Figure walkthrough using recombinant dna technology to make useful products

Abstract

Two-hybrid systems are one of the most popular, preferred, cost effective, and scalable in vivo genetic approaches for screening protein–protein interactions. A number of variants of yeast and bacterial two-hybrid systems exist, rendering them ideal for modern, flexible proteomics-driven studies. For mapping protein interactions at genome scales (that is, constructing an interactome), the yeast two-hybrid system has been extensively tested and is preferred over bacterial two-hybrid systems, given that users have created more resources such as a variety of vectors and other modifications. Each system has its own advantages and limitations and thus needs to be compared directly. For instance, the bacterial two-hybrid method seems a better fit than the yeast two-hybrid system to screen membrane-associated proteins. In this chapter, we provide detailed protocols for yeast and bacterial two-hybrid systems as well as a comparison of outcomes for each approach using our own and published data.

8.1 The Yeast Two-Hybrid System

Two-hybrid system analysis depends on the modular structure of transcriptional activator proteins. Many of these proteins consist of two domains, a DNA-binding domain (DBD) and an activator domain (AD). The DBD recognizes a specific sequence in the DNA upstream of a promoter and the AD stimulates transcription by binding to RNA polymerase (Fig. 15.18). Provided that the two domains interact, they will activate transcription. It is not usually necessary for the two domains to be covalently joined to form a single protein.

Proteins can be screened for binding partners using two-hybrid analysis.

In the two-hybrid system, both the DBD and the AD are fused to two other proteins (X and Y). These two hybrid proteins are referred to as the “bait” (DBD-X) and the “prey” (AD-Y). If the bait captures the prey—that is, if proteins X and Y interact—a complex will form and the gene will be activated. A convenient reporter gene is used to monitor for a successful interaction (Fig. 15.18).

The test proteins (bait and prey) are fused separately to the two halves of a transcription factor. If the bait and prey bind each other they will reassemble the transcription factor and activate the genes it controls.

Two-hybrid analysis was developed in yeast and is being used to generate a complete list of interactions between all 6000 or so yeast proteins. It is thus necessary to examine 6000×6000 combinations. To examine these potential interactions, each open reading frame in the yeast genome was amplified by PCR and cloned into two separate vectors, one carrying the DBD and one with the AD. Thus, each yeast protein is tested as both bait and prey. The vectors are designed to give in-frame gene fusions of each open reading frame with the DBD and AD of a suitable transcriptional activator, such as GAL4 (Fig. 15.19). One vector has a multiple cloning site downstream of the GAL4-DBD and thus gives a 3′-fusion of GAL4-DBD to protein X (GAL4-DBD-X). The other vector has its multiple cloning site upstream of the GAL4-AD and gives a 5′-fusion of GAL4-AD and protein Y (Y-GAL4-AD).

The bait and prey fusion plasmids are transformed into yeast cells of different mating types. This results in two sets of approximately 6000 transformants. All possible matings are carried out between the two sets using a laboratory robot to manipulate the colonies. When the two yeasts mate, the diploid cell will have a bait plasmid and a prey plasmid. If the two fusion proteins X and Y interact, the reporter gene is switched on. In yeast, the HIS3 or URA3 genes are usually used. The genes for β-galactosidase (lacZ), some fluorescent proteins (EGFP), or the biosynthesis of other nucleotides (ADE2) or amino acids (LEU2, LYS2) are used as well. If the reporter gene is not activated, the yeast strain cannot grow unless provided with histidine or uracil, respectively. If the reporter gene is turned on, the cells can grow on medium without histidine or uracil. Thus, the diploid cells from the 6000×6000 matings are selected on medium lacking the chosen nutrient (Fig. 15.20). Only those combinations where proteins X and Y interact yield viable colonies.

