Team:Marburg/Results

Results

Proof of Concept Different Species

In our project - OpenPlast - we aimed to develop plant chloroplast cell-free systems (CFS) from different species as prototyping platforms for part characterization. We proudly present the successful development of highly efficient cell-free extracts for the chloroplast of the model plant tobacco and spinach. Moreover, we are even more proud to announce that we were able to establish the first working chloroplast CFS of wheat as an important crop species, as well as of oak, allowing us to characterize the first biobrick of a tree species.

Figure 1:Cell-free expression of the Nanoluc luciferase by three different plant species
Luminescence values are given as arbitrary units and the data is presented on a logarithmic scale. The reaction was set up in a total volume of 10µl. Negative controls using only the plasmid DNA or the crude chloroplast extracts have been included respectively in order to verify the expression

Expression of the NanoLuc luciferase can be observed in extracts of tobacco, spinach and wheat chloroplasts in Figure 1. The tested construct contains the following basic parts: the T7 promoter, the gene10 5’UTR, the Nanoluc CDS and the rrnB 3’UTR from E. coli. The T7 promoter has been utilized in our project as a tool to prototype our extracts, since its exogenous transcription allows for high gene expression across distinct species.

Figure 2:SBOL scheme of the used T7 Universal Test Construct 5.0

The species we used in our project as our model organisms have all been chosen for specific reasons

Tobacco is the model chassis for chloroplast engineering as the transformation protocol is the most sophisticated and efficient known today. For this reason, we decided to include tobacco as a model for our cell-free systems to characterize our novel parts and compare our in vitro data to actual in vivo experiments.

To make our workflow of using cell-free chloroplast systems as accessible as possible for future iGEM teams and researchers around the globe, we explored a low resource approach, which would not require any greenhouses or other plant growing capabilities. To do so, we developed cell-free extracts of spinach chloroplasts with plant material bought freshly from the market. In our experience it is furthermore possible to use spinach bought in supermarkets, although fresh plants are to be preferred.

The common wheat is the leading source of vegetable protein and the most planted food crop in the world with a production of over 765 million metric tons [1]. However, engineering wheat is not just quite difficult and slow, but chloroplast engineering in this important crop is currently close to impossible. Therefore, we are delighted to have successfully developed a cell-free system of wheat chloroplasts, which now allows high-throughput testing of novel wheat parts. This system could also be applied to screen for various reagents to be used in developing efficient regeneration protocols to finally make wheat chloroplast transformation a reality!

Obtaining these highly efficient chloroplast extracts we now present was not a simple task. Several rounds of the DBTL-Cycle were necessary to accomplish our goal.

Proof of Concept First Cell-Free Extracts

The first extracts we created did not show high expression, so we went back to the drawing board and treated both the extract preparation and the construct optimization like an engineering problem, testing different variables separately. After our first proof of concept for working cell-free extracts of tobacco and spinach chloroplasts (Figure 2), we were able to use these extracts for our subsequent optimization efforts.

Figure 3: Proof of concept for chloroplast cell-free extracts
Luminescence values are given as arbitrary units and the data is presented on a logarithmic scale. The reaction was set up in a total volume of 10µl. Negative controls using only the plasmid DNA or the crude chloroplast extracts have been included respectively in order to verify the expression

Magnesium and Potassium Optimization

We began our optimization endeavors by varying the concentration of magnesium and potassium in our extracts, as it has been reported that these buffer components have a deciding impact on expression in cell-free systems [2].
For this measurement, we varied the concentrations of the two components systematically, keeping everything else the same. We were able to observe clear differences in expression levels between the concentrations used, showing curves with different optima for both salts. We want to highlight that such an optimization is a particularly important step in the creation of any new cell-free system, as required concentrations of these two salts can be vastly different across species.

Figure 4: Optimization of the different buffer conditions of the chloroplast cell-free reaction
Luminescence values are given as arbitrary units and the data is presented on a logarithmic scale. The reaction was set up in a total volume of 10µl. Negative controls using only the plasmid DNA or the crude chloroplast extracts have been included respectively in order to verify the expression.

DNA Titration

As our next step, we set out to optimize the DNA concentration we use for our measurements and found that expression plateaus quite fast, which can be advantageous for further experiments, as it allows us to save plasmid DNA stocks so that we are able to set up more reactions. We used our Universal test construct 5.0 for this measurement, which uses NanoLuc expression as an output (Figure 5).
Comparing expression levels of various constructs could lead to faulty conclusions if working with non-optimal concentrations, as slight changes in concentration could make the difference in expression. Additionally, variation of technical replicates can be reduced for the same reason.

