Sciclips's Blog

Are large protein fusion tag based technologies, including protein arrays, truly useful in functional proteomics or systems biology or drug discovery research?

Posted on: October 5, 2010

There are several technologies developed for studying protein-protein interactions and several products available in the market using some of these technologies. These technologies include tandem affinity purification and protein microarray based methods/kits for functional proteomics and system biology applications. These technologies look very promising and elegant. However, careful scientific analysis of these technologies raises some concerns and questions. These concerns and questions can be debatable, but, these are important issues to be discussed.

Multiprotein complex interaction studies: Tandem affinity purification (TAP) or similar techniques are mainly used for studying multiple protein-protein interactions in protein complexes or protein network analysis. TAP has significant applications in functional proteomics, systems biology and drug discovery studies. The methods used in TAP can be broadly classified into two categories: 1) protein fusion tag based methods and 2) Immunoprecipitation approaches. Several fusion tags such as Strep-tag, calmodulin binding peptide (CBP), histidine-tag, GST, MBP, HaloTag® etc. have been used for protein-protein interaction studies, including TAP. The use of fusion tags, especially large fusion tags such as GST, MBP or HaloTag®, raises some questions regarding the validity of these large fusion tags in analyzing protein-protein or protein-nucleic acid interactions (e.g. chromatin immunoprecipitation). If a large fusion partner is attached to a bait protein (the protein which is used to find its partner/s), this fusion protein will be significantly different from the natural protein. If this is true, can naturally interacting proteins recognize this new fusion protein? Take an example where calmodulin is fused GST (26KDa) and is used to identify calmodulin binding partners. The natural calmodulin interacts with several proteins at different stages of various cellular events. If the natural calmodulin is fused with a protein like GST, calmodulin-GST fusion protein will not be similar to the natural calmodulin. If this is true, do calmodulin interacting proteins can recognize GST-calmodulin? Probably not all proteins that interact with calmodulin, because protein-protein interactions within a cell are tightly regulated and conserved. The fusion protein will never have the natural three dimensional structural integrity or surface charge, energy state, charge state and so forth. We can argue that, there are several published reports on identifying interacting partners using this technology. How do we know that these proteins are really interacting proteins, except model proteins? Does this mean that the current technology will not determine all proteins that interact with natural calmodulin? A single mutation in calmodulin gene or a modification of a single amino acid in calmodulin (e.g. oxidation) can drastically affect the function of calmodulin. We can argue that fusion partners did not affect the structural integrity of protein partner, may be using a linker. If this is true, how can we explain the reports that some proteins loose it’s function or aggregates once you remove the fusion tag or some protein are functionally inactive if it is fused with a fusion tag or some proteins are folded correctly if is fused with fusion tag (e.g. GST,GFP, MBP,HaloTag® etc). These reports clearly demonstrate that a fusion tag can significantly affect the three dimensional integrity of proteins. A systematic and comparative analysis of fusion proteins vs. non-tagged proteins are needed to confirm the application of large fusion tags in protein-protein interaction studies. This analysis should not be done with model systems; rather it should be done with complex samples. Until we confirm this, antibody or small fusion tag (FLAG/FLASH/Step-tag/histidine tag) based technologies may be more suitable than large fusion tags. Non-tag approaches coupled with antibody based technologies (immunoprecipitation) will be the ideal method. The fusion tags were discovered for helping in recombinant protein purification; any applications beyond this should be critically evaluated based on strong scientific principles.

Protein array based protein-protein interaction studies: The use of large fusion tag proteins for oriented immobilization of proteins have been claimed as a better method for studying protein-protein interactions using protein microarrays (e.g. GST, MBP, HaloTag® fusion proteins). These claims have to be re-visited based on proven scientific principles. If we immobilize a protein onto a surface using a large fusion tag, how does this approach guarantees that the immobilized protein can identify its interacting proteins similar to the non-fusion protein inside a cell? Proteins are dynamic in nature and there are several sequential physical and biological factors involved in complex protein-protein interactions within a cell. Immobilization of a protein onto a confined position, as in protein microarrays, will limits its ability to interact with real protein partners. Immobilized proteins on any surface may not mimic protein’s function in real cellular context. As we mentioned earlier, the fusion protein will never have natural three dimensional structural integrity or surface for interactions etc. like the “wild type” non-tagged protein. Immobilizing a fusion protein will further jeopardize structural and functional integrity of the protein. It is true that there are several reports and successful products in the market demonstrating that immobilized fusion proteins can be used for enzymatic functional screening. In other words, immobilized proteins are functionally active. Studying the enzymatic function of a protein may not be an issue using protein arrays. The functionality of immobilized proteins can be validated using specific assays. This may not be the case with protein-protein interaction studies or complex interaction analysis. There is no control system to test these interactions, other than using a model system using single or multiple proteins. Any assumptions made on current protein array technologies using large fusion tags should be critically evaluated. Protein arrays may not be a viable tool for studying protein-protein interactions in critical areas like drug discovery and clinical diagnostics, exceptions are antibody arrays. Probably, we will never be able to identify real protein partners using protein arrays or any methods that involves protein immobilization onto a surface.

