Engineering work normally relies on selection for a desired genotype by antibiotic resistance. Agrobacterium has been found to show a quite high natural resistance to some antibiotics (Khafagi et al., 2012), however detailed information for specific strains are often unclear. Therefore we tested the Agrobacterium strains used in this project to their natural sensitivity to various antibiotics in the absence of antibiotic resistance genes.
The antibiotic sensitivity assay performed with A. rhizogenes ARqua1 and A. rhizogenes K599 showed that both strains show natural tolerance to some antibiotics.
Our results show sensitivity of both strains to ampicillin, chloramphenicol and tetracycline, with A. rhizogenes K599 being more sensitive to those antibiotics. A. rhizogenes ARqua1 has been reported as a streptomycin-resistant derivative of the R1000 strain (Thompson et al., 2020), which also fits our observations. Surprisingly, the K599 strain was not resistant to streptomycin as reported by Alzohairy et al., 2013, but tolerant to spectinomycin. Whereas ARqua1 strain showed to be resistant to high concentrations of streptomycin but not to spectinomycin. When culturing Agrobacterium in the lab, we therefore used streptomycin to select for the native ARqua1 strain and spectinomycin to select for the native K599 strain. During our project, we introduced the 35S:RUBY plasmid into both strains with the purpose to identify successfully transformed plants. This plasmid carries a bacterial antibiotic marker that allows selection for this plasmid using streptomycin or spectinomycin. In view of the natural sensitivity of the used A. rhizogenes strains, we used the respective other antibiotic. Further tested antibiotics not listed in figure 1 were found to be usable for selection of a desired plasmid carrying the respective antibiotic resistance gene. These antibiotics include gentamicin, hygromycin and kanamycin. Whenever we introduced a second plasmid into either Agrobacterium strain, we used plasmids with a pABCa or pSRK backbone, that hold a gentamicin resistance and therefore allowed us a reliable selection for the introduced plasmid.
One of the main challenges when creating SynBio tools outside the classic prokaryotic model organisms is the lack of thorough characterization of basic parts. This leads many researchers to rely on E. coli part characterization data, despite the fact that they do not always constitute the optimal tool for the task at hand. In order to avoid this problem, our team started from the beginning of our project to work on characterizing basic parts that could be used in our target organism Agrobacterium.
The characterization construct expressing the LuxABCDE reporter used for our experiments is harbored in a pABCa plasmid backbone. This vector was originally created for expression in Sinorhizobium meliloti, and incorporates a single copy repABC-type origin of replication (Döhlemann et al., 2017). The version used by our team was further adapted by one of our PI’s lab to be compatible with Golden Gate assembly (Meier et al., 2023). The assembled plasmids for the promoter characterization were used by one of our PI’s lab in different experiments and kindly provided to us (Seiler, 2023).
The Anderson promoter collection has been thoroughly characterized in the chassis most prevalent in iGEM, such as E. coli. For this reason, the collection is extensively used as a standard for constitutive expression in prokaryotes. However, little to no data is available when it comes to non-model prokaryotic organisms. Therefore, our team set out to characterizing their expression strength in Agrobacterium rhizogenes strain ARqua1, which was used for most of our plant transformations.
Based on the measurements and comparisons to the data originally produced in E. coli by iGEM Berkeley 2006, it was observed that the promoters display distinct expression levels in the two organisms. Most of the promoters tested induced only weak expression in Agrobacterium, including ones that promote quite high expression in E. coli, like J23104.
Surprisingly, the strongest promoter of the collection was not J23100, as it is the case for E. coli, but J23102, with more than twice the expression levels observed in J23100. These results serve to further drive the argument for the necessity of characterizing basic parts when establishing an organism as a synthetic biology platform.
