The essence of our circuit lies in its simplicity. It is a “switch” that will enable the expression of a protein when a signal molecule is present in the environment. This switching mechanism is divided into two modules; A detection module and a nutrient uptake module. The first consists of the TetR-KRAB silencing complex, which is under the regulation of a synthetic riboswitch, an RNA element positioned in the 3’UTR region of the TetR-KRAB transcribed area, able to detect the presence of microcystin-LR (biomarker). The second module is a TetR-regulated PHT1 inorganic phosphate (Pi) transporter overexpression cassette. Once the proper environmental signal, microcystin-LR, is available to bind in the aptamer moiety of the riboswitch of the detection module, its conformation will change, resulting in blocking the transcription of the TetR-KRAB protein, thus allowing the overexpression of our protein of interest positioned in the nutrient uptake module.

Figure 1. Representation of our synthetic device in the absence (left) and presence (right) of microcystin-LR.

RIBOSWITCH

After outlining the general concept that we wanted to base our design on, we started looking into possible candidate molecules that could act as indicators for eutrophication. These should fulfill certain requirements:

1. be present only in eutrophic waters. We did not want our system to interfere with the natural balance of any ecosystem, except for the polluted ones.
2. be detectable by the plant that we want to install our system into (i.e. Phragmites australis).
3. not pose significant risk (toxicity) to the plant’s development.

Based on those criteria and available literature, we chose microcystin to be our biomarker, a cyanotoxin particularly dangerous to both nearby plants and animals (humans included)1. Named after the microcystin-producing Microcystis aeruginosa cyanobacterium, this 995,5 Da cyclic non-ribosomal heptapeptide acts as an inhibitor to PP1 (Protein Phosphatase 1) enzymes. We based our idea on the existence of a microcystin detection method that relied on aptamers2. Aptamers are relatively small nucleic acids, or peptides, that can bind with high affinity and specificity to certain molecules. It so happens that a nucleic acid aptamer sequence that binds to microcystins (microcystin-LR to be exact, hereafter MC-LR) exists and is widely used for this method.
Furthermore, the plant that we intended to use (P. australis) was found to be both a native resident of eutrophic freshwater habitats and an excellent accumulator of MC-LR, from which it suffers only minor toxicity effects 3.

Ancient relics in modern chassis

The next step was to find a way to incorporate this aptamer in our detection system. Aptamers, as far as we know, do not exist in organisms as free molecules. They are, instead, parts of riboswitches: one of the oldest mechanisms devised by primordial life forms to regulate metabolic pathways4. Riboswitches function by exerting a regulatory effect upon ligand-binding, and most of them have a repressive action. Although very common among bacteria, in eukaryotic organisms their presence appears to be limited to one family: the thiamine pyrophosphate (hereafter TPP) riboswitches5. TPP riboswitches are widely spread in almost all plant species, hence we chose a member of this family as the basic part of the detection system to engineer and make it responsive to MC-LR. They are located at the 3’UTR region of TPP-biosynthesis genes and act as splicing riboswitches, controlling 5’ and 3’ splicing site sequestration, thus linking intron retention to the presence of ligands. Whenever TPP (or MC-LR) is present, it binds to the aptamer region of the riboswitch and induces a structural change that enables splicing. The intron contains a poly-A signal. Its removal destabilizes the mRNA and leads to degradation, hence acting as a NOT gate6.

Our riboswitch engineering

A riboswitch is a delicate structure. Even a single point-mutation in the whole aptamer region can inactivate it. Thus, changes due to engineering must be as limited as possible, striking a balance between completely replacing the aptamer and keeping any innate structural properties in respect7.
Initially, we obtained the TPP riboswitch8 and MC-LR aptamer2 sequences. We, then, replaced the aptamer region with one targeting MC-LR and tried to notice any significant structural changes through the RNAfold program of the ViennaRNA package. We tried to keep the minimum free energy of the engineered riboswitch within the boundaries of the previous, natural one (-20 to -25 kcal/mol depending on plant species). To do that, several changes (additions and replacements) were made in the non-conserved part of the riboswitch. This way, a small library of new sequences was created, which was later screened in silico for possible candidates, based on various parameters.