The yeast two-hybrid system has led to the discovery of thousands of protein interactions involved in plant responses, vertebrate limb regeneration, pathogen–host protein interactions, cell signaling, and many other systems. However, the original two-hybrid system has several limitations. For example, it relies on proteins expressed from sometimes non-native promoters and forced interactions within the nucleus of yeast. Membrane proteins often misfold when localized in the nucleus. Conversely, other proteins are only correctly modified when present in the cytoplasm. Non-yeast proteins might also misfold in the non-native environment. Toxic effects and steric problems with very large proteins may also cause some interactions to be missed. Furthermore, many proteins bind RNA and/or rely on small molecules to alter their conformation so promoting protein–protein interactions. In general, the yeast two-hybrid system addresses whether two proteins can interact but does not tell the entire story of the interaction. Regardless, the system has provided an excellent starting point for discovering protein interactions, but ultimately independent validation and functional analyses are needed.

Abstract

The two-hybrid system is an artificially constructed genetic system intended to facilitate the detection and assessment of protein–protein interactions. In the two-hybrid system a host organism, typically yeast or bacteria, is engineered so as to contain three components. These are a first protein fused to a DNA-binding domain of known specificity (hybrid 1); a second protein fused to a transcriptional–activation domain (hybrid 2), that can interact with the first protein, constituting a functional, albeit composite, transcription factor; and one or more reporter genes transcribed based on the binding of the composite transcription factor. Many permutations of the two-hybrid paradigm have been developed, and two-hybrid systems have become a mainstay of proteomic investigations.

A. Structure–Function Analysis

Two-hybrid analysis has been used extensively by the PCD community. Many novel members of the PCD machinery have been isolated in two-hybrid screens and many interaction domains have been mapped in detail.9 Several excellent reviews of the theoretical and technical details of two-hybrid analysis have been written and the reader should consult those for technical details.10 , 11

In addition to two-hybrid analysis, many researchers have capitalized on the fact that the interpretation of experimental observations in yeast is simplified by the absence of known homologs of the core PCD pathway members. It has therefore been demonstrated in S. pombe that the BH3 domain of Bak is necessary for lethality12 and that the lethality of Bax in S. cerevisiae is lost if Bax is intentionally localized to an intracellular compartment other than the mitochondrion.13 Interestingly, it has also been demonstrated that a form of Bcl-2 that is unable to interact with Bax in a two-hybrid test is able to rescue S. cerevisiae from Bax-mediated death.14

3.1 Two-Hybrid Screening

Two-hybrid screening approaches are high-throughput complementation assays that test for protein–protein or protein–DNA interactions. The assay is typically performed by introducing proteins of interest pairwise into yeast, with each protein fused to a transcription factor that has been split into two complementary fragments. Conventionally, the protein fused to the N-terminal DNA-binding domain of the transcription factor is referred to as the “bait” and the protein fused to the C-terminal activation domain as the “prey.” When brought into close proximity to one another through interaction between the bait and prey proteins, the binding and activation domains of the transcription factor function to activate transcription of a reporter gene. The reporter gene may encode for antibiotic resistance, such that interacting clones can be selected by applying antibiotic pressure. Alternatively, the reporter gene may code for a lethal gene, such that a physical interaction results in a reduction in colony size. Although two-hybrid approaches are typically performed using yeast, these assays have also been adapted to bacterial and mammalian systems (Joung et al., 2000). The disadvantages of two-hybrid approaches are that they often have high false-positive and false-negative rates. Producing proteins at far higher abundance than is biologically relevant can lead to spurious interactions that elevate the false-positive rate. False negatives may occur if N- or C-terminal fusions disrupt interaction interfaces, or if proper protein folding, processing, or posttranslational modifications cannot be recapitulated.

Two-hybrid approaches have been used to comprehensively characterize interactions between host proteins and proteins derived from a variety of viruses, including Kaposi sarcoma-associated herpesvirus, varicella-zoster virus, murine γ-herpesvirus 68 (MHV-68), vaccinia virus, SARS coronavirus, influenza virus (Friedel and Haas, 2011). In the case of influenza virus, an integrated approach was used to identify and validate interactions between viral and human proteins by complementing a comprehensive yeast two-hybrid assay with additional large-scale experiments (Shapira et al., 2009). This included the measurement of cellular transcriptional responses following transfection with influenza viral RNA, IFN-β treatment, and infection with an influenza strain lacking the NS1 gene (responsible for inhibiting the innate immune sensing of viral RNA and downstream IFN production). A set of genes found to be regulated in either the two-hybrid screen or the gene expression screens were also tested in siRNA knockdown screens measuring influenza replication and IFN-β production. Integrating the data resulting from these various assays revealed that viral polymerase subunits were enriched for interactions resulting in the positive regulation of IFN production, suggesting that the viral polymerase, in addition to NS1, plays a role in inhibiting the IFN response.

