjust the pH to 7.5 or 8.0 (as desired) with concentrated HCl. Add sterile ddH2O to a final volume of 1000 ml, mix, autoclave, and store at room temperature.
Recipe 5: 0.5 M EDTA
Dissolve 186.1 g of EDTA in 800 ml of ddH2O and adjust the pH to 8.0 with NaOH pellets (about 20 g). Adjust volume to 1000 ml with ddH2O, sterilize by autoclaving, and store at room temperature.
Note: The EDTA will not dissolve until the pH approaches 8.0.
Recipe 6: 1× TE, pH 8.0
Reagent Amount
1 M Tris-HCl, pH 8.0 (Recipe 4)2 ml
0.5 M EDTA, pH 8.0 (Recipe 5)0.4 ml
Add sterile ddH2O to a final volume of 200 ml, mix, autoclave, and store at room temperature.
Recipe 7: 1 M LiOAc
Dissolve 20.4 g of LiOAc in 200 ml of ddH2O. Adjust pH to 7.5 with 5% acetic acid. Sterilize by autoclaving and store at room temperature.
Recipe 8: 50% PEG 4000
Dissolve 50 g of PEG 4000 in ddH2O to a final volume of 100 ml. Mix, autoclave, and store at room temperature.
Recipe 9: PEG-LiOAC Master Mix
Reagent Amount
50% PEG 4000 (Recipe 8)240 μl
1 M LiOAC (Recipe 7)36 μl
Herring sperm single-stranded DNA (10 mg/ml)5 μl
These volumes are for one yeast transformation. This mix must be prepared immediately before use.
Recipe 10: Yeast Synthetic Drop-Out Plates and Liquid Medium [SD−(Amino Acid)]
Reagent Amount
Drop-out mix1 g
Bacto yeast nitrogen base6.7 g
Glucose20 g
Add 1000 ml of ddH2O and dissolve the components in a bottle with a magnetic stir bar. Autoclave at 121°C for 15 min.
For plates, add 20 g of Bacto Agar to the bottle of medium and autoclave at 121°C for 15 min. Pour the plates while the agar is still warm (~50°C).
Note: Drop-out mix is stored at 4°C. For use with the MbYTH, you will need several drop-outs, as suggested in the Protocol.
Recipe 11: Yeast Colony PCR Lysis Buffer
Reagent Amount
0.1 M NaPO4 buffer, pH 7.41 ml
4800 U/g Lyticase11 mg
Mix these two reagents. Use freshly prepared.
Recipe 12: 50-μl Yeast Colony PCR Reactions
Reagent Amount
10 μM 5′ 20-mer0.7 μl
10 μM 3′ 20-mer0.7 μl
10× Mg2+-free PCR buffer5 μl
25 mM MgCl23 μl
mM dNTP mix0.5 μl
5 U/μl Taq polymerase1 μl
ddH2O7.3 μl
Mix in a microcentrifuge tube and store on ice until use. This volume assumes use of 3 μl of template DNA.
Recipe 13: Yeast Lysis Buffer
Reagent Amount
2 M sorbitol136 μl
0.5 M EDTA (Recipe 5)60 μl
1 M DTT15 μl
4800 U/g Lyticase42 mg
Add sterile ddH2O to a final volume of 300 μl, mix. Use freshly prepared.
Recipe 14: LB (Plates and Liquid Medium)
Reagent Amount
Bacto tryptone10 g
Bacto yeast extract10 g
NaCl5 g
Add 1000 ml of ddH2O and dissolve the medium in a bottle with a magnetic stir bar. Autoclave at 121°C for 15 min.
For plates, add 20 g of Bacto agar to the bottle of medium and autoclave at 121°C for 15 min. Pour the plates while the agar is still warm (~50°C). Where necessary, add the appropriate antibiotics.
Recipe 15: Protein Extract Lysis Buffer
Dissolve 1 tablet of the Complete Protease Inhibitor Cocktail in 7 ml of 50 mM Tris-HCl, pH 7.5.
Recipe 16: 10× Loading Buffer for SDS-PAGE
Reagent Amount
1M Tris-HCl pH 6.812 ml
SDS4 g
87% Glycerol4.7 ml
Bromophenol blue10 mg
2-ΜΕ4 ml
25 mM DTT0.077 ml
Adjust with ddH2O to a final volume of 20 ml. Store at -20°C.
Recipe 17: 10× PBS
Reagent Amount
NaCl80 g
KCl2 g
Na2HPO4 14.4 g
KH2PO4 2.4 g
Dissolve in 800 ml of ddH2O, adjust pH to 7.4 with HCl, and adjust volume to 1 liter with ddH2O.
Recipe 18: Agarose X-Gal Mix
The final solution contains 1× PBS, 0.5% (w/v) agarose, and 0.1 mg/ml X-Gal.
Note: X-Gal is previously dissolved in NNDMF at a concentration of 10 mg/ml in 1/100 of the final volume. This X-Gal mix should always be freshly prepared.
To make the solution, add agarose to 10× PBS (Recipe 17), adjust to final volume with ddH2O, and heat in a microwave with occasional swirling until the agarose has dissolved. Incubate in a cooling water bath until the container has cooled to the touch. Add X-Gal, mix well, and use immediately.
Note: Cooling is very important, because adding the heat-labile X-Gal substrate solution when the mix is still too hot will degrade X-Gal, producing a false negative result in the X-Gal filter assay.
Recipe 19: 15% Glycerol-YPAD
Mix 15 ml of 100% glycerol with 85 ml of YPAD (Recipe 3). Autoclave and store at room temperature.


Construction of Cub-TF and NubG Fusion Plasmids

Membrane proteins are classified in many ways, one of which is based on the orientation of their N and C termini. A typical type I integral membrane protein presents its N terminus in the lumen, or extracellular side, and its C terminus in the cytosol; the reverse is true for the type II class of proteins, irrespective of the number of times they span the membrane. In principle, the MbYTH system can be applied to both types of integral membrane proteins; however, our focus has been on type I membrane proteins, and the detailed methods described here apply to their analysis.

At the beginning of an MbYTH study, the cDNA sequence encoding a bait protein of interest is cloned into the vector such that it is in frame with the Cub-TF cassette placed either upstream or downstream ("bait vector") (Fig. 4). The pCYC-Ste2L-bait-Cub-TF (pCCW-Ste) vector allows expression of proteins with their C termini attached to the Cub-TF (or Cub-LexA-VP16) and facilitates cloning of type I transmembrane proteins. It is also possible to use a pCYC-TF-Cub-bait (pNCW) vector, which results in the TF-Cub preceding the N terminus of the polypeptide and facilitates cloning of type II transmembrane proteins.

