A creative project is a moving target. You never end up where you start.- Evangeline Lilly
BoNT C - Licence to enter
In modern medicine treatment options involve many substances modified from natural sources, occasionally even toxins. We modify botulinum toxin in a way that leads to its detoxification. Thus, it can be coupled with a variety of other substances while not losing its specific shuttle mechanism for neuronal cells. In detail, we develop a library of different detoxified botulinum toxin derivatives which can accommodate other proteins, small molecules, and fluorochromes by specific linkers. To investigate the influence of the point mutations leading to detoxification in the active site, we conduct MD simulations. Since our shuttle mechanism could potentially be used in patients, we remove the most prevalent immune epitopes by a theoretical bioinformatics approach. Ultimately, our system is supposed to be utilized for therapeutic strategies and specific neuronal targeting in basic research. With our project, we want to encourage future teams to think outside the box while keeping safety in mind.
Botulinum Neurotoxins (BoNTs) are proteins produced by the bacterium Clostridium botulinum. These proteins are very potent neuro-toxins which prevent muscle contraction via inhibition of motor neurons. The bacteria themselves can be mostly found in forest soil and start to proliferate under anaerobic conditions. These conditions are met after uptake by animals in their intestine. After release of the toxin, infection of motor neurons and finally the animal’s death, bacteria proliferation even increases in the rotting process of the carcass. After uptake of the rotten flesh by scavengers the cycle begins anew. For humans, botulism - the disease caused by BoNT - becomes mainly a threat when tinned food is insufficiently conserved which provides an anaerobic environment for the clostridia. Left untreated, the disease can lead to respiratory paralysis and death. The BoNTs belong to the family of AB toxins and can be divided into the serotypes A-H. As an AB-toxin, BoNT consists of two protein chains, the heavy chain (HC) and light chain (LC), which are connected by a highly conserved disulfide bond and multiple noncovalent interactions. The mechanism underlying the cell specific infiltration by BoNT is partly understood, yet many questions remain: For cell recognition, BoNT uses gangliosides in the presynaptic membrane to select for neuronal cells (Benson et al 2011). Besides the gangliosides themselves, which are enriched in neuronal cells, binding of BoNT requires a linkage between gangliosides and N-acetylneuraminic acid (Sia5) (Strotmeier et al 2011). The subsequent infiltration process is performed by the HC binding to the membrane and further unknown receptors. HC and LC then undergo endocytosis until only the LC is released into the cytosol. Lastly, the LC, being a zinc metalloprotease, cleaves Snap-25 and other proteins of the SNARE complex depending on the toxin’s serotype. This leads to prevention of neurotransmitter release into the synaptic cleft and therefore muscle contraction at the neuromuscular end plate. Since BoNT is so highly specific for neuronal cells, its potency and therefore toxicity is among the highest for peptide toxins in the world. For BoNT serotype C the lethal concentration (LD50) in humans is approximately 1ng/kg (parenteral application) and 1µg/kg N (oral application). The high toxicity would therefore prevent from any usage of the specific shuttle mechanism. In 2017 three missense mutations in the active site of the LC were reported that conserve the overall structure and binding affinity of BoNTC, however reduce the activity and overall toxicity by the factor 106. (Vazquez-Cintron et al 2017) This option of an efficient detoxification provides us with a very powerful tool that might be used to transport a variety of molecules into neuronal cells with a high specitivity: Synthetic drugs, therapeutic peptides and labeling proteins for fundamental research imported this way might revolutionize neuronal targeting strategies to increase our overall knowledge about the complexity of the brain and the chances to treat severe neuronal diseases.
The neuron-specific shuttle system of aToxic BoNT offers a wide range of application possibilities, so we developed a modular toolbox to offer a wide range of options for the user to choose from. In our toolbox we coupled various parts of different sizes, chemical properties and functionalities to show the dimensions our shuttle-system offers the science community.
Since different parts require different linking strategies we suggest three different linkers that give the option for individual design and personalized adaption to the desired field of application.
This mode of coupling is specialized for proteins and peptides. The target protein is coupled to the N-terminus light chain (LC) of the aToxic BoNT via a series of Glycine-Serine (GS) repeats. Additionally, we added prolines to the linker to generate a possible protease recognition site. Through this site the cargo can be separated from the LC in the target neuronal cells to fulfill its desired function.
