Methods to study Protein-Protein Interactions - Berthold Technologies GmbH & Co.KG

Classification of in vivo methods used to study protein-protein interactions

There are many ways to classify in vivo methods that are used to study protein-protein interactions. If the focus is particularly on the detection and quantification of the interaction, it can be classified using the properties of the reporter system used by the respective methods.

All methods presented here are based on a reporter system consisting of two parts: one part is linked to one of the proteins of interest (sometimes called the “bait” protein) and the other one to the other protein of interest (sometimes called the “prey” protein). When both proteins interact with each other, the two parts of the reporter system are brought into close proximity and have a clearly detectable effect. The terms “bait” and “prey” are normally used if there is a protein of interest and the objective is to “hunt” for interacting partners. However, in many cases both partners are already known and the objective is to study the interaction in more detail (for example, its regulation).

The methods mentioned below are discussed in more detail in Xing et al. 2016.

Methods according to the functionality of the reporter system

An important difference is whether the two parts of the reporter system are fully functional by themselves or not.

Protein fragment Complementation Assays (PCA)

In the protein fragment Complementation Assays (PCA), the reporter is a single protein that has been cleaved into two fragments. The fragments by themselves are not functional, but the protein can reconstitute itself when both fragments are brought into close proximity and restore its functionality. For example, in the split luciferase complementation test, the luciferase molecule is split into two fragments, neither of which can produce light, but if both fragments are close enough, the luciferase is reconstituted and can produce light again. Examples of PCA methods are the split ubiquitin system and the split luciferase complementation assay.

Two-hybrid assays

In two-hybrid assays, each part of the reporter system is a protein or protein domain that is fully functional, but a specific, measurable effect is produced when both parts are close enough. For example, in yeast two-hybrid assays, both the DNA-binding domain that is linked to one of the proteins of interest and the transactivation domain that is linked to the other protein are independently stable and functional, but the reporter gene is only expressed if they are in close proximity to each other. Similarly, both fluorescent proteins used in a FRET assay are fluorescent, but the acceptor protein emits light only if there is interaction between the proteins of interest.

Methods according to the detected signal

Another important classification is the type of signal generated by the reporter and used to detect the interaction. This affects the instrumentation required for detection, the possibilities it provides for accurate quantification, and suitability for high-throughput analysis and the like.

Methods based on gene expression

In this type of method, if both proteins of interest interact, the combination of the DNA-binding domain and the transactivation domain drives the expression of the reporter. Reporters are normally of two possible types: prototrophic markers, such as HIS3 or ADE2, enabling survival on depleted growth medium, or chromogenic enzymes such as ß-galactosidase, which can be used for the selection via blue/white coloring of the colonies. Both the yeast two-hybrid assay and the split-ubiquitin system (see below) are based on gene expression.

This type of reporter is not capable of following dynamic changes in the interaction, and can only be quantified by absorption and by using a chromogenic enzyme. However, such reporters can be used to evaluate large numbers of binary interactions by mixing two haploid mating types that have been transformed with different sets of proteins of interest. Hence, they are very useful for large scale screening of protein-protein interactions.

In the classical version of this method the N-terminal DNA-binding domain of the transcriptional activator GAL4 is bound to one protein of interest, and the C-terminal transactivation domain to the other one. When both proteins interact, their association produces a chimeric transcription factor, that activates gene expression downstream of the upstream activation sequence GAL1 (Fields and Song 1989). Further split probes have been successfully established. The yeast two-hybrid assay is easy to perform, but can only be used to study interactions between proteins with nuclear localization.

This system is similar in concept to the yeast two-hybrid system and is just as easy to implement. However, it has the additional advantage that it can be applied to virtually any protein, not only to nuclear proteins. As explained above, the split-ubiquitin system is a kind of PCA: the protein ubiquitin is split into two non-functional fragments, each of which is fused to one member of the protein pair of interest. When both fragments interact, ubiquitin reassembles and becomes functional. This promotes cleavage via the ubiquitin-specific protease. In the classical version, this cleavage leads to the release of a transcription factor, that then activates the expression of the reporter gene.

Methods based on fluorescence intensity

Fluorescent proteins are used to detect interactions. Although all fluorescence-based methods have a higher sensitivity and a higher dynamic range than methods based on gene expression of chromogenic enzymes, not all of them are suitable for following dynamic interaction changes.

The most widely used method of fluorescence is FRET. In this method, the fluorescent donor protein is coupled to one of the proteins of interest and the acceptor to the other protein. If there is no interaction, only the fluorescence of the donor should be detected. However, if an interaction takes place (with both fluorescent proteins at a distance of 10 nm or closer), the fluorescent donor protein can excite the acceptor via resonance energy transfer (dipole-dipole coupling). It is normal that some emission of the acceptor is detected even without interaction. This happens because often a small portion of the excitation light is also able to excite the acceptor if there is a spectral overlap in the excitation of the two fluorescent proteins - even if suitable filters are used. Therefore, the FRET ratio (emission of the acceptor divided by emission of the donor) is normally used for quantification, with a significant increase in the ratio indicating an interaction. This type of FRET is sometimes referred to as Sensitized Emission FRET (SE-FRET or seFRET). However, since SE-FRET is the only common method used to measure FRET with a microplate reader, it is usually referred to simply as FRET. Other types of FRET used with fluorescence microscopes are FLIM FRET (see below) and FRET acceptor photobleaching (where the increase in fluorescence of the donor is measured after the photobleaching of the acceptor).

