DogTag - a genetically encoded proximity labelling strategy to capture problematic protein-protein interactions

DogTag - how to find a protein's partners in living organisms

Proteins are the key functional molecules in the cell, forming structural building blocks and performing the many specialized functions a cell requires to survive. Making sure that individual proteins perform the right task at the right time in the right place is largely achieved by regulating which other proteins it interacts with.

Defining these interactions is one of the core problems of modern molecular biology. All methods used so far suffer from a number of disadvantages:
1. Disruptive -the cells of the organism must be broken apart before the interactions can be assessed. This leads to many interactions falling apart and being missed. This is particularly true for weak or transient interactions such as those occurring between enzymes and their substrates or interactions with proteins found in the cell membrane where detergents have to be used to break the membrane. Enzyme-substrate and membrane protein interactions are some of the most important for understanding disease and developing drugs making this knowledge an important priority in biology.
2. Out of context - many methods use a different organism (heterologous system) to test interactions as it is easier and faster (e.g. yeast is used as a surrogate system for investigating interactions between human or plant proteins). In a heterologous system many components many be missing, this means that many large protein complexes composed of multiple interactions cannot be assessed.
3. Non-physiological - many systems use cell culture as a substitute for whole organisms. While necessary in these instances they ignore the context of the tissue, organism or physiological conditions and therefore miss the real interaction patterns of proteins in a particular situation.
4. Outside input needed - Many systems used to look at protein interactions require some form of outside input to reveal the interactions. This means that whole organisms or tissues cannot be used in these circumstances.

We have designed a strategy to get around many of these problems, this therefore represents a milestone in protein-protein interaction analysis and opens the door for more physiologically relevant analyses to be performed and reveal more relevant data to researchers.

By taking a protein of interest and linking it to a bacterial enzyme capable of adding a small tag onto other proteins we can define the protein environment surrounding our protein of interest. Even more importantly the enzyme and tag are both proteins themselves so can be encoded as DNA and placed into the genome of the study organism. This means that, for the first time, it is possible to identify protein partners in the proteins native cellular environment in a tissue of interest (plant seed, animal liver, etc.) in a whole organism (whole plant, animal, etc.) under physiologically relevant conditions or stresses (low oxygen, drought, etc.) without any outside input.

The aim of this work is to take our proof of principle data and demonstrate that this system can be used to discover novel protein interactions from organisms such as fungi, plants and animals and therefore fuel novel primary research in hitherto inaccessible study areas.

Defining protein-protein interactions is a cornerstone of understanding how cellular systems work. We have developed a novel genetically encoded strategy to identify proteins proximal to a protein of interest in vivo and requires no outside input.

The work uses a non-specific bacterial peptide ligase to add a short peptide to proteins in an analogous manner to ubiquitination. Using a genetic fusion of a protein of interest to this ligase imparts specificity to the ligase activity and therefore label interacting partners of the protein of interest. We have engineered the peptide substrate to contain an affinity handle (TwinSTREP) to allow for purification and identification of interacting proteins by MS/MS or western blot.

Our main objectives are to:
1. Optimize and define parameters for discovery scale proteomic work in eukaryotes.
2. Perform proof-of-principle discovery screens for novel substrates of S-palmitoyl transferases in plants and yeast.

To optimize the system we will develop means to efficiently express the peptide tag. We will do this by using fusions to ubiquitin followed by self-cleaving viral 2a peptides to improve free peptide substrate production. We will assess efficiency of mature peptide production and processing by western blot.

Proof of principle screening will be performed using ligase fusions to 2 yeast S-palmitoyl transferases AKR1 and ERF2. Expression of the peptide tag and isolation of tagged proteins will be followed by MS/MS protein identification. Candidates will be expressed in WT and erf2/akr1 mutants as appropriate. S-palmitoylation state of candidates will be assessed by Acyl-RAC. Genuine substrates will show reduced S-palmitoylation in the respective S-palmitoyl transferase mutant . As proof of principle in plants we will perform pairwise screening of each of the 24 Arabidopsis S-palmitoyl transferases against the S-palmitoylated receptor-like kinase FLS2. This will identify which enzyme S-palmitoylates FLS2

Understanding how proteins interact with each other is critical to understanding how organisms function. This knowledge can be used to understand the molecular basis of how existing chemicals can affect organisms (medicinal drugs, toxins, herbicides, fungicides, etc.) and how to ameliorate or enhance their effects, how to design new pharmacological or chemical treatments, understand the basis of genetic diseases or how pathogens promote disease in humans or plants. This work is therefore of immediate and general use to the whole biological community, in particular fundamental researchers. However specific short to medium term non-academic beneficiaries include:

Pharma- and agro-chemical companies and associated industries
1. Bioactive compound discovery and validation - knowledge of how proteins interact and the outcome of disrupting a proteins interactions is a key factor in rational design of drugs or chemicals to control human and animal disease or combat pathogen infection of plants and animals. Discovering interactions of known drug targets provides mode of action while novel interactions provide information on future candidates for pharmacological disruption for medical or agronomic benefit. DogTag provides a means to identify and assess these interactions.
2. Difficult bioactive compound target validation and identification - many drugs have effects that are hard to determine due to the nature of the proteins affected. In particular membrane proteins are very hard to work with due to their physical characteristics. Membrane proteins are disproportionately important as drug targets (only 40% of proteins are membrane proteins but over 60% of approved drugs target membrane proteins) making their study of particular importance. DogTag allows for easy in vivo analysis of membrane protein interactions and will make future analysis faster and more reliable allowing for higher throughput work and the potential for greater knowledge about drug effects before trials. This can make drugs safer for the consumer by eliminating potentially harmful candidates and prevent poor drugs progressing through the expensive R&D process thereby saving money.
3. Transient or weak interactions - many biologically important interactions between proteins are weak or transient, such as those between enzymes and their substrates, making their identification by traditional means difficult. As DogTag reports on in vivo interactions it does not suffer from the limitations that hinder other methods; transient and weak interactions are therefore visible. This allows for target screening or assessing the efficacy of a compound on a particular interaction of interest such as blocking an enzyme modifying one of its substrates while continuing to modify others normally.

Staff employed on the project
Research Staff - Staff on the project will be trained in public speaking, presentation preparation, presenting data and information to expert and lay audiences, analytical processes, accurate record keeping and collaborative work. These are widely transferable skills applicable to all employment sectors.

Long term beneficiaries include
1. Policy makers - This work will help keep policy makers informed of progress towards drug discovery, alternative pesticide/herbicide development.
2. General public - work of this kind allows for the development of medicines, treatments and chemical control agents in a faster, more controlled and better understood manner. This will therefore result in higher standards of public safety and healthcare as well and helping to move food production towards reduced environmental impact and reduce harmful chemical use on food crops.

Timescales
Due to the fundamental nature of this work the initial impact will primarily be within academia. In the medium to long term (5-25 years) non-academic parties will likely benefit from translation of these important fundamental data.