Team:Montpellier/POC
PROOF OF CONCEPT
While modelling in silico the activity of KARMA, we predicted that compared to a protease alone, KARMA was much faster and more efficient when degrading a specific substrate. Encouraged by these in silico results, we endeavoured to test KARMA experimentally.
Our proof of concept aimed to demonstrate improved specificity of a protease by fusing it to a specific antibody. Specifically, we wanted to limit degradation of all non-target proteins by reducing the intrinsic activity of a protease, then add a target-specific antibody in order to maintain an effective activity against one specific target protein.
We first aimed to test if fusing a protease to a VHH increases the overall degradation rate of the VHH’s target as compared to the protease alone. Then, we tested if a mutation-induced reduction in protease activity can be recovered by a VHH-directed increase in specificity. Finally, tested the specificity of our tool by measuring its activity against a non-target protein.
We designed a reporter assay that allows us to easily and quantitatively measure the enzymatic activity of our protease in vivo via variations in a fluorescence signal. Upon expression of our reporter constructs in E. coli, we acquired activity data directly by measuring fluorescence signals using both a plate reader and a flow cytometer.
Composition of KARMA for this proof of concept
In general, KARMA is composed of two parts, the recognition part (the antibody) and the scissor part (the protease). In order to target a specific protein, we must carefully choose the antibody and the protease, but also a linker that fuses these two components without changing their native conformations, so that they maintain their individual functions. Additionally, in order to maintain the integrity of KARMA, the protease shouldn’t cleave or degrade itself nor the antibody.
For this proof of concept, we work with a well-known protease, the TEV protease, which cleaves a specific sequence of amino acids. It was important to have a specific cleavage site in order to simplify the measurement of the protease action on this particular site. Concerning the antibody, we chose a VHH for reasons that we expose below.
The TEV protease
The Tobacco Etch Virus (TEV) Protease is a site-specific protease found in the Tobacco Etch Virus. The optimum recognition site for this enzyme is the sequence Glu-Asn-Leu-Tyr-Phe-Gln-(Gly/Ser) [ENLYFQ(G/S)], also called TEVcs (TEV cleavage site), cleavage occurs between the Gln and Gly/Ser residues.
In order to reduce nonspecific cleavage of untargeted proteins (proteins who also contain the TEVcs but that are not targeted by our antibody) we worked with TEV mutants with hindered activity: TEV PE10 and TEVPH21. The optimal temperature for a wild-type TEV’s cleavage is 30°C, we worked at 37°C to reduce the activity. A low activity would also make it easier to observe an improvement of the enzymatic activity.
The MBP protein
During brainstorming sessions, one of our instructors advised us to express the TEV along with a Maltose Binding Protein (MBP) that would facilitate the TEV’s folding. Indeed, MBP increases the solubility and stability [1] of TEV when expressed together.
The VHH antibody
The objective of KARMA for this proof of concept is to improve the cleavage of the TEV protease on a particular cleavage site (TEVcs) that is placed near the GFP (Green Fluorescent Protein). We looked for an antibody that targets the GFP protein, so that when the antibody is bind to GFP, it brings the TEV protease closer to the TEVcs. This would theoretically improve the cleavage activity in all TEVcs near to GFP.
There are many types of antibodies that we could have used (scFV, Fab, igG, VHH, etc). In order to be very specific, we decided to use a VHH. A VHH antibody (or nanobody) is the antigen-binding fragment of heavy chain only antibodies. They come from camelids and they have are smaller, more stable and much more specific than other types of antibody. Moreover, VHHs do not require any post-translational maturation or modification, so it is easy to express functional VHHs in E. coli. Their little size is also an advantage, as it minimizes the steric hindrance that might interfere with protease activity. The sequence of the anti-GFP VHH that we used was taken from an article (Kubala et al. s. d.) [2]
The linker
The Linker is a key component of our tool, the optimisation of KARMA requires us to find the best possible linker to fuse TEV and the VHH without disturbing their individual functions. Several questions arose during the design, should the linker be rigid or flexible? Which residues is it better to use? And especially which linker length?
Thanks to mathematical modelling and the 3D visualization of chemical structures in software such as Pymol, we were able to compare several linker sizes. We were not sure of the optimal size, so we selected 4 possible linker lengths (11.4 nm, 9.5 nm, 7.6 nm and 5.7 nm) in order to make a small library and perform a functional screen. This would allow us to determine experimentally which linker size was the best.