Figure 5: DNA Titration in chloroplast cell-free system of Spinach and Tobacco
Luminescence values are given as arbitrary units and the data is presented on a logarithmic scale. The reaction was set up in a total volume of 10µl. Gene expression depends on the amount of DNA template added. Graphs are shown for N. tabacum and S. oleracea separately. Negative controls using only the plasmid DNA or the crude chloroplast extracts have been included respectively in order to verify the expression

Downscaling of Reaction Volume

The first optimization experiments resulted in major improvements of expression levels compared to our first extracts. So far, we had been working with 10 µL reaction volumes to reduce pipetting errors and to ensure enough supply of all reagents to our system. Nonetheless, we were now confident that we could downscale our reaction volume to 4 µl total volume. This would not only reduce the amount of reagents needed, but more importantly save up more of our precious chloroplast cell-free extracts.
We were successful in downscaling the reaction volume, as can be seen in Figure 6. Using smaller volumes did result in a reduction of the overall signal, although we hypothesize that this is largely due to the smaller total amount of protein produced in this smaller volume. The expression is still sufficient to be used for part characterization. Here, we would recommend using a ratiometric approach to normalize the luminescence output to a second reporter. This should allow for proper part characterization even in the 4 µL reaction. However, we do not advise to go lower than 4 µL, as this will not be accurately possible to pipet by hand.
Consequently, we thought about using the Echo liquid handler to downscale the reaction even further, as this would allow for a precise transfer of the cell-free reaction components. We ultimately did not follow up with this idea due to time constraints but would be glad to hear from anyone who wants to give it a shot (pun intended)!

Figure 6: Downscaling of the cell-free reaction volume
Luminescence values are given as arbitrary units and the data is presented on a logarithmic scale. The reaction was set up in a total volume of 10µl. Gene expression depends on the amount of DNA template added. Graphs are shown for N. tabacum and S. oleracea separately. Negative controls using only the plasmid DNA or the crude chloroplast extracts have been included respectively in order to verify the expression.

Codon Usage

There are several factors that affect translation efficiency in the chloroplast, as we have discussed in our design section. This includes codon usage, mRNA structure at the end and the beginning of the CDS, as well as the amino acid identity at the positions 2-5. As codon optimization is nowadays widespread practice to adapt certain genetic parts to a desired system, our first characterization goal was to test if different codon optimizations would influence the expression within our cell-free system.

Figure 7: Optimization of the codon usage of the Nanoluc reporter gene
Luminescence values are given as arbitrary units and the data is presented on a logarithmic scale. The reaction was set up in a total volume of 10µl. Negative controls using only the plasmid DNA or the crude chloroplast extracts have been included respectively in order to verify the expression.

Indeed, we could show that the same reporter, optimized for different chassis, shows significantly different expression levels (Figure 7). To our surprise, we found that our NanoLuc part, which was optimized for the chloroplast of the green algae Chlamydomonas reinhardtii, shows higher expression in a tobacco cell-free system, than the same part specifically optimized for the chloroplast of Nicotiana tabacum. There are many possible explanations for this observation; one option is the mRNA at the beginning of the coding region forming different secondary structures depending on which codon optimization was chosen. This is a first example of the practical use of our system, as in a living system a 3-fold difference in expression can make a difference between a gene product being detectable or not - especially when working with species that do not show high expression levels in general and are still hard to work with.

With that result at hand we implemented it in the design of our Universal test construct . If you want to read more about the Universal test construct, click here.

Linear DNA Template

It has been shown that DNA templates can be added to cell-free systems as a linear template, instead of going through time consuming cloning rounds. Even though this is a great advantage of such systems, additional steps are needed to effectively use linear fragments due to unwanted DNA degradation through nucleases present in the extract.

To test if we can implement the use of linear DNA fragments into our measuring workflow, we made a naive experiment using a PCR product as a template for our reaction and were pleasantly surprised that expression from the linear template even surpassed the expression from plasmids!
For this experiment we used chloroplast extracts of tobacco and performed the PCR on the plasmid we wanted to compare the linear template to.