Fig. 4.

Bait and prey vectors for application of type I and type II integral membrane proteins in the MbYTH. Two sets of bait and prey vectors allow different orientations of the reporter cassette relative to bait and prey. (A) A bait protein with its N terminus in the lumen and its C terminus in the cytosol (type I transmembrane protein) is fused C-terminally to the Cub-TF reporter cassette in the vector pCCW. A prey protein with its N terminus in the cytosol is fused N terminally to the NubG cassette of pDSL-Nx. (B) A bait protein with its N terminus in the cytosol (type II transmembrane protein) is fused N terminally to the TF-Cub cassette of pNCW. A cytosolic prey protein is fused C terminally to the NubG cassette in pDL2xN. The two sets of bait and prey vectors allow productive fusions with almost any cytosolic or integral membrane protein, the only prerequisite being that the Cub-TF and NubG moieties are located on the cytosolic face of the membrane.

Both bait vectors contain the weak CYC1 promoter, which drives low levels of heterologous protein expression. Furthermore, the bait vector for type I transmembrane proteins contains, at the N terminus, the yeast Ste2 leader sequence, resulting in an improved targeting of the heterologous bait protein to the yeast membrane. Both bait vectors are centromeric plasmids containing an autonomously replicating sequence (ARS) origin of replication and one centromeric locus (CEN), which results in one to two copies of the plasmid per cell. These low-copy-number plasmids autonomously replicate in both E. coli and S. cerevisiae and contain the KanR and LEU2 genes for selection of plasmid-bearing cells on medium containing kanamycin (bacteria) or lacking leucine (yeast), respectively.

The pADH-Ste2L-prey-NubG (pDL2-xN) and pCYC1-NubG-prey vectors (pDSL-Nx) ("prey vectors") (Fig. 4) allow cloning of a library of prey genes or specific genes of interest in frame with NubG at either the N or the C terminus of the protein, respectively. Both prey vectors express the fusion protein from the strong constitutive promoter ADH1 and are replicated by the high-copy-number 2μ origin of replication. They are selected for by the AmpR and TRP1 genes, allowing growth on medium containing ampicillin (bacteria) or lacking tryptophan (yeast), respectively. The use of the AmpR marker for E. coli facilitates the reisolation of these vectors after screening with respect to the bait vector, which has a KanR marker.

Cloning of the desired cDNA is performed by conventional restriction and ligation strategies, in vitro recombination approaches (for example, GATEWAY by Invitrogen), or direct in vivo recombination in yeast. Details for direct in vivo recombination in yeast will be described, because it is most commonly used in our laboratory. In this system, a sequence-verified bait construct is transformed into any of the yeast reporter strains suitable for the MbYTH.

The cDNA encoding the protein of interest is amplified by PCR with 60-nucleotide (nt) primers having 5′ ends that are identical to the vector, to allow cloning into the vector by in vivo recombination in yeast. A one-step PCR protocol is used to create a cDNA extended by homology regions at its 5′ and 3′ ends. These 38- to 40-nt homology regions are DNA sequences shared by the vector, whereas about 20 nt are gene specific. The bait vector can be digested with either PstI or HindIII, and the prey vector (pDL2-xN) can be digested with SmaI. Cotransformation of the PCR product and the "empty" linearized (digested) plasmid into yeast results in homologous recombination and gap repair, yielding the "bait construct" or "prey construct." Figures 5 and 6 illustrate the process from PCR amplification through homologous recombination.

Fig. 5.

Schematic representation of in vivo recombination strategy for the construction of the bait vector. Amplification of the ORF encoding the protein of interest by PCR using 60-mer primers (40 nt specific to the vector sequence and 20 nt specific to the ORF) is followed by integration through in vivo recombination in the suitable yeast reporter strain.

Fig. 6.

Schematic representation of in vivo recombination strategy for the construction of the prey vector. Amplification of the ORF encoding the protein of interest by PCR using 60-mer primers (40 nt specific to the vector sequence and 20 nt specific to the ORF) is followed by integration through in vivo recombination in the suitable yeast reporter strain.

For type II transmembrane bait proteins, the fusion of the corresponding cDNA has to be performed C-terminally to the TF-Cub-encoding moiety of the pNCW vector (Fig. 4), resulting in the "TF-Cub-bait chimeric protein." To that end, the corresponding cDNA is PCR amplified and cloned into pNCW using one or two of the following unique vector restriction sites: PstI, NcoI, or SacII (Fig. 4). However, a protocol for this protein class is not included, because our laboratory's focus has so far been on type I transmembrane proteins.

Primer design for type I transmembrane proteins

Forward primers should contain about 40 nt identical to the sequence upstream of the chosen restriction site within the vector, followed by 18 to 20 gene-specific nt. To maintain the Ste2 leader sequence, the forward primer includes the corresponding DNA sequence, which encodes the first 13 amino acids of Ste2p; thus, the 5′ primers for the "bait vector" (Cub-TF fusion) and the "prey vector" (NubG fusion) are identical, because both vectors contain the Ste2 leader sequence (Fig. 3).

For the reverse primer, it is essential to omit the native stop codon of the bait or prey sequence. Furthermore, the fusion between the protein-encoding sequence and the 40-nt homology region should be designed so that the reading frame of the cDNA of the bait or prey is in frame with the Cub-TF sequence (Fig. 3).

PCR amplification of bait or prey fusion protein-encoding cDNAs

1. Prepare 50-μl PCR Reactions (Recipe 1) for each bait and prey cDNA

2. Amplify the cDNA sequences by PCR using a proofreading DNA polymerase (such as DyNAzyme EXT) and cycling parameters specific to the open reading frame (ORF) of interest (elongation time) and the sequence of the primers (annealing temperature).

3. Purify the PCR products by gel elution using a commercially available gel extraction kit (such as QIAquick Gel Extraction Kit) according to the manufacturer's instructions.

Note: 0.5 μg of the purified PCR product is required for homologous recombination. Purified PCR products can be stored at −20°C.

4. In separate reactions, digest about 1 μg of the bait and prey vectors with 3 to 5 U of the appropriate restriction enzyme in a volume of 25 μl for 2 to 3 hours at the recommended temperature for the enzyme.

6. Combine 1 μl of 10× Loading Buffer for Agarose Gels (Recipe 2) with 6 μl of ddH2O, and add 2 μl of the digested vector.

7. Size-fractionate the vector on an agarose gel to verify that it is completely digested.

8. Purify the linearized vector by gel elution as in step 3, above.

Note: Purified linearized vector can be stored at −20°C.