In addition to the characterization of commonly used constitutive promoters as a proof-of-concept, our team also wanted to employ inducible systems in the project, which were characterized next. Inducible systems are essential for biotechnological applications and for studying gene regulation, since they allow the creation and control of more complex pathways as well as programming desired responses to environmental cues (d’Oelsnitz et al., 2022). Similarly to the constitutive promoters described above, inducible systems that work efficiently in E. coli are not necessarily suitable for use in alpha proteobacteria (Kretz et al., 2023). Therefore, we selected 11 inducible promoter systems for characterization, with the goal of identifying a suitable promoter for VirG overexpression (see below).
There are several factors that need to be taken into consideration when characterizing an inducible system:
Input/output - Each inducible system consists of a promoter that is repressed by a regulator which, in turn, can be relieved by binding to low molecular weight metabolites.
Leakiness - Unlike electrical systems, biological circuits do not work in a strict ON/OFF mode. Promoters, even when not activated by an input, still have a basal level of expression. Promoters with a high expression level in their OFF state are considered “leaky".
Dynamic range - The expression level of an inducible system can span orders of magnitude from their OFF state to the maximum output strength.
Cross-talk - While a single inducible system covers the needs of many projects, several inputs may be necessary to control more complex gene networks. Therefore, it is essential that the different small-molecule inducers are only active on their specific inducible system.
Cytotoxicity - In addition to their interference with other inducible systems, the molecules used as inducers may also interfere with essential cellular processes.
In our effort to establish an inducible system that takes into consideration all the above-mentioned factors, we selected 11 inducible systems that are widely used to drive gene expression and/or that had previously been shown to work well in closely related bacteria.
In this experiment, we aimed to screen as many inducible systems as possible, as we were aware that some may not work in Agrobacterium. We selected 9 promoters from the “Marionette Collection", which contains a number of inducible systems highly optimized for high dynamic range and low leakyness in E. coli (Meyer et al., 2019). Additionally, Ptac and Ptau were also included (Mostafavi et al., 2014; Stukenberg et al., 2021). Another consideration made when selecting the promoter systems to characterize was to include ones that use non-phenolic compounds as inducers (Ptau, IPTG, Pbetl, and Pbad), in the hope of minimizing cross talk with the native VirA/VirG two component system that responds to phenolic substances like acetosyingone, vanillin, or others (Bolton et al., 1986; Cha et al., 2011).
Based on this data, we decided to further characterize 4 promoters that seemed the most promising: Ptac, Pvan, Pnahr, and Ptau.
| Inducer | Solvent | Maximun induction |
|---|---|---|
| Naringenin | DMSO | 1 mM |
| Vanilin | EtOH | 100 µM |
| Cuminic acid | EtOH | 100 µM |
| IPTG | H2O | 1 mM |
| Sodium salicylate | H2O | 100 µM |
| Anhydrotetracycline | 1:1 H2O EtOH | 200 nM |
| DHBA | EtOH | 1 mM |
| Choline Chloride | H2O | 10 mM |
| Taurine | H2O | 10 mM |
| IPTG | H2O | 1 mM |
| Arabinose | H2O | 4 mM |
The results presented in figure 4 show the luminescence produced by induced vs not induced cultures in the mid-log phase. Out of the eleven promoters tested, only 4 displayed a significant difference in expression level between induced and not induced cultures: Ptac, Pvan, Pnahr and Ptau. Pbad and Pcym also had some response to the induction, albeit extremely faint. From those four promoters, we observe that Ptac and Pvan had the highest overall basal expression, and Ptau the lowest. In addition, Ptau did display the lowest leakiness of all systems tested. Overall, PnahR appeared to have a good middle ground between expression strength and expression tightness. However, upon further investigation, it was found that growth of A. rhizogenes ARqua1 cultures in liquid medium was inhibited by sodium salicylate (Figure 4). In Agrobacterium tumefacies C58, PttgR, Ptet and Pcym have been shown to have a stronger response than the one observed by our experiment, additionally the toxicity of sodium salicylate was not observed (Qian et al., 2021; Schuster & Reisch, 2021).