We furthermore introduced a chemical linker into the aToxic BoNT to connect cargo with the help of click chemistry. This has the advantage of a user friendly one pot reaction with which one can attach a wide range of molecules and proteins to the LC of our shuttle. This was made possible using the expanded genetic code. We first introduced an amber stop codon into the light chain on the positions 216 and 423, respectively. This codon is recognized by the engineered tRNA aminoacyl-synthetase PyIRS, which then, during transcription, incorporates propargyl-lysine into the LC of the BoNT protein. Propargyl-lysine carries an alkyne functional group, which usually does not occur in proteins. Thereby, it is perfectly suited to perform biorthogonal chemistry. Via copper(I)-catalyzed azide-alkyne-cycloaddition (CuAAC, click reaction) any cargo with an azide functionality can be covalently bound to the shuttle.
This method involves the coupling of the target protein and the BoNT-shuttle by the enzyme sortase. If the desired target protein contains a sortase recognition sequence, it can be coupled to the atoxic BoNT-shuttle. To realize this, we included an N-terminal Glycine, which is needed by the enzyme to fulfill the reaction. To couple the target protein to the BoNT-shuttle, both are incubated with the sortase enzyme which then can link both together
Planned Fusion Partners
The cyclic recombinase (Cre) is a tyrosine recombinase which can carry out site-specific recombination events of DNA sections labeled with LoxP recognition sites. Since the Cre recombinase is a widely-used tool in molecular biology to specifically carry out gene knock-outs, knock-ins or translocation experiments, we decided to include it into our toolbox. Furthermore, we decided to chose a Cre recombinase from the iGEM distribution kit from 2013 (BBa_K1680007).
Omomyc is a dominant-negative mutant of Myc, a family of proto-oncogenes encoding for transcription factors which regulate growth, proliferation, tumorgenesis, and apoptosis. In cancer Myc often is constitutively expressed, leading to increased proliferation signals and thus to excessive tumor growth. Through its ability to induce apoptosis in cells overexpressing Myc, as various cancers do, Omomyc may be subject to new therapeutic opportunities in the future.
Learn more about omomyc.
pHluorin2 is a ratiometric, pH-dependent GFP. Its excitation spectrum varies as the pH increases/decreases. This allows pHluorin2 to be used as an accurate biosensor. A special use case is the tracking of proteins that move between different cell compartments and encounter varying pH environments.These properties made pHluorin2 very interesing for our toolbox
Learn more about pHluorin2
Eslicarbazepine is a third-generation pharmacological drug, used for the treatment of epilepsy. It blocks voltage-depended sodium channels by enhancing channel slow-inactivation. This leads to less sodium ions to enter the cell and consequently makes it less excitable. We chose this drug, since it has a simple and good electrophysiological read-out via the patch clamp method. Secondly it specifically targets cells with a voltage-dependent sodium channels. Our shuttle guarantees the specific distribution of the drug at its site of action (the neuron) which prevents possible side effects caused by off-target reaction.
Fluor488 is a phluorophor which we planned to couple via click chemistry to the atoxic BoNT shuttle. This will be a proof of our concept and furthermore give new possibilities to label neurons fluorescently.
Nano-Glo® HiBit Extracellular Detection System
The Nano-Glo® HiBit is a two component-system consisting out of a small 11-amino acid peptide (HiBit) and a protein (Large bit). Upon interaction of the two elements a form a protein with luciferase activity. Using this system has several advantages. Firstly, it is easily applicable, highly efficient and fast to implement. Furthermore, it offers the possibility to detect proteins with high sensitivity in vivo in real-time. This system is usually used to quantify HiBit-tagged proteins expressed on the cell surface. We however planned to transfect the neuronal cells with the Large Bit and to bring in the small HiBit part into the cell with the help of the BoNT-shuttle. The read out over luciferace activity would be another proof of our concept, to show that BoNT is able to enter a neuronal cell with a small peptide attatched to it.
The main method for our cloning work was the Gibson Assembly. This method allows to include up to six inserts into one vector. It has the possibility to build larger DNA fragment or even whole genomes. One should mention some arrangements, which had to be banned before the work could start. The used dsDNA fragments needed to have complementary overhangs with adjacent fragments. This was essential to anneal the fragments in the right order during the Gibson Assembly. Furthermore, the secondary structure should be clean, and hairpins avoided. Additionally, the overlaps should be unique and should exist in a length of 15 to 80 base pairs.
For a good Gibson Assembly, a good and clear preparatory work is very important.