Figure 8: Proof of concept: Usage of linear DNA compared to plasmid as a DNA template
Luminescence values are given as arbitrary units and the data is presented on a logarithmic scale. The reaction was set up in a total volume of 10µl. Negative controls using only the plasmid DNA or the crude chloroplast extracts have been included respectively in order to verify the expression

We hypothesize that this phenomenon might be due to supercoiling of the plasmid, which is therefore less accessible for transcription. Either way, we were really happy to see that the linear template DNA resulted in higher luminescence than the plasmid template. This allows future teams to use this method for even more rapid prototyping of parts in different plant chassis, without the need of tedious cloning procedures.

Design, Synthesis and Test of Synthetic 5’UTRs

Every iGEM team has direct access to free DNA synthesis by IDT and Twist. Building on our positive experience with linear DNA templates, we wanted to test whether DNA synthesis could directly be used as a DNA template for the cell-free reactions. This would especially make sense in cases where cloning of the genetic parts of interest would require complicated procedures or if no template for a desired sequence can be found in nature. Thus, we designed, synthesized, and tested fully synthetic 5’UTRs, which do not exist in nature. If you want to know about the design principles we used to create these parts, please go to our Design page.

In short we divided UTRs function in two elements:

  1. translation-initiation
  2. mRNA stabilisation by RNA binding Protein motifs (PPR proteins, read more about PPRs, here)

We combined a synthetic ribosome binding site with different combinations of these RNA binding protein motifs and synthesized them as full transcriptional units, which can be directly utilized for the cell-free measurement.

We are enormously proud to present the first ever fully synthetic 5’UTRs that have been successfully tested in cell-free systems of chloroplasts from Nicotiana tabacum. Figure 9 clearly shows that our different UTR designs result in varying NanoLuc expression levels. Since the constructs only differed in the RNA binding Protein motifs and everything else was kept consistent, we conclude that these RNA binding Proteins highly modulate the expression level in the chloroplast. We propose to use such sequences as design elements to build genetic logic gates by combining various binding sites. This will need to be tested more extensively in the future.

Figure 9: Proof of concept:Synthetic 5’UTRs
Luminescence values are given as arbitrary units and the data is presented on a logarithmic scale. The reaction was set up in a total volume of 10µl. Negative controls using only the plasmid DNA or the crude chloroplast extracts have been included respectively in order to verify the expression

These results make us confident that our cell-free systems from chloroplasts can be used to design, build, and test synthetic chloroplast parts. Synthetic sequences will allow for more complex genetic designs that are diverse and less prone to unwanted homologous recombination, which could occur when using the same part in one genetic system.

One Day Cell-Free Prototyping

Inspired by our successful approach of using linear DNA templates, we wanted to further optimize our workflow of prototyping genetic designs in a high throughput manner. To this end, DNA synthesis is not yet an option for larger libraries of constructs due to budget limitations.

Thus, we successfully developed a workflow that combines our efforts in developing our Golden-Gate based modular cloning toolbox with our linear DNA template approach by skipping the whole in vivo part of the cloning process.

Figure 10: One day cell-free prototyping workflow
The workflow uses direct PCR amplification from the golden gate reaction mixture, instead of conventional in vivo cloning using E. coli

The way we envision it, a PCR is performed directly on the Golden-Gate reaction with standardized primer pairs. A similar approach has already been reported for different assembly types, using PCR amplification after the Golden-Gate to directly use it for Gibson Assembly, allowing to build high level constructs within one cloning cycle [3].
We made use of that approach by directly using the PCR product of the Golden-Gate assembly as a template for measurements in our cell-free systems on the same day. This workflow allows for high throughput testing of chloroplast parts within 7 hours. Starting with ideas for a genetic design in the morning, followed by the Golden-Gate reaction and PCR, we can acquire data from our cell-free systems in the evening of the same day.