9. Inoculate a single colony of the appropriate yeast strain (L40 or THY.AP4 or NMY32) into 5 ml of liquid YPAD (Recipe 3) and grow overnight at 30°C with shaking.

10. The next morning, measure the OD600 of overnight culture and dilute the culture in liquid YPAD (Recipe 3) to a final OD600 of 0.3.

Note: Each transformation will require 2 ml of yeast culture, so for three transformations, prepare 6 ml of YPAD.

11. Grow the yeast at 30°C with shaking to an OD600 of about 0.8 (3 to 4 hours).

12. Pellet the yeast for 5 min at 700g, then resuspend and wash the pellet in 10 ml of sterile 1× TE, pH 8.0 (Recipe 6).

13. Pellet the yeast for 5 min at 700g and resuspend in 1/20 volume of sterile ddH2O (for example, each transformation utilizes a 2-ml culture, which would be resuspended in 100 μl of ddH2O).

14. For each insert and vector combination, set up each of the following transformation reactions in a microcentrifuge tube:

Transformation A: 500 ng of PCR product + 100 ng of digested empty bait vector (or prey vector) for in vivo recombination.

Transformation B: 100 ng of digested empty bait vector (or prey vector) serves as a negative control.

Transformation C: Undigested empty bait vector (or prey vector) serves as a control for the transformation efficiency.

15. Add 300 μl of PEG-LiOAC Master Mix (Recipe 9) to each transformation reaction and vortex briefly.

16. Add 50 μl of the yeast cells from Step 13 to each transformation reaction and vortex for 1 min to thoroughly mix all components.

17. Incubate each transformation reaction in a 42°C water bath for 15 min.

18. Pellet the transformed yeast for 5 min at 700g at room temperature.

19. Resuspend the pellet in 100 μl of ddH2O and plate each transformation onto one SD−Leu plate (for the bait transformants) or SD−Trp plate (for the prey transformants) (Recipe 10).

20. Incubate for 2 to 3 days at 30°C.

Note: The number of yeast colonies on the transformation control plate (transformation C) may be between 50 and several hundred colonies, depending on transformation efficiency. Transformation A should have substantially more colonies than transformation B, which represents background generated by gap repair of the vector or undigested vector.

21. Pick eight isolated colonies from each plate containing putative recombinant vector (transformation A) and restreak them onto a single fresh SD−Leu plate (for the bait transformants) or SD−Trp plate (for the prey transformants) (Recipe 10).

22. Incubate for 2 days at 30°C.

23 Suspend a loopful of each of the eight yeast colonies in microcentrifuge tubes, each containing 12 μl of Yeast Colony PCR Lysis Buffer (Recipe 11).

24. Incubate the suspensions at 37°C for 45 min, then at 95°C for 5 min.

25. Dilute this lysate with 12 μl of ddH2O.

26. Confirm insertion of the bait or prey construct in the eight restreaked colonies in 50-μl Yeast Colony PCR Reactions (Recipe 12) (17) using appropriate primers for the bait or prey vector (Fig. 3) and the following PCR conditions:

95°C 1 min

35 cycles of:

95°C 15 s

50°C 20 s

70°C 2 min

Last cycle:

70°C 5 min

4°C hold

27. After analyzing the PCR reaction, inoculate 5 ml of liquid SD−Leu (for the bait transformants) or SD−Trp (for the prey transformants) (Recipe 10) with one positive colony from each of the eight (from step 21) that were restreaked, and grow overnight at 30°C.

28. Proceed with plasmid rescue from the yeast using standard methods (18).

29. Prepare DNA using any common mini-prep method for yeast [for example, using Yeast Lysis Buffer (Recipe 13)] and transform this plasmid DNA into an E. coli strain using KanR or AmpR selection, for bait or prey, respectively.

30. Grow transformed E. coli on LB Plates (Recipe 14) containing 30 μg/ml kanamycin for bait plasmid or 100 μg/ml ampicillin for prey plasmid, by standard methods overnight at 37°C.

31. Inoculate single transformants for growth in selective LB Liquid Medium (Recipe 14) at a scale appropriate for DNA mini-preps.

32. Isolate the bait or prey plasmid using any common DNA mini-prep method for E. coli.

33. Check the E. coli mini-prep DNA by restriction digestion and by sequencing with appropriate primers to confirm the presence and sequence of the insert.

Verification of Bait Proteins Expression and Membrane Insertion in Yeast

Testing the expression of the bait protein by Western blot analysis is recommended before attempting interaction tests. We recommend first preparing total protein extracts from the bait-bearing strains and, in a second step, preparing membrane extracts to check if the bait protein is properly inserted into the yeast membrane.

We always check the expression of bait proteins; however, it is tedious to do so for prey proteins in the case of a prey library transformation. A Western blot analysis can be performed to examine the expression of a single known or putative interacting prey protein according to the following protocols.

Preparation of total extracts

1. Inoculate a single colony from each bait-bearing yeast strain from the plate in Step 20 of "PCR amplification of bait or prey fusion protein encoding cDNAs" into 10 ml of SD−Leu (Recipe 10) and incubate overnight with shaking at 30°C.

Note: Cells should be grown to an OD600 of 1.0.

2. Centrifuge at 700g for 5 min at room temperature in a 15-ml tube and decant the supernatant.

3. Add 200 μl of Protein Extract Lysis Buffer (Recipe 15).

4. Add 100 to 200 μl of glass beads (size range 425 to 600 μm, unwashed).

5. Vortex vigorously for 30 s, cooling for 1 min on ice between vortex steps. Repeat vortexing and cooling seven times.

6. Centrifuge the extract at 700g for 20 min at 4°C.

7. Transfer the supernatant to a fresh microcentrifuge tube.

8. Determine protein concentration by Bradford assay.

Note: If necessary, extracts may be stored overnight at 20°C before electrophoresis.

9. Boil extracts containing 30 to 50 μg of total protein in 10× SDS-PAGE Loading Buffer (Recipe 16) at a 1× dilution.

10. Load extracts onto a SDS-PAGE gel and perform a Western blot using any standard method.

Note: Suitable antibodies for detection of the bait fusion protein are protein-specific antibodies, mouse monoclonal antibodies directed against LexA, or a polyclonal rabbit antibody directed against VP16 (Fig. 7).

Fig. 7.

A Western blot analysis showing the membrane, cytosolic, and crude fractions of a yeast total protein extract probed with an antibody specific to the transcription factor moiety of the bait fusion protein. As a control, the lower part of the same blot was probed with an antibody specific to yeast ER membrane protein, Wbp1p, demonstrating a clear separation between the membrane and cytosolic fractions.

