Results

In this section we present the key results of the BananaGuard project (figure 1). By combining both experimental and modelling techniques, we have designed, implemented and tested a system that protects banana plants from Fusarium oxysporum. The BananaGuard system senses fusaric acid and, in response, produces fungal growth inhibitors to prevent infection of the banana plant. In addition we have also implemented a Kill-switch that disables our system when fusaric acid is not present. For this, we have optimized the circuit design, assessed the potential of BananaGuard in the soil, and analyzed the robustness of the system using different mathematical models.

Sensing

While fusaric acid dependent protection systems have been observed in micro-organisms and predicted in silico after genome sequencing, no evidence of a validated fusaric acid dependent promoter could be found in the literature. In this project we characterised and validated such a promoter. A putative fusaric acid dependent promoter along with its hypothesized regulator (isolated from Pseudomonas putida KT2440 (P. putida)) were cloned in front of GFP (BBa_E0040), transformed in P. putida KT2440. Promoter function was quantified by measuring fluorescence in the presence of different fusaric acid concentrations (figure 2). We were able to validate and characterize a novel fusaric acid dependent promoter (Bba_K1493000).

For more information, read fungal sensing.

Inhibition

Upon sensing fusaric acid, three genes and a gene cluster will be activated that will lead to production of fungal growth inhibitors. They were cloned behind an IPTG inducible promoter. Those genes and their function being:

1. phlABCDE gene cluster, able to produce 2,4-Diacetylphloroglucinol(2,4-DAPG)
2. Methionine-γ-lyase, Dimethyldisulfide (DMDS) and dimethyltrisulfide (DMTS)
3. PfrI, produce pyoverdine in presence of iron
4. Chitinase, overexpresses chitinase activity

Methionine-γ-lyase and PfrI were both made into BiobBicks, Bba_K1493300 and Bba_K1493200 respectively. With both BiobBicks validated, and for PfrI characterized. PfrI has shown to give a four fold increase of pyoverdine production in the presence of iron (31μM) in the growth medium. Pyoverdine is a compound that chelates iron and is naturally only produced when there is no iron available.

All transformants were co-inoculated with Fusarium oxysporum cubense TR4 on agar plates in order to test their inhibition ability. Controls used were plates inoculated with F. oxysporum in the presence and absence of wild type P. putida KT2440.

In general, it was hard to distinguish the increased inhibition effect of the fungal growth inhibitors producing P. putida against F. oxysporum. This is because the P. putida chassis we have chosen is already very good at inhibiting F. oxysporum naturally, which probably makes it hard to observe increased growth inhibition by our synthetic, growth inhibitor producing P. putida strains. However, with the Methionine-γ-lyase(MgL) strain, we have a clear indication that there is an enhanced growth inhibition of F. oxysporum (figure 4).We can say that there is an indication of a slight increase of growth inhibition, on top of the natural inhibitionin, the 2,4-DAPG, chitinase and pyoverdine producing strains. For more information, read fungal inhibiton.

Kill-Switch

Once fungal growth inhibitors are produced and F. oxysporum is no longer in the soil BananaGuard has done it's job and is no longer needed in the soil. Therefore we have implemented a Kill-switch into our system, which works like a toggle switch that senses when fusaric acid is around, and when it has dissipated toxins will be produced that eliminate BananaGuard. Toxins will be produced that eliminate BananaGuard itself, with the kill switch regulatory system in figure 4.

CIλ (induced by rhamnose) has shown to suppress the pcIλ/Tet promoter inhibiting GFP production when induced with rhamnose (figure 5). In addition, CIλ and lacI has also shown to suppress GFP production of the pCIλ/lacI promoter (figure 6). A toggle-switch was constructed ( Bba_K1493702, Bba_K1493703) containing pCIλ/lacI promoter + TetR together with pTet + LacI + GFP. After establishing that induction by IPTG leads to adequately low GFP expression (the off-state), whereas induction by aTc results in high GFP expression (the on-state).The non induced state of the toggle switch shows a fluorescence value between the two induced states, indicating that it might be stable. For more information, read Kill-switch.

In order to characterize promoters that are non-inducible we developed a new characterization method that is based on a rhamnose inducible promter. We used this system to characterize a promoter repressed cIλ repressor and LacI repressor (figure 7.) For more information, read rhamnose characterization.

Promoter design model

The kill-switch design is relatively intricate and therefore requires in silico analysis in order to test and improve its architecture. In order to perform such analysis we exploited statistical mechanics to derive a model of the promoter system. Not unexpectedly, the new insight obtained strongly favoured some adaptations to the current design, which included reallocation of promoters as well as parallel placement of an additional kill-switch, which according to the predictions would yield a more stable system. For more information, read model kill-switch promoter design.

Back to the lab

With new output from the promoter model, new promoters were made with different inhibitor binding sites with the BioBrick standard in mind. These promoters were then assembled in front of GFP and all 5 out of the 6 promoters we have designed have shown to work (Bba_K1493801, Bba_K1493802, Bba_K1493804, Bba_K1493805 and Bba_K1493806), by showing GFP expression (see figure 10). For more information, read promoter design results.

System model

Cost

Having the whole system in P. putida is great however there is always metabolic stress in everything that we want P. putida to produce. Therefore another model was developed to predict the cost of the whole system, using a genome-scale constraint based metabolic model. The model indicates that the metabolic stress introduced by fungal growth inhibitors production should not pose a bottleneck. For more information read system cost.

Performance

Knowing that P. putida is able to cope with the whole system, the next objective is to assess the performance of the system; will the kill-switch function according to our expectations? Will the kill-switch kill P. putida in advance of performing its intended role as a fungicide due to imbalance of the toxin anti-toxin system?In order to answer these questions we created a stochastic whole system model, incorporating metabolic stress, leakiness of each individual promoter and the toxin anti-toxin syswtem. The results of this analysis are depicted in figure 4.

The stochastic model (figure 11) has shown that different basal production levels of CIλ can have different effects on population dynamics, cell growth and the stability of the kill-switch, a point of attention for final construct of the system. Finally, the kill-switch will perform with 98% efficiency given the slow growth rate in the soil predicted by the metabolic model. For more information read model performance.

Green house

We established a collaboration with the Plant Research International group of Wageningen, which gave us the unique opportunity to test the system, not only against F. oxysporum, but in a setting that mimics the situation outside the lab as closely as possible with banana plants. At present we have banana plants in the green house (figure 12) that have been inoculated with our engineered P. putida and which were also infected with F. oxysporum. However, plants grow at a much slower pace than bacteria, so results were not possible to be obtained before the wiki-freeze (see green house ).

In short

We as an iGEM team have achieved quite a lot during these couple of months. Here is a short list of what we have achieved:

• Validated and characterized a novel working fusaric acid dependent promoter
• Proved that pyoverdine can be produced in an environment with iron
• Improve inhibition of P. putida towards F. oxysporum
• Show the proof of concept of the Kill-Switch using the input-output plasmid system
• Extensively characterized the rhamnose promoter
• Developed and used a new characterization method based on the rhamnose promoter
• Create a toggle switch that can be activated
• Combined modelling and wetlab results for promoter design
• Have made new promoters and validated their function for future use in iGEM!
• Built a model that predicts the metabolic cost of BananaGuard on P. putida
• Constructed a model that shows the performance of our whole BananaGuard system

In parallel with obtaining results from the lab, we have also been working towards implementing the BananaGuard system in greenhouse-grown banana plants. Will our engineered P. putida win the fight against F. oxysporum? Is it strong enough to survive in the complex soil rhizosphere? Will it save the banana plants? Sadly, the results were not here before the wiki-freeze, however we will present them at the jamboree! So stay tuned and come to our presentation!