The next step was to evaluate the specificity of the selected promoters and their inducers in an independent assay. Overall, we observed a good degree of orthogonality in the promoters tested. The PnahR promoter showed the strongest activation by other inducers, mainly by vanillin and taurine, while Ptau showed some activation by salicylic acid and vanillin (figure 5). While the molecular similarity between sodium salicylate and vanillin could be the cause of the non-specific activation of PnahR by vanillin, the same is not the case for taurine. In E. coli, PnahR remained unresponsive to vanillin (Meyer et al., 2019).
In addition to the characterization of the commonly used constitutive Anderson promoters and inducible promoters, we wanted to characterize different reporters, codon optimized for Agrobacterium, to facilitate new and characterized parts to the iGEM community. Reporters are essential in molecular biological applications, because they allow to measure promoter activity. Once expressed in cells, the reporter is assayed by either directly measuring the reporter protein itself or assessing its enzymatic activity, correlating the strength of the upstream promoter element with the amount of reporter produced. Reporter genes are used as indicators of transcriptional activity in cells.
Our test constructs for the reporter characterization used the pSRK backbone, the 0-09 RBS3 DM, the 0-11 rrnB, TB0015 terminator, the PMPJ23102 promoter and the respective reporter to characterize. We tested three fluorescent reporters, mNeonGreen, Scarlet and Staygold as well as Nanoluc a luminescent reporter. The advantages of luminescence reporters are higher sensitivity and lower background noise. All reporters were synthesized and codon optimized for A. rhizogenes. Designing our reporter test cassette we chose to use the strongest of the previously characterized anderson promoters, PJ23102. Lacking knowledge about the behavior of our reporters in Agrobacterium, a strong promoter was the most promising option, in order to get qualitative proof of their performance, even for weak reporters.
For the characterization we transformed our constructs in both of our Agrobacterium strains, ARqua1 and K599.
Figure 6 displays the results generated for the fluorescent reporters. The fluorescence intensity was measured using an excitation wavelength of 552 nm for red fluorescent proteins and 488 nm for green fluorescent proteins. The figure shows the expression strengths measured in Arqua1 and K599.
Among the three reporters, mNeongreen demonstrated the highest fluorescence intensity, indicating a robust expression under the tested conditions. It surpassed both mScarlet and Staygold in terms of reporter strength. The high expression in strain ARqua1 suggests that mNeongreen could be a preferable choice especially when characterizing weak promoters or RBS. However, mScarlet represents a viable alternative with moderate expression levels. Staygold, while still functional, may be less desirable when high fluorescence intensity is required, demonstrating only a slightly higher expression compared to the Dummy negative control. In K599, mNeongreen is also the strongest Reporter. However, Staygold has a higher fluorescence intensity in K599 than mScarlet. Recognizing a missing negative control for the K599 strain, those findings should be considered with reservations.
ARqua1 is the strain with the highest fluorescence intensity in all measurements. The overall performance was notably superior compared to K599.
The overall performance was notably superior in ARqua1 compared to K599.
In our RhizoGene project, we aim to increase the efficiency of Agrobacterium rhizogenes-mediated transformation of plants, simplify transformation protocols, and expand the range of usable plant chassis. In order to determine the transformation efficiency of individual strains within the shortest possible time, we adapted a transformation protocol published in 2016 using the model organism Arabidopsis thaliana (Mai et al., 2016).
Arabidopsis, a well-established plant model organism, possesses several attributes that made it an ideal candidate for our research. Its compact size allowed us to efficiently study numerous plants within limited growth chamber space, while its short cultivation time facilitated rapid and precise prototyping of transformation protocols.
To assess transformation efficiency, we initially used Agrobacterium rhizogenes ARqua1 with the fluorescent reporter gene dsRed. However, scoring the fluorescent reporter in the transgenic plants was very time consuming. Also, the actual transgenic area could only be delimited imprecisely. In addition, a fluorescence microscope was necessary for the evaluation. Our goal was to establish a particularly low-threshold protocol for upcoming iGEM teams. To address this accessibility concern, we transformed the 35S:RUBY plasmid into Agrobacterium rhizogenes, offering the advantage of scoring transformation events by naked eye via red coloured betalain, simplifying and expediting the assessment process (He et al., 2020).