Preparation of membrane extracts

1. Inoculate a single colony from each bait-bearing yeast strain from the plate in Step 21 of "PCR amplification of bait or prey fusion protein-encoding cDNAs" into 50 ml SD−Leu (Recipe 10) and incubate overnight with shaking at 30°C.

Note: Cells should be grown to an OD600 of 1.0.

2. Centrifuge at 700g for 5 min at 4°C in a 50-ml tube.

3. Discard the supernatant, add 1 ml of cold Protein Extract Lysis Buffer (Recipe 15), and resuspend by vortexing.

4. Transfer the cells into a 2-ml microcentrifuge tube and fill the tube with 300 μl of glass beads (size range 425 to 600 μm, unwashed).

5. Vortex vigorously for 30 s, cooling for 1 min on ice between vortex steps. Repeat vortexing and cooling seven times.

6. Centrifuge the extract at 700g for 20 min at 4°C.

7. Carefully transfer the supernatant fraction into a fresh microcentrifuge tube and spin down again at 700g for 10 min at 4°C.

8. Transfer the supernatant into microcentrifuge tubes and spin down at 150,000g in an ultracentrifuge for 2 hours at 4°C.

9. Collect the supernatant (cystolic fraction) and add an equal volume of Protein Extract Lysis Buffer (Recipe 15).

10. Resuspend the pellet in 200 μl of Protein Extract Lysis Buffer (Recipe 15) with 1% Triton X-100 and centrifuge at 150,000g in an ultracentrifuge for 2 hours at 4°C.

11. Collect the supernatant (membrane fraction) and add an equal volume of Protein Extract Lysis Buffer (Recipe 15).

12. Freeze extracts in liquid nitrogen and store at −80°C.

13. Boil extracts in a volume corresponding to about 50 μg of protein in 10× SDS-PAGE Loading Buffer (Recipe 16) at a 1× dilution.

14. Load extracts onto a SDS-PAGE gel and perform a Western blot using any standard method.

Note: Suitable antibodies for detection of the bait fusion protein are specific antibodies for the fused protein, mouse monoclonal antibodies directed against LexA, or polyclonal rabbit antibodies directed against VP16 ( Fig. 7 ).

Verification of Bait Protein Expression and Function by the "NubG/NubI" Test

To test for correct expression and functionality of the bait protein in yeast, the bait plasmid of confirmed sequence is cotransformed with several control prey plasmids leading to the expression of the following membrane prey proteins: Alg5-NubG, Alg5-NubI, Fur4-NubG, Fur4-NubI, Ost1-NubG, and Ost1-NubI (Table 1). Because the proteins are expressed ectopically and tagged, the resulting proteins may not always be properly folded. However, if the expressed protein is positive in a NubG/NubI genetic test and is detectable in a Western blot analysis, we assume that the protein is correctly folded and functional. OST1 and ALG5 are resident proteins of the yeast endoplasmic reticulum (ER), and FUR4 is a yeast plasma membrane protein. OST1, ALG5, and FUR4 are used as appropriate controls in this system, because all of the proteins being transported through the secretory pathway should interact with the NubI fusions of the three proteins. In proteins that do not pass through the secretory pathway, suitable controls from the respective compartments should be chosen, such as mitochondrial, vacuolar, or peroxisomal proteins.

Table 1.

NubI/NubG coexpression combinations. If all four negative controls yield a positive result, the bait protein may self-activate the transcriptional reporter without requiring ubiquitin-mediated cleavage.

NubI is the wild-type N-terminal portion of ubiquitin that will interact with any neighboring Cub fusion protein (see Introduction). Therefore, coexpression of one of the three different NubI-bearing plasmids with the bait protein of interest fused to Cub-TF will result in the reconstitution of split ubiquitin if the Cub-TF bait protein is expressed in the ER or the plasma membrane. Therefore, cellular localization of the bait-Cub-TF fusion can be monitored by coexpression of a noninteracting membrane protein fused to NubI (FUR4, OST1, or ALG5). If cleavage of the TF occurs, then the localization of both the two ubiquitin halves and their corresponding protein termini must be cytosolic or nuclear, because these are the only compartments where the UBPs are present.

Transformants expressing the Cub-TF fusion protein (bait protein) together with three different NubI fusion proteins, but not the corresponding three NubG fusion proteins, should grow on SD−Trp, −Leu, −His within 3 days, and should turn blue in the X-Gal filter test. However, failure to grow or exhibit a positive X-Gal result does not always mean that the bait Cub-TF fusion protein is incorrectly located in the membrane. Because OST1 and ALG5 are resident proteins of the ER, and FUR4 is a plasma membrane protein, efficient export of the bait protein to the ER or to the plasma membrane, respectively, may result in insufficient time that the bait protein and the corresponding prey protein (FUR4, OST1, or ALG5) are located in the same membrane compartment. Thus, insufficient amounts of split ubiquitin may be formed, resulting in suboptimal levels of cleavage. If no colonies appear after 3 days, the plates should be incubated for another 1 to 2 days to allow slower growing colonies to form. Negative control plates (bait vector cotransformed with each of three NubG-fusion plasmids or empty prey vector) should be checked to confirm that growth is not due to self-activation of the bait protein, from either overexpression or nonspecific cleavage in the cytosolic portion. Very rarely, it could happen that the bait protein will interact with FUR4, OST1, or ALG5, producing a false positive.

Mistargeting of the bait protein, leading to bait instability and degradation, could result in a bait protein that self-activates the transcriptional reporter signal. If the bait is strongly self-activating (strong growth together with one or all NubG-fusion proteins, or with empty prey vector), the selective medium can be supplemented with 3-AT. This compound is a competitive inhibitor of the HIS3 gene product and, therefore, increases the threshold of selection. Normally, addition of 1 to 30 mM 3-AT in the selection medium is sufficient to inhibit growth due to self-activation.

To confirm the results, one should also check the yeast colonies by performing the X-Gal filter test (Fig. 8).

Fig. 8.

An example of a NubG/NubI test for a "bait protein" of interest using appropriate constructs as controls. The yeast reporter strain was first transformed with the bait plasmid and subsequently with the indicated prey plasmids. Growth of yeast cells expressing a bait protein with various Nub fusions was monitored on agar plates lacking tryptophan (left, SD−T), and tryptophan and histidine (center, SD−TH). Two independent colonies were restreaked on SD−Trp and SD−Trp, −His selective plates prior to assessment of β-galactosidase activity using the X-Gal filter test (right).