A similar integrated approach was used to characterize virus–host interactions for murine γ-herpesvirus MHV-68 (Lee et al., 2011). Using a yeast two-hybrid approach, a library of 84 MHV-68 genes was screened against each other to identify 23 intraviral interactions. The library was also screened against a cDNA library derived from human liver cells to identify 243 virus–host interactions. An affinity purification approach validated 70% of the intraviral interactions, giving an estimate of the false-positive rate of the yeast two-hybrid screen. Network analyses indicated that cellular proteins targeted by MHV-68 had more partners in a cellular protein–protein interaction network than expected by chance. This integrated screening and validation approach therefore yielded viral–viral and viral–host protein interaction networks.

3.2.1 Two-Hybrid Assay

The yeast two-hybrid assay is a classical method for identifying and characterizing protein–protein interactions in vivo (Fields & Song, 1989). The assay is based on reconstitution of a functional GAL4 transcriptional activator and consequent activation of a reporter gene under the control of a GAL4-responsive promoter. The target is cloned into a “bait” vector in frame with the GAL4 DNA-binding domain. The library to be screened (or the second of the two putative interacting proteins) is cloned in another vector in frame with the GAL4 activation domain. The two types of hybrid plasmids are cotransformed into yeast cells. If the two proteins interact, a functional GAL4 activator is reconstituted, and the expression of a reporter gene (lacZ in the case of SFY526 strain) is induced. Reporter activation can be assessed quantitatively by a β-galactosidase assay. Our laboratory has used the two-hybrid system to characterize interactions between an IF-binding domain of BPAG1 and neuronal IF proteins (Leung, Sun, & Liem, 1999). This approach was also used to identify MACF1 as a binding partner of p230/Golgin-245, a TGN protein involved in transporting GPI-anchored membrane proteins to the cell periphery (Kakinuma et al., 2004).

As with Far-Western blotting, it is important to remember that the two-hybrid assay identifies interactions of monomeric proteins rather than filaments. The interaction of a protein domain with a cytoskeletal filament protein in the two-hybrid assay may be different from its interaction with the corresponding filaments in cells or may depend on a specific modification of the filaments.

Prey Protein Subdomain-Dependent Interactions

The two-hybrid system is also very useful to determine the regions of prey proteins that interact with the Gα subunit. Regions or motifs of high homology in a protein family, for which one of the members shows up positively in a Gα two-hybrid screen, are obvious candidates for interaction domains. Finding a positive one-to-one interaction of the isolated region/motif with Gα in the two-hybrid system defines it as a Gα-interacting domain, and producing further deletions or mutations within it may help delineate the minimal requirements for interaction. Generally this delineation is first performed using the β-Gal filter assay, and a liquid β-Gal assay is usually not necessary. This is how RGS domains4 were initially discovered (see Table III). Often the correct folding or the borders of a domain in a newly isolated protein are unknown; hence some trial and error in making two-hybrid constructs (in the prey plasmid) of the putative interaction domain should be taken into account.

The reverse approach, determining the site of interaction within the Gα subunit, has been less useful because several deletion mutants have been shown to misfold. Preliminary experiments to show proper folding of Gα deletion mutants are strongly advised before using them in two-hybrid assays.

1 Theory

The two-hybrid system is an in vivo yeast-based system that takes advantage of the modular nature of the yeast GAL4 transcription factor. GAL4 has two domains whose activities can be separated: the DNA-Binding domain and the transcriptional Activation Domain. The two-hybrid system identifies the interaction between two proteins (X and Y) by reconstituting these two GAL4 domains and thus allowing transcriptional activation of a reporter gene, which has been designed to be a selectable marker. Reconstitution of the DNA-binding domain and the activation domain will only occur if X and Y interact. To achieve this, two fusion proteins are constructed, one of which contains the DNA-binding domain fused to the first protein of interest (DB-X, also called the ‘bait’) and the other of which contains the activation domain fused to the second protein of interest (AD-Y, also called the ‘prey’). DB-X–AD-Y interaction reconstitutes a functional transcription factor that activates reporter genes driven by promoters containing the relevant DB sites. A selectable marker such as HIS3 is used as a reporter gene, and transcription activation resulting from the interaction of DB-X and AD-Y can therefore be assessed by a growth of cells on agar plates lacking histidine. Thus, the yeast two-hybrid system allows for the detection of protein–protein interactions genetically.

