Team:Wageningen UR/Results
Results
In this section we like to present you the main results of the BeeT project. BeeT is engineered to produce a toxin specific for Varroa destructor, produce the toxin on the right time and incapable of escaping the hive alive. To accomplish this, we performed multiple experiments and created different models. The outcome is in short shown on this page.
In order to improve on existing methods, BeeT should effect Varroa mites only. To accomplish this we decided to make use of Cry toxins. These toxins are naturally produced by Bacillus thuringiensis and because of this also known as BT toxins. A functional Cry toxin is only effective when specific binding occurs to the gut membrane of the target organism. Hereafter, the Cry toxins will form pores into the cell membrane, which results in cell death. As cell death occurs, the gut membrane becomes porous. Consequently, the organism dies. 1 To find a Cry toxin active against V. destructor we engineered our own toxins and we searched in nature for one as well.
Due to the parasitic nature of Varroa mites, testing Cry toxins proved to be very problematic. To overcome this problem we developed an in vitro test for Cry toxins. Out of the membranes of the target organism, brush border membrane vesicles (BBMVs) were made and incorporated with 6-carboxyfluorescein. A functional Cry toxins will create pores into the BBMVs, which then results in the leaking of fluorophores out of the BBMVs. Due to self-quenching behaviour of 6-carboxyfluorescein, this can be measured as an increase in fluorescence. As a proof of principle, BBMVs from the gut of Tenebrio molitor were made and loaded with 6-carboxyfluorescein to test the pore formation ability of Cry3Aa, which is known to be toxic to T. molitor larvae 3.
Figure 1a shows how the fluorescence increases of BBMVs incorporated with fluorophores in the presence and absence of Cry3Aa. A kinetic value could be coupled to this process. These values for multiple measurements for BBMVs in the presence and absence of Cry3Aa are shown in Figure 1b. From this can be concluded that the presence of a functional Cry protein results can be measured.
To engineer a Cry toxin active against V. destructor we took the Cry3Aa toxin as a starting point. This Cry toxin consists of three domains, from which one is responsible for the binding. In 1996, Rajamohan et al. demonstrated that mutations in the binding sites of Cry toxins can both decrease and enhance the specificity of a toxin towards its target organism2. Three putative binding sites have been identified after analysing the 3D structure of the binding domain of Cry3Aa. The putative sites were changed with random mutagenesis and the adapted proteins were cloned into E. coli. 144 Cry proteins were produced and tested for activity on BBMVs from V. destructor as described previously. From these measurements 24 candidates were selected to test further. The results are shown in Figure 2. From these results, it can be concluded that the third binding site (amino acids 410-416) seems to be a good candidate for future engineering and specificity adaptation of this particular Cry toxin. Due to the relatively high deviation in reaction speed for the toxins 3.3.3 and 3.3.7, these should not be taken into account, as they are rather inconclusive. This leaves us with one proper candidate – the toxin mutant Cry3.3.5.
In order to find the right specific binding motif, phage display was performed. Phages with a binding motif on their exterior were exposed to the gut membrane of V. destructor. Hereafter, the bound phages were isolated and analysed. The filamentous bacteriophage M13 was used with a 12-mer library (The Ph.D.™-12 Phage Display Peptide Library). The phages were fed to Varroa mites and exposed to BBMVs originating from Varroa mites. The recovered phages were isolated and sequenced. The consensus sequences of the binding motif of the 12-mer both in vivo and in vitro are shown in Figure 3.
Alongside creating a Cry toxin ourselves, we searched in nature for one as well. We gathered 800 death Varroa mites and looked for B. thuringiensis or related species inside these mites that might have been the cause the death. Figure 4 shows the morphology of B. thuringiensis and two found strains. Five out of 106 isolates were identified as Bacillus-like species. One strain, not B. thuringiensis, showed the presence of a large overexpressed protein and was sent for sequencing. We are waiting with excitement for the results.
A constant low level of Cry toxin can facilitate resistances11. That is why when BeeT spreads through the hive, the toxin production should be regulated. We created two main systems that regulate the toxin production. One is a system designed with riboswitches that promote toxin production when Varroa mites are present. The other system, that works in parallel with the riboswitches, uses quorum sensing to start toxin production only when the concentration of BeeT is high.
Riboswitches are pieces of mRNA that can regulate gene expression depending on whether it is bound to a ligand. The ligands for the riboswitches that were used here were guanine and vitamin B12. Both substances indicate the presence of Varroa mites. 95% Of the mite faeces consist of guanine. Vitamin B12 is present in the haemolymph of the honey bees, which is the food source of Varroa mites. Both riboswitches are successfully built into a construct in a way that when the ligand is present, toxin can be produced. Furthermore, they have been tested with RFP as reported gene in the presence of different concentrations of their corresponding ligand. The results for the vitamin B12 riboswitch are shown in Figure 5. As can be seen here, when the concentration vitamin B12 increases, the RFP production increases as well.
The second regulatory system uses quorum sensing. A quorum sensing mechanism enables the bacteria to regulate their expression based on their density. We adopted the lux system originating from Vibrio fischeri and demonstrated this system’s functionality using a newly constructed GFP reporter (Figure 6). When the cell density increases, the cells will sense each other’s autoinducers. These induce, via a complex formation, production of more autoinducers and GFP.