Transformation of bait-bearing yeast strain with NubI and NubG plasmids

1. Grow the bait-bearing yeast strain in 2 ml of SD−Leu medium (Recipe 10) per transformation at 30°C, shaking at 200 rpm, to an OD600 of 0.80.

2. Centrifuge 2 ml of this yeast culture per transformation at 700g for 5 min. Resuspend and wash the pellet in 10 ml of sterile 1× TE, pH 8.0 (Recipe 6).

3. Pellet the yeast for 5 min at 700g and resuspend in 100 µl of sterile ddH2O per transformation.

4. Add 1 to 2 μg of each prey construct or empty prey plasmid DNA to separate microcentrifuge tubes.

5. Prepare the PEG-LiOAc Master Mix (Recipe 9).

6. Add 300 μl of PEG-LiOAc Master Mix (Recipe 9) to each tube of DNA and vortex briefly.

7. Add 100 μl of bait-bearing yeast cells (from step 3) to each tube and vortex for 1 min to thoroughly mix all components.

8. Incubate in a 42°C water bath for 20 min.

9. Pellet yeast at 700g for 5 min at room temperature.

10. Dissolve each yeast pellet in 200 μl of ddH2O and plate 100 μl of each transformation onto one SD−Trp, −Leu and one SD−Trp, −Leu, −His plate (for L40 yeast strain) (Recipe 10).

11. Incubate plates for 3 days at 30°C.

12. Perform the X-Gal filter test (see below).

X-Gal filter test

The filter lift assay is one of the most commonly used methods to assay the activity of the β-galactosidase reporter. Yeast cells are first transferred to Whatman 3MM filter paper, lysed in a freeze-thaw cycle using liquid nitrogen, and then overlaid with an agarose mixture containing the β-galactosidase substrate X-Gal. Yeast expressing β-galactosidase convert X-Gal to an indigo-blue chromophore; thus, the yeast will be blue. This method can also be used to test the Cub-TF fusion protein (bait protein) for self-activation, as well as to screen for interactions.

1. Pick single colonies from the selection plates, streak them out on fresh SD-selective plates (Recipe 10), and incubate for 2 to 3 days at 30°C.

2. Place a piece of circular Whatman 3MM filter paper directly onto the agar plate with the yeast colonies and incubate for 10 min to allow the yeast to stick to the filter.

Note: The 3MM Whatman filter paper should be cut to fit the petri dish.

3. Using forceps, carefully remove the filter paper from the plate, and transfer the filter into liquid nitrogen for 2 min.

4. Put each filter into a petri dish, yeast side up, and thaw for 5 min at room temperature.

5. Overlay the filter with freshly prepared Agarose X-Gal Mix (Recipe 18).

6. Allow the agarose to polymerize and incubate at room temperature until a blue color develops.

Note: Depending on the level of activation of the lacZ gene, this may take between a few minutes to several hours. Weak interactions are often detectable only after overnight incubation.

Construction and Screening of Prey cDNA Libraries

In conventional genetic screening systems, such as the YTH system, a cDNA library of prey proteins is made from cDNAs expressed as fusions to the C terminus of a transactivator domain, such as the GAL4 activation domain or the VP16 transactivation domain. Creating a set of fusion proteins with an invariant N-terminal domain and a variable cDNA has the advantage that initiation always takes place within the optimal context of the ATG of the N-terminal activation domain. The fixed orientation of the activation domain and the C-terminal prey protein is not a problem in most cases where the fusion protein is soluble, although some proteins will presumably lose some or all of their biological activity due to the N-terminal fusion. Because the fusion between the activation domain and the cDNA cannot be precisely controlled, only one-third of all clones will have a cDNA in the same reading frame as the preceding activation domain. Although this is a common problem of cDNA expression libraries, it can be overcome by creating libraries of high complexity (> 2 × 106 independent clones).

In contrast to the classical YTH system, in which the interaction between bait and prey takes place in the nucleus, the MbYTH system detects interactions between integral membrane proteins at the membrane. Both the Cub-TF and NubG moieties must be located in the cytosol to allow reassociation and cleavage of the artificial transcription factor. Because different classes of integral membrane proteins differ in the lumenal (or extracellular) or cytosolic localization of their respective N and C termini, a single cDNA library expressing a fusion of NubG to an integral membrane protein may occasionally localize the NubG part in the extracellular side of the membrane, depending on the orientation of the protein in the membrane. For instance, receptor tyrosine kinases (RTKs) are type I integral membrane proteins with their N termini located outside the cell and C termini located in the cytosol. Therefore, a NubG-prey cDNA library in which NubG was fused to the N-terminal portion of the prey (a NubG-x library) would express a fusion protein with NubG located in the lumen and therefore inaccessible to the cytosolic Cub-TF moiety of the bait. Consequently, one would expect that interactions between bait and RTKs cannot be found when screening a NubG-prey cDNA library. Type II integral membrane proteins have their N termini in the cytosol and therefore correctly display a cytosolic NubG when expressed from a NubG-x cDNA library. Thus, to use the MbYTH system to screen a library, either the membrane topology of the target protein must be known so that the proper fusion protein library can be constructed (NubG-x or x-NubG) or both libraries must be screened (Fig. 4). There are many different methods for construction of cDNA libraries, and it would be beyond the scope of this Protocol to describe them all. Standard methods for constructing directional cDNA libraries starting from either total or polyA+ RNA have been described (19, 20). Alternatively, one may use commercially available kits that already contain all necessary components [for example, SMART cDNA library construction kit (BD Biosciences) or CloneMiner cDNA library construction kit (Invitrogen)]. Below, we will describe some special points that need to be considered when constructing cDNA libraries for the MbYTH system.

When constructing expression cDNA libraries, it is important to ensure that the cDNAs are cloned directionally into the expression vector. Failure to do so will result in low-complexity libraries expressing sequences that do not correspond to ORFs. Conventionally, directional cloning is done either by using two different restriction enzymes for the 5′ and 3′ adaptors during first and second strand synthesis, or by using restriction enzymes that produce unequal overhangs, such as SfiI.

The cDNA libraries for the MbYTH system optimally contain a high fraction of full-length cDNAs to ensure that complete integral membrane proteins capable of inserting into the membrane in their native orientation are expressed. In practice, constructing full-length cDNA libraries has proved challenging. Nevertheless, several promising approaches to construct full-length libraries have been published (21, 22), and there is hope that, as methods evolve, cDNA libraries will encode more and more full-length proteins.

cDNA libraries for the MbYTH system are available from Dualsystems Biotech (

NubG-x cDNA libraries

NubG-x libraries fuse the NubG moiety to the 5′ end of the cDNA (Fig. 4). Such libraries can be constructed either by oligo-dT-primed first strand synthesis starting from total RNA or by random-primed first strand synthesis starting from polyA+ RNA. When using oligo-dT primers for the first strand synthesis, one ensures that a certain fraction of all cDNAs will be full length. When using random hexamers, one may increase the diversity of fragments represented in the library at the cost of average insert size, because randomly primed libraries tend to have smaller cDNAs than oligo-dT-primed libraries. To maximize the chance that an integral membrane protein is correctly targeted to and integrated into the membrane, a major part (or optimally the entire coding sequence) should be expressed from a given prey clone. For this reason, it may be best to rely on oligo-dT-primed libraries.