TET-R DEVICE

In need of a signal inverter

Placing the detection within the parameters of our system showed that, in order to drive expression in the presence of Microcystins, we needed another interpreter, to invert the signal from a negative one to a positive one. To put it in technical terms, we needed to create an inverting Not Gate System, therefore, falling into the need of a repressor.

Tetracycline Repressors for synthetic devices

Repressors have been widely used in synthetic biology as they allow for a precise control of gene expression, producing a desired phenotype. Amid the field, are repressors that can function properly in plants and amongst the most prevalent ones are the Tet Repressors9, as was suggested to us by Prof. Kalliope Papadopoulou. Tet Repressor proteins (or TetRs) are proteins that play a vital role in antibiotic resistance against Tetracycline. By looking into the repressor’s role in Tetracycline-induced transcriptional control, it was understood that the suppression occurred only in the adjacent presence of specific DNA sequences called Tet Operators (or TetOs)10. And so, there were two problems that needed solving; which TetR to choose and how to place the TetO sequences. The registry of biological parts already provided standardized parts for the two of them (BBa_R0040, BBa_C0040), so what remained were a few tweaks in order to achieve proper expression in a plant system.

The hidden hurdle

One of the main problems when working with tetracycline-controlled gene expression in planta is the system’s leakiness. It turns out that ‘leaky’ gene expression is quite a common and rather persistent problem in synthetic biology and often contributes a lot to the poor performance of a genetic circuit11. In general, leakiness seems to be dependent on the host of the synthetic gene as well as the genetic parts of a construct12. In regards to our system, we found 2 different ways to influence the level of leakage, one revolving around the sequence of TetR and the produced transregulators13 - the so-called Tet toolbox - and the other around the number and placement of TetO repeats in close proximity with the promoter.

Our Repressor-Operator Design

Taking all that information in, we concluded our preferred design of the inhibitor - corresponding promoter complex. This would include the addition of a Nuclear Localization Signal (NLS), specifically Simian Virus 40 NLS14, to the TetR sequence, since, even though this is a small repressor (48kDa), which meant that diffusion into the cell nucleus is more than likely15, this would ensure its passage into the nucleus. Following that, the addition of the Krüppel Associated Box protein (KRAB)16 to create a TetR trans-silencer called TetR-KRAB, where, as the name suggests, TetR has been fused with the Krüppel Associated Box protein (KRAB), enhancing the inhibition abilities of the complex by adding another mechanism of silencing13. You can learn more about this part in the Improvement page of our wiki.

Additionally, regarding the TetO sequences we decided to use a combination of the pTight (also known as pTetO7), containing an altered version of the Tet Response Element (TRE) with seven direct repeats of TetO17 and pTriple Op, containing three remote TetO sequences in close proximity to the TATA box18. It should be noted that pTriple Op is a modified version of p35S CaMV, which is a promoter with great expression levels in plants. You can learn more about the sequence modifications and their purpose in the Engineering Success page of our wiki.

Figure 2. Representation of the Detection Module in the absence (left) and presence (right) of microcystin-LR.

As mentioned above, the purpose of our entire synthetic system is the increased absorption of phosphorus, which will only be induced by the presence of eutrophic indicators (e. g. microcystin LR) in the roots of the plant. We now come to the second part of our system, which is the microcystin-LR dependent overexpression of the PHT1 transporter. This module consists of a transcriptional unit that contains the pht1 gene with a constitutive promoter, whose structure is engineered to be inhibited in the presence of the TetR protein.

Phosphorus or Nitrogen?