To classify the upcoming experiments, we established a baseline using our Arabidopsis thaliana protocol by replacing the bacterial strain with Agrobacterium rhizogenes ARqua1 containing the 35S:RUBY plasmid with the RUBY reporter gene and a hygromycin resistance on the T-DNA.
We transformed a total of 179 baseline plants in multiple rounds of experiments, plating about 50 sterilized and stratified seeds for each round. Five days after germination in the dark, the shoots of the seedlings were long enough to be separated from the roots. The wounded shoots were incubated in Agrobacterium suspension for 7 minutes. After 3 days, the Arabidopsis seedlings can be examined for the first time under the microscope for red tissue. New hairy root cells are not formed at this time point. Subsequently, the plants are plated out on petri dishes containing cefotaxime 300 mg/L to kill the Agrobacterium. Seven days later, i.e. 10 days after transformation, red hairy root tissue had formed if transformation was successful. Successful transformation of one rhizodermis cell is required for hairy root cell development. However, the formation of betalains does not occur with every formatting root. This can be explained by different combinations of native T-DNAs and RUBY T-DNA integration into the plant cells.
Three days after transformation, 46% of the 179 seedlings developed the red pigment betalain. This shows the expression of the RUBY reporter, corresponding to a transient or stable transformation of the T-DNA. More than 20% of the seedlings died between the first and second evaluation. One explanation for the high lethality could be the heightened plant defense due to the challenge by the bacteria, another explanation could be the particular stress of the plants due to a rough handling by the experimenters. Especially by incautious handling, microscopy under a coverslip can lead to high stress for the plants, from which some do not recover.
While working on our baseline, we came to consider whether it might be interesting at a later stage to select the transgenic tissue by the inserted hygromycin resistance. Since hygromycin is expensive, we thought of cloning another antibiotic resistance as a selection marker into the 35S:RUBY plasmid that would be more widely applicable by other labs and iGEM teams. After some literature survey we decided to use kanamycin resistance. Even though we did not use the kanamycin resistance in 35S:RUBY:KanR afterwards, we initially ran a second baseline to check if ARqua1 with the modified plasmid gave the same results as the first baseline.
As you can see in Figure 9, 3 days after transformation, we could observe a transformation event in 47% of the 284 Arabidopsis thaliana used. Remarkably, after 10 days, 35% of the plants counted on day 3 died. In fact, since the Hygromycin baseline (Figure 9), several new team members were incorporated in the plant laboratory who were still untrained in appropriate lenient handling. This confirms the hypothesis that a significant proportion of the deceased plants were caused by improper handling and cannot be explained by the virulence of the bacterium alone.
While part of our RhizoGene team focused on targeting Agrobacterium virulence genes using inducible promoters, we looked in parallel for other ways to increase the transformation efficiency in the newly developed protocols for different plant species. During our research, we came across a publication in which different molecules were used to increase transformation efficiency (Cha et al., 2011). For Agrobacterium-mediated gene transfer into plant hosts, the substance acetosyringone is often favored, which is naturally secreted by wounded dicotyledonous plants and induces Agrobacterium virulence genes. However, according to this paper, as an alternative to acetosyringone, other structurally similar phenolic compounds can also induce the virulence genes. For this reason, we decided to test vanillin in Agrobacterium culture to investigate an increase in transformation efficiency. This also offers the advantage that vanillin is significantly cheaper than acetosyringone. To achieve a higher probability of transformation events, we also centrifuged half of the bacterial culture with an OD(600) of approximately 1 and resuspended the pellet in the remaining culture. We added 500 µM vanillin to the bacterial culture with increased titer one hour before transformation.