Our first task was to get a mutated Light Chain (LC) from the pH6-t-BoNTCs-L plasmid we were given by professor Thomas Binz. Therefore, we dismantled the light chain in three parts. The complementary and adjacent fragments included the necessary mutations. Thus, the first Gibson Assembly consisted of three insert and one vector, the pet28a. After the transformation, we got the whole mutated light chain plasmid. The atoxicity was verified through a toxicity assay.
The next step was to insert the heavy chain, the mutated light chain and the respective part into our vector via Gibson Assembly. The specific fragments were connected to each other and combined to a new plasmid.
There are some advantages why this cloning method is commonly used in science. One big advantage is that the Gibson Assembly is very fast and efficient. Multiple fragments can be performed in one single-tube reaction. Another big advantages is that it does not yield adverse cuts in the DNA sequence whereas the restriction cloning does.
Obtaining purified and stable proteins is an important cornerstone for any protein-based experiment. For this reason, we designed our various protein parts with the most reliable tag variants available, the His-tag and the Strep-tag.
In this method, the DNA sequence of the protein to be isolated is supplemented at the C or N terminus by the codons of a His tag (His linker, consisting of six histidine residues).
This His tag binds specifically to a column material with divalent nickel ions. The Ni2+ ion is bound in an octahedral complex with the carrier material and water molecules. These can be displaced by histidine, resulting in a very stable chelate complex. Ni2+ is bound on these columns by nitrilotriacetic acid residues and can interact with two histidine residues of the protein in exchange for water. This specificity ensures that only the fusion protein binds to the column material.
The fusion protein is eluted with imidazole, which competitively displaces the histidine of the protein from the chelate complex. This allows us to obtain proteins that only contain the polyhistidin-tag and thus a purified protein.1
The purification of Strep-tag II fusion proteins begins by applying the cell lysate with the fusion protein to a column of immobilized Strep-Tactin. The fusion protein then binds to Strep-Tactin. This is followed by a washing step in which all unbound proteins are washed off the column with a physiological buffer (e. g. PBS). The fusion protein is then eluted with a low concentration of desthiobiotin, an analogue of biotin, the natural ligand of streptavidin. Desthiobiotin competes with high affinity specifically for the biotin binding site. It has the advantage over biotin that it binds to strep tactin with a lower affinity and can therefore regenerate the matrix.2
In addition, we opted for an expression system based on T7 polymerase in the bacterial strain E. Coli Lemo21(DE3). This E. Coli tribe was deliberately chosen. It is known for its good expression rate of toxic proteins.3 We chose the T7 expression system because the T7 promoter and T7 polymerase were very well suited for our recombinant proteins.
The system is structured as follows. Behind the T7 promoter lies the coding region for a "protein of interest". However, T7 polymerase is a foreign enzyme and is therefore not expressed under normal circumstances. In addition, the sequence for the T7 promoter specific T7 polymerase was integrated into the E. coli strain using the DE3 prophage. This can be recognized by the bacterial strain in the designation DE3.
The expression of T7 polymerase is controlled by the lacUV5 promoter, which can be induced by IPTG or lactose.
Since neuronal properties are very complex and our shuttle system requires highly specific cell parameters to infiltrate cells, we were in need of a cell model that would be easy to handle and could simulate the neuronal phenotype at the same time.
In a suitable cell model we were looking for the following cell parameters: 1. easy accessibility 2. high reproduction rate 3. human origin 4. neuronal phenotype and/or origin 5. high sensitivity to botox and preferably BoNTC (Purkiss et al 2001) 6. well established in scientific research
These requirements left us with a choice between human pluripotent derived stem cells (iPS) or stable glioblastoma cell lines. While it takes a lot of time, money and experience to produce iPS-derived neurons and the outcome can be unreliable, glioblastoma derived cell lines provided a robust human cell model that featured enough neuronal characteristics to serve as a readout model for our project. After looking into a lot of different glioblastoma lines we decided to work with the cell line SHSY5Y. These cells, originally derived from the neuroblastoma of a five-year old female, feature a lot of characteristics common with primary human neurons: eventhough they’re unable to build synaptic connections with each other, the cells have a similar composition of membrane proteins and lipids needed for infiltration by BoNTC (high percentage of gangliosides), are positive for hydroxylase and can show a acetylcholinergic, glutamatergic or adenosinergic phenotype. Most importantly, SHSY5Y was shown to be affected by BoNTA (Purkiss et al 2001) and BoNTC (Rust et al 2016). Additionally, the cell line is widely used in neuroscientific research, can be differentiated to gain more neuronal characteristics in the formation of dendritic connections (Shipley et al 2016) and proliferates mainly via activation of the transcription factor Myc (Savino et al 2011). Therefore, it can be brought to apoptosis by the the anti-oncogene synthetic peptide Omomyc, that was planned to couple to BoNTC. Based on these known parameters, we decided to use SHSY5Y cells as our readout system for the efficiency of the BoNTC constructs in a undifferentiated and differentiated state.