But taking a step back, first we needed to determine a strategy to deal with this phenomenon. According to literature19, enhanced nutrient removal from the water can limit the eutrophication problem, hence we decided to incorporate it into our design.
Recent data show that the most effective method of dealing with eutrophication is a combination of removing the two main nutrients that cause the problem, namely phosphorus and nitrogen19. Unfortunately, due to financial and time limitations, we had to choose one of them to focus on. Phosphorus is known to be more associated with eutrophic phenomena in fresh water 20, with nitrogen contributing most to coastal and estuarine ecosystem deterioration21. Since the main focus of our project is the bioremediation of Lake Karla, we dealt only with phosphorus uptake from freshwater bodies.

Inorganic Phosphate (Pi) Transporters

Therefore, we studied a variety of different plant systems that could eliminate excess phosphorus: by diffusion, Pi root transporters and soil microbes colonizing the roots (mycorrhizal fungi)22. After careful consideration and consulting Dr. Kalliopi Papadopoulou, it was evident that Pi transporters fit best in our design for their high affinity to Pi and simplicity compared with other complex systems.
Moving on to known Pi transporters, we opted for the PHT1 family of phosphate transporters for the following reasons:

1. they have high affinity for orthophosphate (Pi)23 ,
2. they appear to localize primarily to the cell membrane of root cells24 ,
3. their main role appears to be Pi acquisition25 and
4. a lot of members of this family are well characterized in various plants.

AtPHT1;5 and OsPHT1;6 Transporters

After further screening of possible PHT1 transporters, we decided to use the genes of PHT1;5 from Arabidopsis thaliana 26 27 and PHT1;6 from Oryza sativa28 29 Pi transporters, based on the following criteria:

1. Good degree of characterization.
2. Εvolutionary relatedness of source organism to P. australis.
3. Increased Pi uptake in overexpressing plants, regardless of Pi concentration.
4. Little to no toxicity when the pht1 gene is overexpressed.

Design of the second module

The pht1 gene is expressed under the control of a modified version of the widely used Cauliflower Mosaic Virus 35S promoter, called pTriple Op (as described above - Our Repressor-Operator Design).
For the transcriptional termination of the pht1 gene, we tested the functionality of two different terminators, to make sure that this module would work properly. We used the commonly used NOpaline Synthase terminator (tNOS), alone, or coupled with the plant Heat Shock Protein terminator (tHSP) to ensure a strong poly(A) signal. In the case of the NOS terminator, we also added a “stuffer” sequence to distance the promoter controlling KRAB expression from the TetO region in higher level constructs.
It is important to mention that we ordered two variants of genes for the same transporter, one containing the original coding sequence and one containing the coding sequence, codon optimized for the model plant Nicotiana benthamiana. We also added Kozak frame’s elements in the sequences containing the start of the translation for better expression results. For more details, see Engineering Success. If the second module works as expected, then the pht1 gene will be overexpressed only in the absence of the TetR protein, i.e., the presence of the microcystin - LR.

Figure 3. Representation of the Nutrient Uptake Module in the absence (left) and presence (right) of TetR protein.

Overall our design consists of two modules: a detection module and a nutrient uptake module. The purpose of the detection module is to perceive eutrophic indicators in the plant's environment and in response to enable the operation of the second module. For this reason, the detection module consists of a riboswitch sequence, which changes its RNA structure in the presence of microcystin-LR, and thus does not allow the proper expression of the TetR protein, whose gene is located upstream of the riboswitch.
The second module, i.e. the module for the nutrient uptake, consists of the gene of a phosphorus transporter (pht1). This gene, hence the expression of the PHT1 transporter, is under the control of a constitutive promoter, which, however, has been modified so that the TetR protein suppresses its function. Consequently, in the presence of the TetR protein the PHT1 transporter cannot be expressed, while conversely in the absence of TetR the transporter is able to be expressed, leading to its overexpression.

Figure 4. Overall representation of our synthetic system in a plant cell.

In other words, when the plant senses eutrophication, i.e. microcystin-LR, in its cells, it will increase Pi uptake, while otherwise this property will not be induced.