The results 3 days after transformation were outstandingly good: 87% of the plants showed RUBY expression whereas without vanillin only 47% of the plants were RUBY positive. In this transformation, there were no dead plants after transformation. This proves that higher transformation efficiency with careful treatment does not lead to increased plant lethality at all. These overwhelmingly good results gave hope that more plant species could be successfully transformed by this adaptation of the protocol. The amazing results with Bambara groundnut confirmed this assumption (see results below).
In addition to increasing the virulence of Agrobacterium strain ARqua1, we also searched for bacterial strains that might perform better in the plant species we transformed. In a comparison paper between Agrobacterium strains ARqua1, K599 and R1000, K599 showed superior results in transformation experiments compared to ARqua1 with a higher virulence (Foti & Pavli, 2020). Therefore, we tested Agrobacterium rhizogenes K599 in Arabidopsis thaliana to observe the transformation efficiency with this strain. In K599, we transformed our 35S:RUBY:KanR plasmid.
Similar to the increase in virulence by vanillin, 3 days after transformation by K599, there is a increase in RUBY positive plants with 72% compared to baseline with ARqua1 with 47% (Figure 9). A low death rate of treated plants also suggests that the higher virulence of K599 did not have a negative impact on the plants.
Our team also focussed on the construction of composite parts which should enhance Agrobacterium rhizogenes-mediated plant transformation. Each version of our composite part consists of an inducible promoter, a version of VirG and one of 2 different backbones. The general idea was to fine-tune the expression of the transcriptional activator VirG as the master switch of all other virulence genes. Therefore, we were looking for an inducible promoter with a high dynamic range, low leakiness and no cross reactivity. The promoters of our choice were Ptac and Ptau which can be induced by IPTG and taurine (see above). The two versions TiBo542 and super80 of VirG differ in the independence of the sensory kinase VirA of the latter. Whereas the pABCa has a single copy for Agrobacterium. The pSRK entry vector (Khan et al 2008) carries the pBBR1 (broad host range) ori and was initially selected for our VirG overexpression constructs, due to its medium copy number in Alphaproteobacteria and compatibility with E. coli (Blázquez et al., 2023;Antoine & Locht, 1992). Additionally, we wanted to use the pABCa backbone with its single copy ori. We tested different constructs containing various combinations of the previously mentioned parts in plants.
A. thaliana transformation via ARqua1 containing Ptac_TiBo542[VirG]_pSRK was the first construct tested (Figure9). After 3 days we saw a total amount of 32% of plants which expressed RUBY. Seven days later, the number of RUBY-positive plants had increased to 57%. Thus, transformation efficiency after 3 days is 15% lower compared to the baseline with ARqua1 35S:RUBY:KanR. After10 days the efficiency increased 25%, making the transformation 11% more efficient compared to the baseline.
The next construct we tested was the improved version Ptac_super80[VirG]_pSRK. When analyzing the transformation efficiency after 3 days we saw a drop of RUBY positive plants to 29% compared to the baseline (Figure 9). Both results were against our previous expectations, however after several rounds of trouble shooting we came up with the idea that Ptac still shows relatively high leaky expression despite having a big difference in expression levels of induced to uninduced. To evade leakiness we chose the lower expressed Ptau.
When testing Ptau_super80[VirG]_pSRK we observed a slight improvement to 37% RUBY positive plants. Nonetheless, it is important to note that the total percentage was still lower than during the RUBY baseline experiments without ARqua1 containing any additional constructs. Upon evaluating our most recent results 10 days post-transformation with constructs containing the pSRK backbone, we were surprised to witness an unexpected increase in the number of RUBY-positive plants compared to the 3-day post-transformation results. We were especially surprised by the Ptac_TiBo542_pSRK transformation results which showed a transformation efficiency gain of 11% in comparison to the 35S:RUBY:KanR baseline. The current efficiency levels now seem to be on par with the outcomes from the baseline experiments. With the assumption that pABCa behaves similarly, we anticipate obtaining comparable results ten days after transformation, and we anticipate these results to be available within the next week.