Like any expression cDNA library, NubG-x cDNA libraries will have only one-third of their respective cDNAs in frame with the N-terminal NubG. Good-quality NubG-x libraries should have an average insert size of ~1.5 kb and complexities in the range of 1 × 106 to 5 × 106 independent clones.

X-NubG cDNA libraries

These libraries are technically more challenging than NubG-x libraries (Fig. 6) for several reasons. First, oligo-dT-primed first strand synthesis cannot be used, because every cDNA made with an oligo-dT primer contains the 3′ untranslated region, as well as the stop codon of the gene. Therefore, no productive fusions to the NubG at the 3′ end can be made. To circumvent this problem, random hexamers that anneal throughout the length of the RNA are used in the first strand synthesis. A disadvantage is that random-primed libraries have smaller average insert sizes, and therefore fewer integral membrane proteins will be represented as full-length clones.

Furthermore, as there is no NubG coding sequence at the 5′ end to provide an ATG for initiation of translation, an ATG has to be provided by cDNA insert. Unless the library contains a substantial fraction of full-length cDNAs (which is technically challenging), only very few cDNAs will have an ATG within a feasible distance of the 5′ end of the cDNA. Therefore, the majority of cDNAs in such a library will not be translated at all. Of the fraction that is translated, two-thirds will, again, not produce a correct fusion with the downstream NubG cassette, leaving an extremely small fraction of productive cDNA-NubG fusions.

Alternatively, the ATG problem may be overcome by adding a short leader sequence to the 5′ end of the cDNA whose only purpose is to ensure that initiation of translation takes place. In this case, all cDNAs will be translated (as is the case for NubG-x libraries) but, because two fusion points are present in the cDNA clones, at most one-ninth of all clones will make productive fusions with the 5′ leader sequence and the 3′ NubG sequence.

Although highly desirable from a biological standpoint, x-NubG libraries are much more difficult to construct than NubG-x libraries. To create an adequate x-NubG library, one should aim at average insert sizes above 1 kb and at complexities exceeding 5 × 106 independent clones.

Library transformation and screening for interactors (large-scale transformation)

This high-efficiency protocol can be used to screen a bait protein of interest against a cDNA or genomic library fused to NubG (prey library). It is based on a cotransformation approach in which the bait gene carrying plasmid has already been introduced in the yeast strain by a simple transformation, before the yeast strain is transformed with the prey plasmid. It requires 20 μg of prey plasmid DNA and should yield transformation efficiencies in the range of 5 × 105 to 2 × 106 clones/μg DNA.

1. Inoculate a single colony of the bait plasmid-carrying yeast strain from a stock plate into 10 ml of SD−Leu (Recipe 10) and grow for 8 hours at 30°C.

2. Inoculate 100 ml of SD−Leu (Recipe 10) with the entire 10-ml culture and grow overnight at 30°C.

3. Measure the OD600 of the overnight culture. Calculate the amount needed for X ml = 30 OD600 units.

4. Pour a volume equivalent to 30 OD600 units of the yeast overnight culture into 50-ml Falcon tubes and centrifuge at 700g for 5 min at room temperature.

5. Resuspend the pellets in 200 ml of 2× YPAD (Recipe 3) and grow to an OD600 of 0.6 (two doublings) in a 1-liter Erlenmeyer flask at 30°C with shaking at 200 rpm.

6. Split the yeast culture into four 50-ml Falcon tubes and centrifuge at 700g for 5 min.

7. Resuspend each pellet in 30 ml of sterile ddH2O by vortexing.

8. Centrifuge at 700g for 5 min at room temperature.

9. Remove the supernatants and add 600 μl of sterile ddH2O to the yeast pellets.

10. Prepare four additional Falcon tubes with 5 μg of library plasmid and 100 μl of herring sperm single-stranded carrier DNA (2 mg/ml) in each tube.

11. Prepare PEG-LiOAc Master Mix (Recipe 9).

12. Add 600 μl of yeast cells from Step 9 to each DNA containing Falcon tube from Step 10.

13. Vortex briefly to mix.

14. Add 2.5 ml of PEG-LiOAc Master Mix (Recipe 9) to each tube.

15. Vortex for 1 min to thoroughly mix all components.

16. Incubate at 42°C for 45 min.

17. Pellet cells at 700g for 5 min at room temperature.

18. Resuspend each pellet in 3 ml of 2× YPAD (Recipe 3) and pool the contents of all four tubes into one Falcon tube.

19. Incubate at 30°C for 90 min with gentle shaking.

20. Resuspend the pellet in 4.7 ml of ddH2O.

21. Plate 300 μl per plate onto 15 SD−Trp, −Leu, −His plates (Recipe 10).

Note: If necessary, supplement the agar plate with 3-AT (see "Self-Activating Bait Proteins" in Troubleshooting to optimize 3-AT concentration).

22. Use the remaining resuspended cells to prepare 1:10, 1:100, and 1:1000 dilutions, and plate 100-μl aliquots onto two SD−Trp, −Leu plates (Recipe 10) for each dilution. Use these plates to calculate the transformation efficiency.

23. Transformants should appear after 2 to 3 days (SD−Trp, −Leu) or 3 to 4 days (SD−Trp, −Leu, −His +3-AT, for L40 yeast reporter strain) at 30°C. The number of colonies obtained varies for different bait proteins.

24. Perform an X-Gal filter assay on the primary selection plates, as described above in "X-Gal filter test."

Note: During the X-Gal filter assay, perform all procedures in a clean environment using sterile equipment, including the Whatman filter paper, to avoid contamination of the colonies.

25. Pick all colonies from the selection plates that are LacZ+ and restreak onto fresh SD−Trp, −Leu, −His plates (Recipe 10) supplemented with the appropriate concentration of 3-AT. Incubate for 2 days at 30°C.

Note: It is normal for some restreaked colonies to fail to grow. These are most likely false positives.

26. Repeat the X-Gal filter assay. Some colonies will not turn blue anymore (false positives), whereas others will only partially turn blue (mixed colonies).