When the cell density is high enough, the quorum sensing system ensures that more and more toxin is produced. The down side of this is that this will likely kill the BeeT population. This is because the Cry toxin will lyse BeeT, when produced in very high concentrations. It would be beneficial to subdivide this population to keep healthy bacteria, as non-producers. These cells would be able to initiate a new growth phase after death of the toxin-producing cells. The critical requirement for this is that cells respond to the stimuli at different times despite being genetically identical. To create such a system we used two proteins: the first encodes for the protein that inhibits the toxin expression, whereas the other promotes toxin expression. The protein that has the upper hand, determines whether toxin production is on or off. Both proteins are encoded behind the same promoter, however, one of the proteins has a higher turnover rate. The trick is to find the “sweet spot” of the translation rates, at which in some cells one protein takes the upper hand and in other cells the other protein. This sweet spot has been found with a model. Figure 7 is produced by the model and shows the presence of two different subpopulations.
To elaborate further on the toxin production switches, we wanted to add a toggle switch to the system. Looking upon the fact that BeeT might not grow very well in beehives, we decided to couple the toggle switch with a riboswitch. The potential slow growth rate is in disadvantage for the quorum sensing system, because they may not be able to grow to the density required for toxin production. Whereas the riboswitch is not dependent on population density.
The toggle switch we created controls expression of the BeeT’s toxin between an off-state and an on-state. It is switched on by guanine or vitamin B12 and switched off by blue light. The latter combines the optogenic kill switch, which will we explained later in more detail, with the toxin production. In order to create the system, a new hybrid promoter had to be made. The hybrid promoters ensures that toxin production is only possible in the dark. Figure 8 shows the testing results of 5 different hybrid promoters. From this can be concluded that the hybrid promoter BBa_K1913022 functions the best. Although we did not have time to test the system as a whole, we expect it to work since both the riboswitches as the hybrid promoter are functional.
BeeT is intended to use in beehives where bees fly in and out continuously. Hereby it can be spread throughout the environment. Since we cannot be sure about the effect on existing ecosystems if BeeT would be released in the environment, it must be engineered to die if it leaves the beehive. To accomplish this we made use of a light kill switch and a cas9 kill switch.
The optogenetic kill switch is the unification of two different genetic systems: pDawn, an artificially-created promoter system activated by light; and MazEF, a toxin-antitoxin system native to E. coli . The antitoxin, MazE is constituitively expressed. The toxin MazF is only expressed in the presence of light, because MazF is regulated via pDawn promoter system and an invertor system. This means that in the darkness of the beehive - we confirmed instrumentally that the blue-light irradiance in a beehive is practically zero - no toxin is produced, allowing the cell to remain stable. However, in the sunlight, toxin production takes the upper hand and the cell dies. In Figure 9 is demonstrated that the pDawn promoter system works. Alongside the artificially-created promoter system pDawn, we tested another one pDusk, which gave negative results in the lab.
The artificially-created promoter system pDusk and pDawn were modelled in Matlab together with the MazEF toxin-antitoxin system. Within our parameter estimation procedure we found two parameter sets, out of 1000 sampled sets, which satisfy the conservative constraints described in the optogenetic kill switch modelling section. We fitted the model to literature data and thus can conclude that our model describes the behaviour in the lab. The results from those two sets can be seen in the animated Figure 10. Taking a closer look, it can be seen that it takes a few hours for the MazF toxin to take the upper hand in the pDawn system. This means that beekeepers can open and work with their beehives, without immediately destroying the BeeT.
To bulletproof our biocontainment strategy, we added an additional kill switch. As a chassis for BeeT we want to use a bacterial strain developed by Mandell and colleagues (2014)1. This “biocontainment strain” is auxotrophic for a synthetic amino acid, para-L-biphenylalanine (BipA). Several essential proteins of this bacterial strain were engineered to function only when BipA is incorporated in the active site, leading to death of the bacterium when the synthetic amino acid is not available. In case of BeeT, BipA should be applied to the beehive, which will be the only place BeeT can survive given that the beekeeper continues to supply it. To prevent horizontal gene transfer, we proposed to extent the biocontainment strain by creating a switch to cleave heterologous DNA. We aimed to cleave this DNA with a modified Cas9. In our case, when BipA is present, the synthetic amino acid will be built into the active site of Cas9, making it catalytically dead. However, in the absence of BipA, the native amino acid will be used, making the Cas9 catalytically active. This active Cas9 will cut heterologous DNA. Figure 11 shows the Cas9 incorporated with BipA. Unfortunately, we were not able perform in vivo experiments with this Cas9.
Testing BeeT in a Beehive BEEHAVE
Ideally we want to test BeeT in a beehive. This is, however, not an option. Bringing genetically modified organisms free in the environment is far from responsible, moreover forbidden. Because of this, we had to find an alternative way to test BeeT. First we proved with an experiment and with a model that BeeT can survive in the sugar water, via which it will be applied to the bees. Secondly we modelled the influence of BeeT in an open source model called BEEHAVE. We adapted the model in a way that it could predict what the effect of BeeT on virus epidemiology, mite population dynamics, and bee population dynamics is.