Plasmid incompatibility occurs when two plasmids cannot stably coexist in the same bacterial cell line over multiple generations. Typically, closely related plasmids are incompatible, while distantly related ones are compatible. Incompatibility often arises from shared replicons with identical Rep protein specificity or controlling elements. Competition can also stem from similar partitioning systems. This incompatibility may be reciprocal, leading to equal chances of loss, or unidirectional, favoring the plasmid with additional advantages, such as a second replicon. Although classification systems exist for certain hosts, many plasmids defy categorization due to the absence of systematic incompatibility tests on all known plasmids.
Many of the most widely used origins of replication from E. coli, such as ColE1 do not work in Alphaproteobacteria. Hence, some of the plasmid backbones used by our team contained 2 origins of replication: one for viability in E. coli during cloning steps and one for replication in Agrobacterium, or alternatively, a broad host range ori. Additionally, our iterations of the best composite part were designed to co-exist in Agrobacterium with the 35S:RUBY:KanR plasmid carrying the plant selection marker and the pVS1 ori. Therefore, the stability of plasmids carrying different origins of replication was tested in strains also carrying the 35S:RUBY:KanR plasmid.
The pSRK entry vector carries the pBBR1 (broad host range) ori and was initially selected for our VirG overexpression constructs, due to its medium copy number in Alphaproteobacteria and compatibility with E. coli (Antoine & Locht, 1992; Blázquez et al., 2023). However, after the observations from plant transformations, we noticed that strains carrying both plasmids displayed lower transformation efficiency when compared to strains solely carrying the 35S:RUBY plasmid. This led to the suspicion that the two plasmids might be unstable when co-existing in Agrobacterium, negatively affecting cell health and thus decreasing overall transformation efficiency.
In order to verify this hypothesis, we conducted a stability assay in A. rhizogenes ARqua1. A. rhizogenes ARqua1 competent cells carrying the 35S:RUBY:KanR plasmid were transformed with the pM4_1_002 VirG overexpression plasmid and plated on selection media (Gentamicin + Spectinomycin). Colonies were picked and screened via colony-PCR to verify the presence of both plasmids. A cryo stock from one colony was used to inoculate liquid LB medium without antibiotics, gentamicin + streptomycin, and gentamicin + spectinomycin, which were then incubated overnight. On the following day, a 20 µL sample was taken from the overnight culture and frozen away, 1 µL of the overnight culture was used to inoculate fresh medium for a new overnight culture. This process was repeated until samples from 5 days were collected.
The gel images in figure 10 show that the combination of pSRK (pBBR1 ori) and 35S:RUBY (pVS1 ori) is not stable in A. rhizogenes ARqua1. After 4 days, both plasmids were lost in plain LB and LB(gen+strep) cultures. In LB (gen+spec), 35S:RUBY:KanR was still detected on the 4th day and then lost on the 5th. All cultures were visibly thin on day 4 and 5.
Based on these results, we opted to use the pABCa backbone for our VirG expression constructs.
As a final test we therefore decided to use the single copy ori in the pABCa backbone to express our VirG construct. And evaluate the transformation efficiency in Arabidopsis.
When testing Ptac_TiBo542[VirG]_pABCa construct we observed a transformation rate of around 36% contradicting our prior belief of pABCa being compatible with the 35S:RUBY:KanR ori. We experienced similar results when testing Ptac_super80[VirG]_pABCa only seeing around 48% RUBY-positive plants.
In our ambitious desire to improve transformation methods for non-model organisms, we decided to test the recently published cut-dip-budding protocol (Cao et al., 2023). The plants we selected for our work were Bambara groundnut Vigna subterrenea, the garden strawberry Fragaria x. ananassa and the regional dandelion Taraxacum officinale.
In the first experimental run, we cut the roots of the plants grown in a non-sterile environment, incubated the shoots of the plants in Agrobacterium rhizogenes ARqua1 for 15 minutes, and then placed the experimental plants in vermiculite. We evaluated the results after 2 and 4 weeks by removing the plants from the vermiculite and examining the cutting site with the naked eye. Of the 14 Bambara groundnuts used, no plant showed root growth after 2 or 4 weeks. 19 of the 31 dandelion plants showed root growth after 2 weeks, as did 7 of the 8 strawberry plants. Unfortunately, none of the plants developed hairy roots or RUBY expression. Accordingly, a transformation event probably did not occur in the first experiment.