27. Restreak mixed colonies by selecting yeast from the areas that have turned blue in the X-Gal filter assay.

28. Recover the plasmids from yeast using any common DNA mini-prep method.

29. Transform the plasmids into an E. coli strain using AmpR selection (18) and conventional methods.

30. Isolate the plasmids from the bacteria using any common method.

31. Perform restriction analysis to confirm the presence of a prey plasmid.

Bait dependency test

False positive clones can be eliminated by several approaches. False positives are proteins that behave as spurious interactors in biochemical assays and in the context of the two-hybrid system where they do not interact with the bait protein in a physiologically meaningful way. First, it is very important to show that a particular bait and prey combination is still able to interact when reintroduced into fresh yeast. This step proves that activation of the reporter genes actually takes place through the interaction of a bait and a prey protein encoded on the respective plasmids, and not by an artifact, such as genomic mutations in the yeast strain that may give rise to constitutive activation of one or several reporter genes. Once it is confirmed that a given prey protein activates the reporter genes by interacting with the bait protein, it is also important to determine whether the prey protein interacts specifically with the bait, or whether it just interacts nonspecifically with any protein overexpressed in the yeast reporter strain. This is done by coexpressing the prey protein with one or several unrelated control bait proteins and testing whether the yeast coexpressing the particular combination grows under selection or not. Because integral membrane proteins may be localized to particular membrane-surrounded compartments (such as, the plasma membrane, the ER, the Golgi apparatus, or the inner or outer mitochondrial membrane), it is important to choose a control that localizes to the same compartment as the bait protein used in the original screen (Fig. 9).

Fig. 9.

Bait dependency test. The prey protein-encoding plasmid is isolated from the putative positive clones, amplified in E. coli, and transformed into several bait protein-bearing strains, one of which is the strain expressing the original bait protein of interest and the others of which are unrelated bait proteins. Only the latter case should result in growth on selective medium and turn blue in the X-Gal test.

True interactors grow on SD−Trp, −Leu, −His plates when cotransformed with the bait construct, but not when cotransformed with the control bait. In the L40 yeast reporter strain, both the HIS3 reporter (growth on selective plates) and the lacZ reporter (blue staining in the filter lift assay) should be positive. If the yeast strain also bears the ADE2 reporter, this should also be activated (for example, in case of THY.AP4 and NMY32). Some prey proteins may interact strongly with your bait and weakly with the control bait. We recommend that these clones be analyzed further, because they may represent valid interactors.

Prey clones that score positive in the bait dependency assay should be analyzed by sequencing from the 5′ and 3′ ends to determine the identity of the encoding cDNA. Sequencing also allows verification that the cDNA is in frame with the NubG moiety.

1. Set up a 10-ml overnight culture of the yeast reporter strain in liquid YPAD (Recipe 3).

2. Measure the OD600 of the culture. The optimal OD is between 0.6 and 1.0.

Note: If the culture has overgrown (that is, the OD is greater than 1.0), dilute to OD600 0.2 and grow for another 4 to 5 hours.

3. Transfer the culture to a 50-ml Falcon tube and centrifuge at 700g for 5 min.

4. Discard the supernatant and resuspend the pellet in 2 ml of sterile ddH2O.

5. Label two microcentrifuge tubes (tube 1, test bait; tube 2, control bait) and add 2 μg of test bait or control bait plasmid to each tube.

6. Prepare fresh PEG-LiOAc Master Mix (Recipe 9).

7. Add 300 μl of PEG-LiOAc Master Mix (Recipe 9) to each tube of plasmid DNA and vortex.

8. Add 100 μl of yeast to each tube and vortex for 1 min.

9. Incubate at 42°C for 45 min.

10. Centrifuge at 1500g for 5 min.

11. Discard the supernatant and resuspend pellet in 50 μl of sterile ddH2O.

12. Plate onto one 145-mm diameter SD−Leu plate (Recipe 10).

13. Incubate for 3 days at 30°C.

14. Pick several colonies from each plate and inoculate into 100 ml of SD−Leu medium (Recipe 10).

Note: Each prey protein to be assayed will require 2 ml of yeast culture.

15. Grow the liquid culture overnight to an OD600 of 0.6.

16. Centrifuge for 5 min at 700g.

17. Discard the supernatant and resuspend the pellet in 5 ml of ddH2O.

18. Prepare two microcentrifuge tubes per each prey protein to be tested and label (tube 1, test bait; tube 2, control bait). Every tube should be prelabeled for each bait and prey combination.

19. Add 2 μg of prey plasmid to each of the two tubes.

20. Prepare PEG-LiOAc Master Mix (Recipe 9).

21. Add 300 μl of PEG-LiOAc Master Mix (Recipe 9) to each tube of prey plasmid and vortex.

22. Add 100 μl of yeast transformed with the bait construct to tube 1 and 100 μl of yeast strain transformed with the control bait plasmid to tube 2.

23. Vortex for 1 min to thoroughly mix all components.

24. Incubate at 42°C for 45 min.

25. Centrifuge for 5 min at 1500g.

26. Discard the supernatant and resuspend each pellet in 100 μl of ddH2O.

27. For each tube, plate 50 μl onto a 100-mm diameter SD−Trp, −Leu plate and 50 μl onto a 100-mm diameter SD−Trp, −Leu, −His plate (Recipe 10).

28. Incubate for 3 to 4 days at 30°C

Note: Colonies should be visible after 2 to 3 days on SD−Trp, −Leu plates and may require 5 days on SD−Trp, −Leu, −His plates.

29. Perform an X-Gal filter test as described above in "X-Gal filter test" or a quantitative pellet X-Gal assay according to methods described in Möckli and Auerbach (23).

30. Characterize positive clones by restriction enzyme digestion and sequencing after transferring the prey plasmid into E. coli.

31. Make glycerol stocks of the characterized yeast clones by resuspending a loopful of freshly growing yeast colony in 15% Glycerol (Recipe 19), followed by vortexing and freezing in liquid nitrogen. Store these clones at −80°C.


Problems in Detection of Bait Protein

In the MbYTH system, bait proteins encoded by yeast genes generally seem to be well tolerated and are expressed at sufficient levels to be detectable by Western blotting and functional assays, whereas mammalian bait proteins are more susceptible to variations in expression levels. The most commonly encountered problem in using mammalian bait proteins in the MbYTH is that the bait under investigation is not detectable by Western blotting. Currently, the reason for this problem is not clear, but several factors might play a role, such as RNA stability, protein stability, or both; suboptimal integration into the membrane; or activation of the unfolded protein response (24). Depending on the protein under investigation and based on our experiences, we suggest that the following steps can be taken to increase the protein expression levels.