Like most methods based on fluorescence intensity, FRET suffers from a high background and therefore low signal-to-noise ratios, which limits its sensitivity. Several alternative methods have been developed to overcome this limitation: BRET (see below) and TR-FRET. In TR-FRET (time-resolved FRET), fluorophores with a very long fluorescence lifetime (mostly chelates of lanthanides, such as europium, terbium and samarium) are used to bypass interference from molecules with a short fluorescence lifetime or other factors (especially excitation light). Instead of measuring light when excitation light is on, TR-FRET measurements are performed hundreds of microseconds later. The chelates can still emit light, but all other fluorophores in the sample and of course the excitation light have already faded away. Time resolved fluorescence (TRF) is not to be confused with fluorescence lifetime (see below).

BiFC is a PCA that uses a cleaved fluorescent protein (GFP or another). Fragments of the protein are not fluorescent, but the fluorescent protein is reconstituted when both fragments are brought into close proximity. Similar to FRET, BiFC is a fluorescent method for measuring PPI, but there is one important difference: the complementation of the cleaved fluorescent protein fragments is irreversible, making the method unsuitable for investigating interaction dynamics. Another complication with this method is that the cleaved fragments eventually reassemble by themselves, which increases the detection of false positives and requires extreme care in planning the experiment and designing appropriate controls. However, the fact that BiFC is irreversible can become an advantage: It can be used to identify and quantify weak interactions that are difficult to investigate with other methods.

Methods based on fluorescence lifetime

The fluorescence lifetime describes the average time a fluorescent protein spends in the excited state before returning to the ground state by emitting a photon. Methods based on fluorescence lifetime are based on measuring the lifetime of individual fluorophores rather than their emission spectra. Since fluorescence lifetime does not depend on concentration, sample absorption, excitation intensity, or spectral bleed, it is more robust than intensity-based methods. Fluorescence lifetime is used for FLIM (Fluorescence Lifetime Imaging Microscopy) imaging.

FLIM FRET

An interesting phenomenon is that FRET not only increases the fluorescence emission of the acceptor but also reduces the fluorescence lifetime of the donor due to energy transfer. The combination with FLIM to visualize FRET enables PPI to be investigated in vivo with high reliability, enabling highly dynamic spatiotemporal resolution. However, a successful use of FLIM requires sophisticated hardware add-ons and extensive training.

Fluorescence lifetime measurements can also be performed with a fluorometer. While there are a number of single sample fluorometers that can perform this procedure, the availability of microplate readers that can perform high-throughput fluorescence lifetime measurements is very limited.

Methods based on luminescence

Luminescent reporter systems such as the split-luciferase complementation assays are also popular. Here, the reporter is a luciferase and emitted light can be easily quantified using a luminescence reader.

This is a PCA based on the complementation of two cleaved fragments of a luciferase protein and its subsequent detection by substrate conversion and light production. The method is very similar to BiFC but uses a luciferase instead of a fluorescent protein. However, there is an important difference: the reconstitution of luciferase is reversible in most cases, and therefore suitable for monitoring dynamic interactions in real time. Since most model organisms are not luminescent, split luciferase complementation is a very sensitive detection system with a high signal-to-noise ratio. Both Firefly luciferase and Renilla luciferase have been successfully used to study PPI in vivo. Split luciferase complementation is a popular method to study protein-protein interactions in planta by using a plant in vivo imaging system to visualize luminescence on transformed leaves.

BRET (Bioluminescence Resonance Energy Transfer) was developed to bypass some of FRET's background problems. Instead of using a fluorescent protein as a donor, a luciferase is used instead. However, since luciferases produce light through a chemical reaction, an external light source is not required for excitation. BRET therefore has a very small background and does not suffer from problems often associated with FRET, such as autofluorescence, light scattering or photo bleaching.

The classification of BRET as a luminescence method is therefore controversial: while the measured signal represents the emission of a fluorescent protein (e.g. eYFP for BRET1 and GFP for BRET2), the excitation lamp cannot be used (since the acceptor is excited by the donor luciferase via dipole-dipole coupling). For the detection of BRET, luminescence readers are therefore used instead of fluorescence readers, which is why this technique is often classified as a luminescence method for practical reasons.

All methods discussed here are summarized in the table below:

Comparison of In Vivo methods to study PPI

In Vivo PPI Methods
Method	Signal	Dynamic?	Large-scale screening of PPIs?	Comments
Yeast Two Hybrid system	Bacterial survival/color (absorbance)	No	Yes	Inexpensive, but limited to nuclear proteins
Split-ubiquitin system	Bacterial survival/color (absorbance)	No	Yes	Inexpensive
FRET (SE FRET)	Fluorescence intensity	Yes	No	High background
TR-FRET	Time-Resolved fluorescence intensity	Yes	No	Reduced background compared to FRET
BiFC	Fluorescence intensity	No	No	May be useful to detect weak interactions
FLIM FRET	Fluorescence lifetime	Yes	No	Very reliable, but sophisticated equipment required
Split-luciferase complementation	Luminescence	Yes	No	Exogenous substrate needed
BRET	Luminescence/Fluorescence	Yes	No	Reduced background compared to FRET; exogenous substrate needed

More about Instruments used to study PPI