Antiflorigen effectiveness
One of the main objectives of the project is to find an antiflorigen protein that is able to delay flowering. To create a proof of concept, antiflorigen effectiveness testing was divided into two stages.
Stage 1: As there is no existing literature or knowledge about the heterologous production of antiflorigens in a microbial host, the first step was to gather candidate genes, clone them and purify the resulting proteins from E. coli.
Stage 2: Testing purified proteins in planta to obtain data on flowering delay. Testing on fruit trees is of course not possible within the timeline of iGEM, so we went for the next best thing, which is the plant model species Arabidopsis thaliana.
Stage 1 - Design
Even though florigens and antiflorigens from the PEBP gene family are conserved among plant species, there are small differences between them which result in different traits when it comes to transportation and effectiveness [1,2]. As we were unable to find literature on antiflorigens production in a microbial host, we created a list of candidates to test their effectiveness (Table 1).
Table 1: Origins of each of the chosen PEBP gene family proteins.
| Species | Gene name | References |
|---|---|---|
| Arabidopsis thaliana (thale cress) | TFL1 | [1,3,4] |
| ATC/CEN | [1,5] | |
| Chrysanthemum seticuspe f. boreale (chrysanthemum) | CsAFT | [1,6,7] |
| CsTFL | [1,6,8] | |
| Beta vulgaris subsp. Vulgaris (sugar beet) | BvFT1 | [1,9,10] |
| Malus x domestica (apple) | MdTFL1-1 | [11,12,13] |
| Prunus serotina Ehrh. (black cherry) | PsTFL1 | [14] |
| Prunus avium (sweet cherry) | CEN-like 1 | N/A |
| CEN-like 2 | N/A |
This list includes two genes from Arabidopsis thaliana. Since A. thaliana is a model plant, it is one of the best studied species regarding flowering control pathways. A. thaliana is also already proven to show flowering delay once antiflorigens were overexpressed in it. Since we also use A. thaliana as our experimental organism, these two proteins can function as a positive control in the proof-of-concept in vivo plant experiments.
Furthermore, there are two genes from another well-documented plant species: Chrysanthemum seticuspe f. boreale. The antiflorigens from C. seticuspe f. boreale showed flowering delay once overexpressed. Which makes it interesting to compare it to the results of our in vivo experiments.
Following, we have a gene from sugar beets which is evolutionarily quite distant from A. thaliana and could offer insights into the conservedness and inter-species applicability of the PEBP protein family.
Next is a gene from an apple tree, which is among the few well-studied tree species. Interestingly, the gene has already been overexpressed in A. thaliana, proving the flowering delaying properties in another plant species.
Then we selected a gene from a cherry species of which the protein structure is reported to be very similar to other Prunus species like apricots and peaches, which makes it promising for the application that is central to our iGEM project.
Lastly, we selected two genes from a sweet cherry species. However, we found this gene by comparing sequences from different species as there is no literature that provides a complete antiflorigen gene from sweet cherries. As such, these genes are regarded as putative and will be considered a wild card in our experiments.
The genetic sequences of the antiflorigens were obtained and codon optimized for expression in E. coli. To make sure that a his-tag can be added at the N-terminus of the proteins, it is helpful to know its 3D structure. As can be seen in Figure 1, which depicts a TFL1 protein as a representative for antiflorigens, the terminal ends are accessible. This suggests that the his-tag will not interfere with folding and protein function and should be free to interact with a purification column.
Figure 1: 3D protein model of TFL1 [17], which is representative of how antiflorigens are folded.
The designed gene constructs contained a 5’ NdeI and a 3’ BamHI restriction site to ligate the genes into the pET-16b backbone using an old school restriction digestion approach. This backbone was chosen as it adds a his-tag sequence at the N-terminus of an inserted protein. The final design can be seen in Figure 2.
Figure 2: Final plasmid design for TFL1 in pET-16b, which is representative of how the other genes were designed into the pET-16b vector.
The designed and created plasmids, were transformed into Escherichia coli strain Dh5α to construct a plasmid library. After the library was made, the plasmids were transformed into another E. coli strain called BL21, which is specifically designed for protein purification and therefore better suited for the job than the Dh5α strain, which is mostly used for plasmid transformation purposes like constructing a plasmid library.
Protein production was induced by adding IPTG to a growing bacterial culture. After which the proteins were purified under native conditions using the attached his-tag to the protein and a his-tag resin to which it will bind.
Stage 1 - Results
Out of the nine chosen candidate genes, seven were correctly cloned in the plasmid backbone (Figure 3) and transformed into E. coli Dh5α.
Figure 3: Whole plasmid sequencing results of the successfully transformed plasmid constructs.
Unfortunately, CsTFL and MdTFL1-1 were never correctly cloned into the plasmid backbone. This might indicate that the resulting proteins are detrimental to the bacterial host, for example because of the formation of inclusion bodies. Even though the proteins should not be produced yet, the lac operon is well known to be leaky. So, even without the inducer IPTG, some amounts of the proteins will be produced. Even though no toxicity can be expected from the proteins, it still might be that the product of the designed gene construct limits the survivability of the bacterial host.