We should think of alternative technologies to unravel complex or simple protein-protein interaction networks within a cell. A technology that does not require manipulation of proteins has to be considered in identifying true protein-protein interactions. Antibody based technologies are an immediate solution to this, but it comes with several technical challenges such as non-specificity. Emerging non-invasive technologies such as Rama spectroscopy or similar technologies (IR/laser etc) may come-up as powerful tools in future. Innovative technologies that can address true biological interactions have to be developed based on strong scientific principles. Development or use of technologies or products that lack true scientific context may jeopardize scientific innovations; loss of time and money are not to mention.


3 Responses to "Are large protein fusion tag based technologies, including protein arrays, truly useful in functional proteomics or systems biology or drug discovery research?"

I agree with most of the concerns expressed on the use of large tags. Unfortunately, the same is true for small tags and the use of antibodies.
If we look at the small tags such as various versions of 6xHis or Flag we can easily see that they are usually charged and they can affect binding and sometimes even proper folding of the fusion partner.
Antibodies can also bind to protein epitopes that are involved with binding protein partners. Lets not forget that most of the “good” antibodies have high affinity for the epitope on the target protein and will most likely displace a potential binding partner to the wild type target protein.
In my former job I have worked on development of small tag that has overall neutral charge and also developed antibody based arrays. In my experience (and according to the opinion of folks that used the array) in some cases one could pool down specific protein complexes but this is not always the case. The only comprehensive way of dealing with a problem like that in a multiplex format is utilizing one two platforms:
1. Antibody based array on which every target protein has multiple immobilized antibodies directed at different epitopes distributed over the entire 3D structure of the protein.
Obvious weaknesses of this approach are cost, need for similar affinity of all binding antibodies for the target protein, ability to detect the binding partners in a multiplex format (my dream to combine the platform I developed with MS downstream ;-).
In my experience a huge benefit is the significant stability of the antibody based array (we had arrays that were fully functional even after 4 Years of storage).
2. Protein based array based on random immobilization chemistry of the wild type proteins.
The difficulties of this approach are quite numerous. Getting functional and stable purified wild type proteins is time consuming and expensive (and sometimes impossible with the current purification principles). Getting a functional immobilized wild type protein is another difficulty, which is even further complicated by the significant differences in stability of the proteins. We have observed these issues with some antibodies as well. If the use of the array is carried out right after its preparation one could minimize the problem. However, in our experience, some proteins are “dead on immobilization” no matter how careful is one that carries the immobilization process. The use of 3D surface for immobilization, while helping with stability in some cases, inevitably results on higher background (i.e. non specific binding).
Let me put a problem that is even more challenging to address – how to do this with transient protein-protein interactions? After all, these are the most interesting targets to study ;-).
All the stuff I share is based on our experience and does not necessarily agree with what other groups observed.

“The first example of a protein acting as a post-translational modifier was ubiquitin, but over the past, 15 years, a series of ubiquitin-like modifiers (UBLs) has also been discovered. Compared with small molecule modifiers such as phosphoryl or methyl groups, ubiquitin and UBLs provide much larger and more chemically varied surfaces. Thus, these covalent modifiers can act as flexible adaptor modules for altering protein conformation
or protein–protein interactions. UBLs are also more plastic in an evolutionary sense: duplication and diversification of UBL-encoding genes can give rise to multiple related molecules that can acquire new functions and participate in distinct cell regulatory mechanisms.”


Schwartz DC, Hochstrasser M. A superfamily of protein tags: ubiquitin, SUMO and related modifiers. Trends Biochem Sci. 2003 Jun;28(6):321-8.

LifeSensors is dedicated to providing researchers with cutting edge technologies for recombinant protein and ubiquitin research.

for the interaction studies adoption various tagging proteins, is the only practical way of doing large scale studies. so as a reader of such study, I think onl should only take the large quantity of interaction data as a starting point( or a hint) for the more specific indepth look by other in vivo based studies. and several studies addressing same interaction will also give a more strong suggested interaction possible.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

Comprehensive Cancer Bomarker Database

Sciclips consultancy

Diagnostic & Prognostic Biomarker Database

Bioprotocol database (open access)

Combination therapy database (open access)

Therapeutic drug target database (open access)

Bioinformatics databases (open access)

Stem cell researchers database

A comprehensive database for stem cell researchers, Stem cell research reagents, Stem cell patents, Stem cell grants, Stem cell clinical trials, Stem cell statistics, Stem cell news and Stem cell open innovation ideas. Please visit

Sciclips consultancy

Proteomics researchers database

A comprehensive database for proteomics researchers, Proteomics research reagents, Proteomics patents, Proteomics grants, Proteomics clinical trials, Proteomics statistics, Proteomics news and Proteomics open innovation ideas. Please visit

Share or sell research reagents prepared in your laboratory

Drug target database (open access)

Pharmacogenomics protocols online

Proteomics and Mass spectrometry (MS) protocols online

Therapeutic drug targets database (open access)

Drug discovery, Biopharmaceuticals, HTS assays protocols online

Protein display/Protein engineering/Directed evolution protocols online

Molecular biology and Cell biology protocols online

siRNA protocols miRNA protocols online

Video protocols

Cell culture video protocols, Cytotoxicity assay video protocols, Stem cell video protocols, Molecular biology video protocols

Sciclips consultancy


Bioprotocols online (open access)

Combination therapy database (open access)

Therapeutic drug targets database (open access)

Bioinformatics databases (open access)

HTS assay protocols online (open access)

%d bloggers like this: