Protection

Design

After P. infestans penetrates the cell wall of potato, it will exploit the potato and in turn infect potatoes nearby within 3 days. Since it will infect other potatoes in such a short time that there is no effective biological method to react and inhibit the development of the disease, we decide to prevent the disease at the very beginning by rendering the potatoes the ability to prevent the invasion of P. infestans. Under certain conditions, the zoospores of P. infestans will attach to the surface of potato leaves, penetrate the cell wall by high turgor pressure and some enzymes, and secrete some effector protein, such as Avr3, into potato cells. The effector protein needs to bind to a transmembrane receptor called PI3P, which can mediate its entry into potato cell to translocate into host cell. It will suppress plant resistance gene-based immunity so that P. infestans can enter potato cells without any resistance. To stop the effector protein from entering potato cell, we found through literature research that reduction in translocated effector is a promising way to decrease the virulence of pathogens and improve disease resistance in potatoes. In research done by other scientists, FYVE protein domain from Hrs or EEA1 can also bind to PI3P receptor strongly in animal cells. We then decided to construct a FYVE protein domain with high affinity that can compete with the effector protein to inhibit the entry of P. infestans.

FYVE protein domain

FYVE protein domain is well conserved PI3P binding domain in various organisms with only 141 amino acids. FYVE protein domain originally existed in EEA1 and Hrs proteins in human and mouse, respectively. However, EEA1 and Hrs protein are too large and may take long time for the plant to degrade, we then decided to extract the protein domain from Hrs protein. On the other hand, monomeric FYVE has far lower affinity to PI3P than Hrs and it is not so stable. Therefore, we decided to construct a dimeric FYVE which has a higher affinity and is much more stable than monomeric FYVE.

Promoter choice

Though the depletion of PI3P may be an effective way to prevent p/infestans from invading the potato, the constitutive expression of FYVE protein domain may cause some physiological effect to potatoes since PI3P is an important receptor to plants. Thus the timing of the expression of FYVE is the key to successful implementation of this technology.

We choose a promoter, Gst1, in potato that will be activated within 24 hours when p.infestans infect the potato and the production of the promoter will down-regulate after the infection. Also, this promotor will be activated mildly when the plant is wounded, in which the plant is vulnerable to late blight.

Circuit design

Our goal in this part is to inhibit the entry of P. infestans effector protein, so the dimeric FYVE will be constitutively expressed in plant cell. The circuit will have a viral constitutive promoter CAMV35S following the coding sequence of dimeric FYVE.
The circuit is shown on the right.

Experiment

To check whether dimeric FYVE, a protein domain in mice, will work well in plant, we fused GFP with dimeric FYVE in vector pSAT1-Venus-C and transfect the plasmid into a tobacco cell, BY-2. If the dimeric FYVE works well, we can see green fluorescence on the endosome of BY-2.

Detection

Introduction

One major problem in controlling potato late blight is that there is no simple and convenient way to detect the disease. If we want to make sure that our potato has not been infected by P. infestans, we have to examine the potato in the laboratory for approximately a week. We want our detection system to be simple and convenient, and can instantly report the infection to the user. So the solution we reached is to design a soil based microbial fuel cell (SMFC) that can detect salicylic acid, a chemical that is produced when potato is injured or infected. With this device, we can know immediately at home whether the potato is infected.

We choose to utilize the Mtr (metal reduction) pathway of Shewanella oneidensis MR-1 to build our SMFC. Shewanella oneidensis MR-1 is a gram negative bacteria that is widely used for constructing microbial fuel cells because how it produce electricity is well characterized. The Mtr pathway contains 4 proteins: CymA, MtrA, MtrB, MtrC. CymA is a transmembrane protein that can transport electrons out of the cell membrane, and can then activate proteins mtrA, B, and C consecutively. From literature research, mtrB gene plays a pivotal role in stabilizing other component in this pathway. Therefore, in our project we utilize mtrB to create our biosensor by detecting changes in electric signals.

Design

To create this long term biosensor, we first knock the endogenous mtrB gene in Shewanella oneidensis MR-1. By reintroducing mtrB gene under the control of sensor (nahR), we can control the bacteria to generate electricity when it detects salicylic acid. We managed to design our sensor system using mtrB. However, the mere expression of mtrB is not enough. As electric signals in the soil can also be detected by the soil-based microbial fuel cell we created and can constitutively produce electric currents. This electric signal caused by soil itself somehow come as a background noise. When the electric signals emitted by the plant is not intense enough or when the number of bacteria in the soil drops, users may easily confuse it with the signals produced by the soil and make wrong judgments on pathogen control. Thus, we incorporate the oscillator into our circuit design. Even when the currents are not strong enough, users can still easily tell the difference between the electric signals emitted by the soil and those by the infected plant by recognizing the oscillating pattern of the latter. (See our modeling page)

Main circuit

As the circuit showed below, we first tested the oscillator and sensor with GFP as reporter. Then we replaced GFP with MtrB to test the expression of MtrB will also oscillate.

Cure

Overview

Our goal is to kill P. infestans without harming the potatoes and environment. Fungicides used to kill P. infestans nowadays contain lots of heavy metal, such as Cu2+ that will do harm to both the plant and environment. At first we try to use some antimicrobial peptides; however this peptide will not only kill P. infestans but do harm to potatoes as well. Therefore, we try to look for the defensin that can weaken or even kill P. infestans from other plants. Although there are other chemicals that might be effective in inhibiting the growth of P. infestans, such as 2, 6-dichlorobenzonitrile, but some of these may also do harm to bacteria itself. Based on the above reasons, we finally choose lm-defensin, a defensing from maca that can effectively weaken and inhibit the growth of P. infestans

Circuit design

To purify defensin, we put a His tag on the N-terminal of the defensin. Also to produce lots of defensin, we choose T7 promoter and E.coli BL21 (DE3) which is used to express huge amount of proteins.

System

Our defense systen contains three different: prevention, detection, and cure. So how do we connect this three part together to form an impeccable defense system? First, we plant the genetically modified potato in the farm to prevent the disease from infecting the whole farm in a short period of time. Secondly, we implement the soil based microbial fuel cell that can detect and report whether the potato is infected or not immediately, in case that the potato can’t fend of the disease. When the SMFC detects the disease, it will send an electric signal that can trigger the spraying system. The spraying system will then spray the environmentally friendly defensin that we produced automatically. Man power is not required in the whole process except planting the potatoes.
Also, the defense system will also connect to a phone app that we created, so that we can know whether the potatoes are healthy or not. The defense system is localized, socialized, and mobilized.

FYVE inhibition

This model was designed to investigate the competitive binding between FYVE protein domain and PI3P. Before we construct the circuit with FYVE, we have to determine whether the affinity between FYVE and PI3P is strong enough to compete with Avr3 effector protein secreted by P. infestans. We characterize this agonist-antagonist competition model by Gaddum-Schild equation. This model consists of two parts: monomeric FYVE and dimeric FYVE. According to paper research, monomeric FYVE is not so stable but the affinity of both monomeric and dimeric FYVE is quite high. The model is used to determine how much FYVE protein domain we should link with a linker so that the affinity of FYVE will be high enough to compete with Avr3.

Result

Conclusion

As we can see in the models dimeric FYVE is not only more stable but also have higher affinity to compete with avr3, the effector protein secreted by p.infestans. To create FYVE protein domain with higher affinity, we can link as much FYVE domain artificially as we want. However, it takes a lot of time to add a linker protein between FYVE and connect two FYVE together. Moreover, the longer the FYVE is, the longer it takes for the palnt to degrade the protein we created, which might have a negative impact to the potato. By constructing this model, we can find out the simplest way to create a competitive inhibitor and save the time for trial and error.

Parameters

Reference:

1. Structural Basis for Endosomal Targeting by FYVE Domains
2. Phosphatidylinositol 3-Phosphate Induces the Membrane Penetration of the FYVE Domains of Vps27p and Hrs
3. External Lipid PI3P Mediates Entry of Eukaryotic Pathogen Effectors into Plant and Animal Host Cells

Oscillation

We construct a Lux/Aiia quorum-sensing oscillator so that S. oneidensis-MR1 will generate oscillating current which can help farmers tell if the potatoes are infected. We build two models to characterize the outcome of the genetic oscillator. The first model is used to predict whether the genetic oscillator will work. Since it is possible that some of the engineered S. oneidensis we inoculated on anode won’t survive, the second model is used to predict whether the gene expression will oscillate when there are only a few bacteria left on the anode.

Population simulation with delay differential equation

Since our circuit design is based on the circuit published by Danino et al.[1] our first model is a slight modification of the equation from the supplementary information. This model consists of four delay differential equation. We added one more delay differential equation to this model to simulate the expression of MtrB gene.

Parameters

Conclusion

As you can see in the graph, the production of Aiia and LuxI will oscillate as the time goes by. However, the production of MtrB will build up instead of oscillating. This is because MtrB is too stable and it is not totally degraded before the next period of oscillation. Thus, it is necessary to be targeted for proteolysis by adding a degradation tag in MtrB coding sequence.

Single cell simulation with ODE

If we inoculate the engineered S. oneidensis MR-1 into the anode of our SMFC, chances are that some of the bacteria will not survive in soil. Thus, it is necessary to simulate if the oscillator will work with only a single bacteria or a small population of bacteria.

With a small amount of bacteria, there will not have a significant time delay, and that’s why we chose a set of ODE to simulate the genetic oscillator.

Conclusion

As we can see in this model, even if there is only a small amount of bacteria left on the anode, the oscillator can still work. However, the current output will be much lesser than that of a large population of bacteria.

Parameters

Reference

1. Danino. T, Mondragon-Palomino, O, Tsimring, L. & Hasty, J. “A synchronized quorum of genetic clocks.” Nature 463, 326-330 (2010).
2. Petros Mina, Mario di Bernardo, Nigel J. Savery, Krasimira Tsaneva-Atanasova. “Modelling emergence of oscillations in communicating bacteria: a structured approach from one to many cells.” Published 7 November 2012
3. Danino. T, Mondragon-Palomino, O, Tsimring, L. & Hasty, J. “A synchronized quorum of genetic clocks.” Nature 463, 326-330 (2010).