In the second experimental run, we increased the incubation time in Agrobacterium rhizogenes ARqua1 culture. We used 9 strawberry plants, 18 dandelions and 12 Bambara groundnuts. In this experiment, spontaneous root formation also occurred in strawberries and dandelions. None of the plants showed signs of positive transformation events or transgenic tissue in the phenotype.
In accordance with the promising results in the study with our workhorse A. thaliana, we increased the titer of actively dividing Agrobacterium rhizogenes ARqua1 in the medium in experiment number 3. We also added 500 µM vanillin to the culture one hour before incubation of the plants to induce virulence of the bacteria. Fortunately, between experiments 2 and 3, we were able to begin work in another greenhouse that had stable temperature and humidity. Since we were also able to transform in this greenhouse, we were able to minimize environmental stress to the plants.
Experiment number 3 was conducted with 9 strawberry plants, 19 Bambara groundnuts and 6 dandelion plants. Surprisingly, after two weeks, 3 Bambara groundnut plants showed root growth for the first time, with the roots also showing hairy patches. This result was highly exciting, as this is phenotypic evidence of transformation events. Four weeks after transformation, as many as 8 Bambara groundnuts had developed roots. However, no RUBY marker was visible. After moving to the new greenhouse, we observed a particularly strong drought stress of the plants after transformation, as the solar radiation was more intense during summertime. Most of the strawberry and dandelion plants died within 2 weeks, so no RUBY expression or hairy roots developed.
After 4 weeks, we reexamined the experimental plants trial no. 3, which by this time had developed a mature root system. To our amazement and excitement, two of the Bambara groundnut plants developed ruby red root sections. This means that the T-DNA with the reporter gene of 35S:RUBY:KanR was successfully transformed into Bambara groundnut which means that our adapted protocol suits to transform this poorly investigated tropical crop plant plant. This is a real breakthrough that makes us proud, especially since Bambara groundnut has not been successfully transformed according to our literature survey at that time.
During the implementation of the cut dip budding protocol, we came across a method by Sebastian Cocioba via social media by chance. In fact, while working with the CDB protocol, we had thought about a way to further reduce the stress of the plants, as the roots suffered even when being carefully removed from the vermiculite. In the new method, rockwool is soaked with Agrobacterium culture and placed in a centrifuge tube filled with tap water. The plant, stripped of its roots, is placed in this assembly with the cutting side down and co- cultured with the bacteria. If root growth occurs, the roots can be assessed through the transparent tube wall. Expecting to follow up on the positive results with Bambara groundnut, we transformed dandelion, strawberry, and Bambara groundnut with Agrobacterium rhizogenes ARqua1 using the rockwool method. This involved adding 500 µM vanillin to the culture as in our previous, very successful CDB run and centrifuging the culture to double titer.
After the impressive transformation efficiency of Agrobacterium strain K599 in A. thaliana, we decided to transform our crop plants with this strain as well. We proceeded as described above. For now we are still waiting for upcoming results.
In a great collaboration with Lars Opgenoorth from the PhytOakmeter project at the University of Marburg, we got the opportunity to work with the oak clone DF159. Since the clone is axenically propagated, it was necessary to develop a sterile transformation protocol specifically adapted to this propagation procedurel. If you want to read more about our self-developed protocol, you can click here for our "Oak diary".
In the first round of experiments, we used the Agrobacterium strain ARqua1 to transform 53 oak plantlets. After the transformation day, we transferred the oaks to a medium containing 500 mg/L of cefotaxime and activated charcoal two days later. We then conducted an evaluation once or twice a week for three months, documenting health status and new, potentially transformed tissue.