We chose a well-known linker that is composed of a tandem repetition of a sequence of 4 glycines and 1 serine (GGGGS), we repeated this motif several times until we reached reach the desired linker sizes (3 to 6 repetitions). This linker model was proposed during a meeting with Dr Pierre Martineaux from the "Criblage functional et ciblage du cancer" team at the Montpellier Cancer Research Institute. We thank him for sharing his expertise and offering us his precious help.
Two strategies for testing KARMA
In order to ensure that our results are actually accurate and trustworthy, we decided to perform two strategies in parallel. Both compare the action of KARMA and the TEV alone, the difference lies in the reporter that allows to measure the TEV activity.
First strategy: Proteolysis tag removal
This strategy relies on the ability of the endogenous ClpXP protease to recognize and degrade protein tagged by a specific signal motif called ssrA in E. coli. We build a fluorescent reporter in which a sfGFP gene is linked to a ssrA tag preceded by a TEV cleavage site (sfGFP-TEVcs-ssrA). TEV is a site-specific protease cutting with high specificity at a specific location within its cognate cleavage site.
Fusing an SSRA tag to sfGFP results in lower cell fluorescence signals because the fluorescent protein is degraded faster and do not accumulate in the cell. In the presence of TEV, the ssrA tag can be cleaved from sfGFP before it is processed by ClpXP, thus yielding a more stable protein and increased fluorescence. This implements a kinetic competition between degradation of sfGFP by ClpXP and cutting of the ssrA tag by TEV. In this framework, increased TEV activity results in higher cell fluorescence.
This idea was inspired by one of our instructors’ work, Hung-Ju Chang, who had worked on the recovery of fluorescence due to the action of the TEV protease.
Second strategy: Quencher removal
In this design, the ssrA is replaced by a fluorescence quencher called ShadowG. The proximity of the quencher is expected to reduce the fluorescence intensity of sfGFP. Cutting of TEV at its cutting site should dissociate the quencher and enable effective fluorescence of the freed sfGFP. Here again, more active TEV protease can process more quenched protein and result in the observation of higher green fluorescence intensity.
In both strategies, the objective is to monitor the enzymatic activity of the TEV protease and that of KARMA (TEV-antibody) by an increase of green fluorescence. This gain comes after the separation of GFP from the proteolysis tag or the quencher.
The utility of KARMA being its SPECIFICITY, we measured unspecific cleavage by replacing the sfGFP gene by an mRFP1 gene. The RFP protein is completely unrelated to GFP so it should not be recognized by the anti-GFP VHH. We further checked the absence of that epitope by amino acid sequence alignment . Thus, any increase in red fluorescence signal would correspond to the background activity of the TEV protease not enhanced by specific targeting.
Expected results
For both strategies, we expect that in comparison to TEV alone, KARMA will produce a gain in green fluorescence (specific cleavage) while maintaining or reducing the red fluorescence (nonspecific cleavage).
If this happens, it means that the an anti-GFP antibody improved the specificity of the TEV protease. By binding specifically to the targeted sfGFP, KARMA’s antibody would increase the chance that the TEV protease meets and cleaves the TEVcs placed next to the sfGFP.
[1] Raran-Kurussi, Sreejith, et David S. Waugh. 2012. « The Ability to Enhance the Solubility of Its Fusion Partners Is an Intrinsic Property of Maltose-Binding Protein but Their Folding Is Either Spontaneous or Chaperone-Mediated » éd. Bostjan Kobe. PLoS ONE 7 (11): e49589.
[2] Kubala, Marta H., Oleksiy Kovtun, Kirill Alexandrov, et Brett M. Collins. 2010. « Structural and Thermodynamic Analysis of the GFP:GFP-Nanobody Complex ». Protein Science 19(12): 2389‑2401.
[3] Fernandez-Rodriguez, Jesus, et Christopher A. Voigt. 2016. « Post-TranslationalControl of Genetic Circuits Using Potyvirus Proteases ». Nucleic Acids Research 44(13): 6493‑6502. Sarah Guiziou et al. 2016. “A part toolbox to tune genetic expression in Bacillus subtilis” Nucleic Acids Research, 2016, Vol. 44, No. 15 7495–7508.
[4] Murakoshi, Hideji, Akihiro C. E. Shibata, Yoshihisa Nakahata, et Junichi Nabekura. 2015. « A Dark Green Fluorescent Protein as an Acceptor for Measurement of Förster Resonance Energy Transfer ». Scientific Reports 5(1): 15334.