For our alkynyl-botulinus toxin we planned to use a fluorescent molecule as a proof of principle for bioorthogonal conjugation. We already tried to create a drug-conjugate useful for further investigations by using the Na+-channel blocker Eslicarbazepine.
Since there are already commercially available azide-linked fluorophores we saw no need to functionalise them further. We obtained Azide-Fluor488 as a kind sponsorship from Sigma-Aldrich. You can see the azide-linked fluorophore in the next picture.
To see if a site-specific modification is indeed beneficial, we also employed a BODIPY-fluorophore as NHS-ester to nonspecifically modify solvent-exposed lysines of our botulinus toxin. This NHS-ester is also commercially available and was obtained as a sponsorship from Lumiprobe. You can see the aBODIPY-fluorophore as NHS-ester in the next picture.
Having an actual drug as a conjugate is more complicated than the use of a fluorophore by several orders of magnitude. For instance a 50 kDa peptide covalently linked to a 250 Da small molecule would likely inhibit it’s function to act on the target proteins. Therefore we chose to investigate a cleavable linker. As already applied by the group of Peter G. Schulz in 2015 we chose a disulfide linker that would be cleaved by the intracellular glutathione. Therefore, the general strategy from Eslicarbazepine was: R-OH to R-SH to R-SS-R’-NH3 to R-SS-R’-N3 to create an azide group for bioorthogonal chemistry linked to our drug via a disulfide linker.
A method for a one-pot synthesis of sulfhydryls from hydroxy compounds has already been published. Therefore we planned to follow their strategy to employ Lawesson's reagent in Dimethoxyethane. We then wished to use a disulfide exchange to create a mixed disulfide using 2,2'-Dithiodiethylamine. A similar reaction has been used to create a disulfide linker with antibody drug conjugates. An alternative option was found with S-tosyl mercaptoethanolamine. A similar reaction has been used in the total synthesis of ajoene in 2018.
Azide-Donor For the final step we would have to change the amine group to an azide group to enable this compound for azide-alkyne click reactions. These transformations usually involve rather dangerous substances, such as triflyl azide or imidazole-1-sulfonyl azide which is known to be explosive. The latter of which, however, can be prepared as a hydrogen sulfate as a stable and safe reagent. As this compound is to our knowledge not commercially available for a reasonable price, we decided to synthesise it ourselves from imidazole, sulfuryl chloride and sodium azide.
Working with genetically engineered organisms demands special awareness for safety of humans and environment. Therefore, we had to assess possible risks while working in the lab. We had to constantly adjust assumptions we made during the project to ensure the safety of all people working on our project and the environment which could be affected by our work. This included the necessity of planning every step of our research ahead of time and coordinating the work, respectively. To follow Good Scientific Practice and federal regulations, we profoundly researched all available information about our used gene blocks and possible complications we might face in the course of our work with genetically modified organisms. To minimize evolving risks, we decided to limit ourselves to tools, parts, and organisms within the Biosafety level 1 and therefore also to limit our work to a BSL 1 laboratory. Before we started our work, we contacted the representative for biological safety and security of the University of Tuebingen, Dr. Kittel, to confirm that we follow all necessary provisions and keep our work at BSL1.
Working in the lab
Our project demanded all kinds of diverse expertise. As our team consists of different backgrounds and very individual talents, we decided to allocate the different tasks accordingly. We did not only rely on the expertise of our group members but also contacted and cooperated with various professors and workgroups. The cooperation started with providing materials and lab space and included support for both theoretical and practical procedures. Luckily, we could organize different parts of the project in separate labs to react on different needs according to the subprojects. For our chemical synthesis, we joined the Lab of AK Maier located in the Department of Chemistry in Tübingen. They provided us with all required equipment for our chemical synthesis. In addition, we were pleased about our close cooperation with the Gust lab.