In the first week after transformation, almost all oaks lost their leaves due to the enormous transpiration stress the plants had suffered under the sterile bench. Only after about 3 weeks, lateral buds sprouted and new leaves formed. At 4 weeks after transformation, we were pleased to observe root formation in two of the oaks, but no hairy roots. In fact, the culture in activated charcoal medium for evaluation turned out to be very impractical, as the formation of new tissue could only be observed inaccurately. At 12 weeks after transformation, we removed the oaks from the medium, evaluated them by visual inspection, and microscopically examined cross-sections of the roots and shoots in hopes of detecting ruby red tissue. At that time 27 of the 53 plantlets had been lost due to contaminations. Of the 26 plants that were not contaminated, 4 oaks had formed new roots. Two of these oaks had also developed hairy roots. Unfortunately, we were unable to detect RUBY in the oaks, indicating the transfer of one of the endogenous T-DNAs of the ARqua1 strain. Nonetheless, using a completely self-created transformation protocol, we were able to detect transformation events in two Quercus robur, which is tremendous success, since we were not aware of any report on an existing oak transformation protocol. This gave us hope that further improvement of the protocol might provide even better results.
After the promising results of the first experiment, we repeated the transformation again with 51 oak plantlets. This time we made some adaptations to reduce the drought stress during handling under the laminar flow hood. The detailed description of the performed method can be found on the experiment page. As a bacterial strain we used Agrobacterium K599 andinduced the virulence according to the promising results in Bambara groundnut with 500 µM vanillin. In addition, the bacterial culture was spun down and subsequently resuspended to increase the bacterial titer of the suspension twofold. At the time of wiki publication, the transformation was 19 days ago, therefore no statement about the transformation events can be made yet. Nevertheless, it can already be seen that the further developed method has had a positive influence on the oaks. On the one hand, only one oak shed its leaves after the transformation, and new buds also formed within the first week. After 2 weeks 21 oaks have already built calluses. As an innovation, we have the oak's cutting edges on the outer wall of the new test tube.
Agriculture relies mainly on a handful of monocotyl crops such as maize, wheat and rice. Despite their immense agricultural and nutritional significance, monocots, which represent the most crucial food crops in terms of both quantity and calories, have been explored by fewer than five iGEM teams throughout the history of the competition. When aiming to facilitate Agrobacterium-mediated plant transformation and broaden the spectrum of genetically accessible plants it is also important to think about testing in monocots. Since monocots don't release phenols like acetosyringone after wounding the virulence genes of Agrobacterium do not get activated and therefore the monocot does not get transformed (Cao et al., 2023). To circumvent this, phenolic additives like vanillin and acetosyringone are added to the bacterial culture and also the growing medium of the infected plants. In this context decoupling of VirG from VirA might result in an induction of the general virulence without the need of additives. When trying to work with the emerging model-organism for monocots Setaria viridis, commonly known as green millet, we encountered several challenges. At first we had to overcome the hurdle of seed sterilization. For this we had to come up with a way of completely removing the husks around the seeds so that our sterilization solution can reach every part of the seed. After successfully plating out several rounds of seeds which did not contaminate had virtually no seeds germinating so we looked for ways to break seed dormancy. After troubleshooting and literature research we found a way to break seed dormancy by adding a gibberellic acid (Sebastian et al., 2014). After our first trial of transforming the seedlings we surprisingly got 40% RUBY-positive plants after 3 days and even 53% after 10 days, slightly exceeding the baseline results for Arabidopsis. All following germinating trials were unsuccessful which forced us to ultimately change to a close relative of our chosen plant: Setaria italica, commonly known as foxtail millet. This plant is cheap, readily available and germinates even without the need of added gibberellic acid. However, after several trials with multiple constructs and conditions we couldn't detect the RUBY reporter. This could be due to the fact that the 35S:RUBY:KanR promoter is relatively weak in monocots (Jang et al., 2002). Sadly we couldn't explore the possibility to exchange the promoter and optimize RUBY expression in Setaria during the time of our iGEM project.