Yeast Two-Hybrid Assay

A yeast two-hybrid assay investigates protein–protein interactions by exploiting a transcription system normally used by yeast cells. In a normal yeast cell, a transcription factor called Gal4 binds to a promoter region called an upstream activating sequence (UAS). Gal4 is composed of a binding domain (BD), which binds to the UAS, as well as an activation domain (AD), which initiates transcription of a target gene.

In a yeast two-hybrid assay, a scientist uses recombinant DNA technology to divide the Gal4 protein into these two separate domains (Figure 15.15A). The Gal4 binding domain is fused to a protein of interest thought to interact with another protein. This protein can be considered the bait in that its role in the experiment is to attract other binding partners. The scientist fuses the Gal4 activation domain to a potential binding partner. This binding partner can be considered prey in that it could potentially bind with the bait. If the two proteins do not interact, the BD and AD fragments will be physically separated, and no transcription will occur. However, if the two proteins are able to bind, the Gal4 BD and AD fragments will become physically close enough to cause transcription of the downstream coding sequence (Figure 15.15B; Box 15.3). If the gene adjacent to the UAS is a reporter gene, such as lacZ, the scientist will be able to identify the interaction between the two proteins.

BOX 15.3

Walkthrough of an Intracellular Signaling Experiment

Let us say that you work in a laboratory that studies a specific aspect of neural development: how a neuron’s axon correctly extends to its postsynaptic target. The tip of an axon is referred to as the axonal growth cone, and it typically reaches its destination by responding to extracellular guidance cues produced by other cells. There are three general steps that occur to guide the growth cone: (1) an extracellular guidance cue is secreted to guide the axon’s position; (2) a receptor on the axon’s surface recognizes that cue and transduces it into an intracellular signaling cascade that affects other proteins within the developing neuron; and (3) various downstream organelles receive and respond to the signal, including the cytoskeleton that must be reorganized to move the growth cone toward the site where the initial cue was detected. If you know the identity of the receptor and the extracellular cue, you might be interested in figuring out how the receptor translates the cue into a response. What is the intracellular signaling pathway that leads to the cytoskeletal changes necessary for the growth cone to turn toward the cue? Let us focus on identifying the next step of signaling that occurs after the receptor is activated: the activation of a second protein that binds to the intracellular domain of the receptor.

There are multiple approaches for finding proteins that interact with the receptor. For example, you could immunoprecipitate the receptor and then perform mass spectrometry to identify any other proteins that co-precipitated. However, this approach would not tell you if the two proteins interacted directly or were simply part of a larger complex of proteins. Another approach would be to perform a yeast two-hybrid assay. This technique is used to identify protein-binding partners that directly interact with the receptor. To perform this assay, you would begin by genetically engineering a construct that attaches the receptor’s intracellular domain to the DNA-binding domain of a transcriptional activator, such as Gal4, to act as the bait. Then you would create or use a cDNA library to attach different cDNA molecules to the activation domain of Gal4 to act as the prey. Finally, you would introduce the bait and prey constructs into yeast cells that express a reporter gene, often lacZ, under the control of a promoter that is recognized by the transcriptional activator (UAS for the Gal4 activator). Thus, the bait will bind to the promoter but cannot initiate transcription of the reporter until a specific prey with the activation domain interacts with the bait to create the full activator. Then, the reporter gene will be transcribed and can be detected.

To detect yeast cells that express lacZ, you could grow thousands of cells on an agar plate that also contains X-gal, the substrate that causes a visible blue byproduct to form in those cells. The final step would be to pick blue yeast colonies off the plate, extract and sequence the DNA of the prey, and identify the protein that interacts with the receptor based on the cDNA sequence. This screen allows you to conclude that the prey you found can interact with the intracellular domain of the receptor. Most likely, you will have discovered a number of positive hits to sequence and should confirm the interaction’s specificity to determine which would be a good candidate for a follow-up.