All seven plasmids were successfully transformed into the BL21 production strain and the proteins were purified. The antiflorigens are all roughly the same size and weight, around 180 amino acids and 20 kDa. So, when visualizing the purification fractions on the SDS-page gels, we were looking for a band around the 20 kDa size. The first purification trial was done under denaturing conditions to test if antiflorigens could be produced in a microbial host. The results are presented in Figure 4.
Figure 4: Protein purification results of the two A. thaliana proteins TFL1 and ATC/CEN under denaturing conditions. U: Protein production uninduced fraction, I: Protein production induced fraction, L: Lysate fraction after cell lysis, FT: Flow-through fraction of the purification process, W: Wash fraction of the purification process, E1: First elution fraction of the purification process, E2: Second elution fraction of the purification process using a different buffer, E4: Fourth and last collected elution fraction from the purification process.
There are clear bands visible around the 20 kDa mark, implying that we successfully for the first time created a microbial host that produces eukaryotic antiflorigens. Additionally, you can see a band around 20 kDa at the uninduced fractions, confirming the statement made earlier that the lac operon is being known to be leaky.
The next step was getting these proteins purified in their native form. We were able to purify two proteins a week, and so a choice was made to try to produce and purify six antiflorigens: TFL1, ATC/CEN, CsAFT, BvFT1, PsTFL1 and CEN-like 1. The results can be seen in Figure 5.
Figure 5: SDS-page results of the His-tag purification of six antiflorigens in native form. U: Protein production uninduced fraction, I: Protein production induced fraction, L: Lysate fraction after cell lysis, FT: Flow-through fraction of the purification process, W: Wash fraction of the purification process, E1: First elution fraction of the purification process, E2: Second elution fraction of the purification process using a different buffer, E4: Fourth and last collected elution fraction from the purification process.
There are real clear bands again around the 20 kDa range, so even in native form the proteins can be extracted from their microbial host. However, these proteins are now still in the elution buffer. For the in vivo plant experiments the proteins need to be in PBS. So, a buffer exchange was performed, to get them in the right buffer. Sadly, in this process the BvFT1 proteins were lost due to bad separation of the proteins from their imidazole containing elution buffer to the PBS buffer. In the end, we continued to the in vivo plant experiments with five remaining purified proteins in PBS solution.
However, these proteins are now still in the elution buffer. For the in vivo plant experiments the proteins need to be in PBS. So, a buffer exchange was performed, to get them in the right buffer. Sadly, in this process the BvFT1 proteins were lost due to bad separation of the proteins from their imidazole containing elution buffer to the PBS buffer.
In the end, we continued to the in vivo plant experiments with five remaining purified proteins in PBS solution.
Stage 2 - Design
When setting up the experiments, again we encountered no previous literature about in vivo plant experiments with heterologously produced antiflorigens. So, we had to look for literature about florigens instead and found one patent, on which we based our set-up as well. This patent used cell-penetrating peptides to get florigens inside the plant cells of A. thaliana [15]. As florigens are different than antiflorigens, we adapted the protocol to reflect these differences in our experimental set-up. So, in the end all five remaining antiflorigens are tested in three different concentrations (25, 50 and 100 μM) to get an indication on what dosage can be effective. 50 μM is the concentration already proven to be effective for florigens in the mentioned patent. The antiflorigens are combined with a cell penetrating peptide called R9 in a 1:10 ratio concentration (250, 500 and 1000 μM) as already mentioned to be most effective [15]. R9 is a peptide made up of nine consecutive Arginine amino acids and apparently aids proteins diffus e into plant cells without the need of an injection, as shown in Figure 6.
Figure 6: Impression of the macropinocytosis cell penetrating mechanism of the R9 peptide. This mechanism is reported to be the cause of diffusion induced by the R9 peptide [16].
This is incredibly helpful, since the A. thaliana plants are small at the moment of application, so injection is not really an option. Furthermore, the combinations of antiflorigen protein and R9 peptide resides in a PBS solution, to make it biologically applicable. This was then applied on the shoot apical meristem , where flowers develop.
Unfortunately, the high antiflorigen concentration of 100 μM couldn’t be reached for the proteins ATC/CEN and CEN-like 1. So, both are only tested at the remaining two concentrations.
At the moment of application, the A. thaliana plants are only a few days old and thus very small. Because after this age, the genetic and molecular switch of the flowering cascade is already activated. Eventually a flowering delay is expected on the moment it normally flowers, which is after 4 to 5 weeks. This set-up is depicted in Figure 7.
Figure 7: Simplistic depiction of the in vivo plant experimental set-up, which will test various concentrations of antiflorigens on each a set of 10 A. thaliana plants
Stage 2 - Results
Unfortunately, due to plants needing quite some time to grow, the in vivo plant experiments are still ongoing at the time of writing, as can be seen in Figure 8. Hopefully, in the near future, we will have the results we are aiming for.
Figure 8: A. thaliana Col-0 plants inoculated with antiflorigens growing in the climate room as we speak.
Conclusion At this moment it is fair to say that we can confirm the possibility of antiflorigens being produced in and purified in native form from a microbial host. Thereby, successfully finishing stage 1 of our proof of concept of the effectiveness of antiflorigens.
Stage 2 of our antiflorigen experiments has already been started, however the much-anticipated results of these tests remain to be exposed on a later time-point.
To be continued...