The cDNA encoding the bait protein can be cloned into a bait vector with a stronger promoter. Although most mammalian bait proteins reach sufficient expression levels when cloned into the standard bait vector with the weak CYC1 promoter, stronger promoters, such as the ADH1 or TEF1 promoters, often help to increase the expression levels.

When working with a type I integral membrane proteins, it is helpful to remove the endogenous N-terminal cleavable signal sequence and replace it with a cleavable signal sequence derived from a yeast type I protein. We routinely use signal sequences derived from the invertase (SUC2) or mating factor alpha 1 (MFα1) genes.

Finally, addition of a short leader derived from the yeast Ste2p also helps to increase protein expression in many cases. This leader has no targeting function, but presumably increases expression of the bait protein by providing an ATG that is in a good context for initiation by the yeast translation machinery, such as AAAAATG (25).

Self-Activating Bait Proteins

Self-activating baits lead to a situation opposite that with undetectable baits, because in most cases, the problem is overexpression or protein instability. The most straightforward way to reduce self-activation by a bait protein is to increase the screening stringency by the addition of 3-AT to the selection plates. Please note that it is absolutely crucial to perform a library-scale transformation to determine the level of self-activation of any given bait. It is not sufficient to titrate a bait using small-scale transformations, because the transformation efficiency (and hence the number of sampled transformants) is much smaller than in a library transformation. However, if a given bait already displays self-activation in a small-scale transformation, one can say with reasonable certainty that the background in a library transformation will be too high for screening.

To titrate a self-activating bait protein, perform a library-scale transformation using the empty library vector instead of the cDNA library and plate the transformants on selective plates with increasing concentrations of 3-AT (for instance, increase linearly from 2.5 mM to 30 mM 3-AT). Select the 3-AT concentration at which no colonies are visible after 4 days of growth. This is the appropriate concentration for the library screen.

If your bait still displays self-activation on the highest 3-AT (that is, 30 mM) concentration tested, you may have to subclone your bait into a vector with a weaker promoter to decrease the expression level.

Because the overexpression of a membrane bait protein could lead to the above-mentioned self-activation phenomenon, it is preferable in the case of yeast proteins to use endogenously Cub-TF-tagged ORFs as baits. In this case, the expression of the endogenously tagged bait membrane proteins is under the control of its native promoter, hence reducing the chances of self-activation. Such an approach has recently been shown to be a valuable tool to study membrane protein interactions in a large-scale format (26).

NubG-x and x-NubG cDNA Libraries

When searching for integral membrane proteins as interaction partners of your protein of interest, it is important to realize that NubG-x and x-NubG libraries yield different types of interaction partners in a library screen. Because the NubG moiety must be located at the cytosolic face of the membrane to interact with the Cub-TF cassette, NubG-x libraries contain productive fusions only with soluble proteins or type II integral membrane proteins whose N termini are located in the cytosol. Type I integral membrane proteins, on the other hand, would locate the NubG on the lumenal side of the membrane, as the NubG is fused to the N terminus of the protein, which is located in the lumen. Consequently, to identify type I integral membrane proteins, an x-NubG library, not a NubG-x library, must be screened. Cytosolic interaction partners can be found using both types of libraries, although NubG-x libraries might work better, because the number of productive fusions between NubG and the cDNA is higher than in x-NubG libraries.

Identification of False-Positive Interactors

One of the most frequently encountered problems in the conventional YTH system is the isolation of false-positive interactors in a library screen. Nearly any YTH screen yields a certain number of false positives, but most (not all) of them can be eliminated by control assays.

An MbYTH screen also yields false positives; that is, certain proteins tend to be isolated in a screen, because their inherent properties make them self-activating and not because of their specific interaction with the bait protein. For instance, N-terminal fragments of wild-type ubiquitin have been found in several screens using x-NubG libraries. These fragments efficiently bind the Cub moiety of the bait fusion protein, and the RNA encoding ubiquitin is abundant in most eukaryotes, making this a common false positive. Interestingly, such fragments are found much less frequently in NubG-x library screens.

False positives can be eliminated by using one or several control bait proteins, as described in the Instructions section. However, care must be taken to choose appropriate controls. In the conventional YTH system, baits and prey are forced into the same subcellular compartment (the nucleus), whereas no such mechanism exists in the MbYTH system, in which properties inherent in the bait or prey protein determine its subcellular localization. In a bait dependency assay, it is therefore advantageous to choose control proteins that are most likely sorted to the same compartment as the bait protein under investigation. Thus, when studying a bait protein that is normally localized to the plasma membrane, a control protein with a similar localization should be used. It may also be helpful to conduct bait-dependency assays using not just one, but several control baits.

Related Techniques

In the past, several other yeast-based techniques have been developed that enable researchers to examine the interaction behavior of membrane-anchored proteins. These techniques include the Son of Sevenless (SOS) recruitment system (SRS) (27), Ras recruitment system (RRS) (28), the reverse RRS (29), heterotrimeric guanine nucleotide-binding protein (G protein) fusion technology (30), and the rUra3-based split-ubiquitin system (31). Several detailed descriptions of these techniques have been published (32, 33). However, to our knowledge, none of these assays have yet been reported to work as a screening system for membrane proteins.

Notes and Remarks

During recent years, the MbYTH system has become a broadly used method to examine the protein interactions of membrane-anchored proteins (32, 33). The major advantage of the MbYTH over other methods is that it can be used for many types of full-length integral membrane proteins independent of their localization in the cell. In the MbYTH system, the bait protein of interest must be a membrane protein; the other prerequisite is that the Cub-TF and NubG modules, which are fused to the protein of interest, be located in the cytoplasm. Another advantage of the MbYTH is that it can be used to monitor the interaction either between two membrane proteins or between a membrane and a cytosolic protein. In contrast to the classical YTH, there is no need for a nuclear localization signal, because the MbYTH system relies on the transport of a cleaved artificial and functional transcription factor into the nucleus. In general, the small ubiquitin moieties should present a minimized steric hindrance between the interacting proteins. Furthermore, the MbYTH approach also allows the isolation of novel protein interactors from cDNA libraries using an integral membrane protein as bait (11, 34, 35).

In conclusion, with the MbYTH, an in vivo system for monitoring membrane protein interactions has been developed, giving rise to new opportunities to detect interacting proteins and to examine novel therapeutic strategies based on interactions. The MbYTH system is rapidly becoming a preferred method to pursue these purposes by multiple researchers.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
View Abstract

Stay Connected to Science Signaling

Navigate This Article