This yeast two-hybrid experiment is a good start, but you could learn even more about the receptor–protein interaction by attaching different functional regions of the receptor’s intracellular end to use as bait. This would tell you with finer resolution exactly what part of the receptor is interacting with the prey or allow you to capture only prey that will interact with a specific region—say, the kinase domain—or a particular region you know to be functionally important. To convince yourself and other scientists of the interaction between the receptor and intracellular protein, you could perform additional, complementary experiments. For example, co-immunoprecipitation, as just described, is often used to verify yeast two-hybrid results. You could also use protein affinity chromatography. In cell signaling studies, there is no such thing as too much evidence; figures often contain several different methods to validate the same result.

Now that you have a candidate (or several candidates, based on the results of the yeast two-hybrid experiments), you should follow up with other experiments to make sure that your candidate protein makes sense as a protein that mediates the intracellular response to a guidance cue. One question you might ask right away is whether the candidate protein is expressed in the right place (the axon growth cone) and at the right time (the embryonic or postnatal period during which the axon grows toward its target). If the interacting protein is not present in axonal growth cones or at a time when axons are growing out, the interaction may not be relevant to the growth cone turning behavior you’re interested in, or the interaction may not actually occur. To examine protein expression, you could use immunohistochemistry to stain for the expression of the protein in histological preparations of the growth cone at different developmental time points. You could also collect tissue samples containing the growth cones from different time points and perform western blots to examine the expression of the protein at each time point.

If you have determined that a protein binds to the receptor, is expressed in the growth cone, and is expressed during the developmental period in which the growth cone migrates toward its target, you have laid the foundation for many future experiments. For example, you could ask whether the interaction between the receptor and the protein causes the protein to be phosphorylated. You could use techniques described in other chapters, such as RNA interference (RNAi—Chapter 13), to investigate the consequence of knocking down the expression of the intracellular protein. There are many future experiments that can and should be performed, and the more complementary lines of evidence you collect, the more you will know, and the better your study will be received.

A yeast two-hybrid assay is useful to test hypotheses that two proteins interact, but this assay can also be used as a screen to identify potential binding partners for a protein of interest (Figure 15.16). For example, a scientist can identify unknown “prey” for a protein of interest by attaching the activation domain to a large mixture of DNA fragments from a cDNA library (Chapter 10). Individual ligation products are then introduced into yeast cells containing the target protein. If one of the members of the library encodes a protein that interacts with the target protein, the two Gal4 fragments will interact and activate transcription of the reporter. The scientist can then identify the yeast colony expressing the reporter, purify and sequence the DNA attached to the activation domain, and determine the identity of the binding protein. While this assay shows that the proteins can directly interact, it does not show that they interact in vivo or even that they can interact in mammalian cells.

D Role of ELF3 as an Adaptor for COP1 and GI, Possibly in the CRY2-Dependent Flowering Pathway

Yeast two-hybrid analysis and co-immunoprecipitation assays have demonstrated molecular interactions between ELF3 and the E3 ubiquitin-ligase CONSTITUTIVE PHOTOMORPHOGENESIS 1 (COP1), and between ELF3 and GI (Yu et al., 2008). It was proposed that ELF3 acts as an adaptor protein between COP1 and GI. The blue-light receptor CRY2 may negatively regulate COP1 via a direct interaction between the two (Wang et al., 2001). Genetic analysis under LD and SD regimes support this idea. For example, the late-flowering phenotype of cry2 was suppressed by cop1 under LD, suggesting that COP1 may be a downstream factor of CRY2 under various light/dark cycles. Thus, blue-light-activated CRY may stabilize the CO protein in the GI-independent pathway by inhibiting COP1 (Valverde et al., 2004; Yu et al., 2008). Indeed, Jang et al. (2008) and Liu et al. (2008a,b) have shown that COP1 triggers degradation of the floral inducer CO. In this way, the cry2 loss-of-function mutation would cause increased accumulation of COP1, thereby inhibiting the stabilization of CO.