We observed expression from our linear templates , which showed part dependent expression levels (Figure 11, 12)

Figure 11: Proof of concept: Testing parts with linear DNA templates, which differ in their 5’UTR
Luminescence values are given as arbitrary units and the data is presented on a logarithmic scale. The reaction was set up in a total volume of 10µl. Negative controls using only the plasmid DNA or the crude chloroplast extracts have been included respectively in order to verify the expression crude chloroplast extracts have been included respectively in order to verify the expression.
Figure 12: Proof of concept: Testing parts with linear DNA templates, which differ in their 3’UTR
Luminescence values are given as arbitrary units and the data is presented on a logarithmic scale. The reaction was set up in a total volume of 10µl. Negative controls using only the plasmid DNA or the crude chloroplast extracts have been included respectively in order to verify the expression

Part Characterization in a Model Organism

Following our optimization efforts, we continued to pursue our the main goal we had - characterization of genetic parts in cell-free systems of chloroplasts. To this end, we performed slightly different measurements than before, using a ratiometric approach that is in more detail described on Measurement page.
We use a dual luciferase method that utilizes a second luciferase reporter next to our main reporter NanoLuc. Our constructs thus harbour a second reporter, namely the Firefly luciferase, that always keeps the same regulatory elements. This allows to normalize the NanoLuc expression to the Firefly expression that should proportionally be the same in all reactions. Thus, we can reduce the variability in our measurements and increase consistency between experiments, which is especially a challenge for cell-free systems due to batch effects noticeable for extractions performed on different days.

Part Characterisation in a Model Organism

Figure 13: SBOL scheme of our dual luciferase construct
A Firefly cassette with fixed regulatory sequences is included in every measurement to normalize the measured NanoLuc expression to the Firefly signal

Figure 14 and 15 show our first part characterizations using this ratiometric approach. We were able to observe a wide range of expression levels depending on the regulatory element used in the plasmid. This clearly shows the intended use of our cell-free systems to prototype genetic constructs in an in vitro setting.

Figure 14: Part characterization of different 5’UTRs via a ratiometric luminescence measurements
Luminescence values are given as a normalized ratio between the Nanoluc and the firefly luciferase The reaction was set up in a total volume of 10µl. Negative controls using only the plasmid DNA or the crude chloroplast extracts can be found in the non normalized data set.
Figure 15: Part characterization of different 3’UTRs via via a ratiometric luminescence measurements
Luminescence values are given as a normalized ratio between the Nanoluc and the firefly luciferase The reaction was set up in a total volume of 10µl. Negative controls using only the plasmid DNA or the crude chloroplast extracts can be found in the non normalized data set.

A dummy sequence was used in both measurements to set the baseline of the expression. This dummy is supposed to serve as a “neutral” DNA sequence not contributing to gene expression. In our dataset, we were able to see that some parts performed particularly well in our cell-free reactions. These parts included the gene10 5’UTR from the T7 phage, the full version of the rbcL 5’UTR from Tobacco, the Synthetic RBS and the TMV 3’UTR.

Part Characterisation in a Crop

Successfully showing that we can prototype and characterize genetic parts in a model species, our next goal was clear to us - we aimed to develop a CFS of an important crop and use it for part characterization.
This is especially of interest since chloroplast transformation methods are not yet available for most of the major crops [4]. The main reasons are a lack of functional genetic parts, missing selection markers and inefficient plant regeneration via tissue culture.
Two of these problems are key elements we wanted to tackle with our project. Firstly, by characterization of genetic parts for the chloroplast of a crop species and secondly by showing that our system can be used to screen possible selection markers for chloroplast transformation.
Read more on how our cell-free systems can be used to screen for efficient selection markers on our integrated human practices page!.

We attempted to develop extracts from chloroplasts of rice, barley, soy, maize, tomato, and wheat. To read more about our efforts to create such extracts, visit our troubleshooting page.

Although we were in the end not able to get working systems for all plants mentioned before, we can proudly announce that we obtained a highly efficient CFS from chloroplasts of wheat, which has to the best of our knowledge never been reported before! We could furthermore demonstrate that these systems can be used to characterize regulatory parts for expression in wheat chloroplasts. For this characterization we made use of the same ratiometric dual luciferase system as described before.

Figure 16: Proof of concept: Part characterization in a crop chassis
Luminescence values are given as arbitrary units and the data is presented on a logarithmic scale. The reaction was set up in a total volume of 10µl. Negative controls using only the plasmid DNA or the crude chloroplast extracts have been included respectively in order to verify the expression

This showcase of our wheat chloroplast CFS is especially valuable, as in vivo characterization is currently not possible in wheat. Wheat is one of many recalcitrant plants whose chloroplast genome has not yet been shown to be reliably transformable and we hope that our efforts will bring the whole chloroplast research community one step closer to this goal.

Species Comparison

Chloroplast gene expression is highly conserved between plant species, due to the singular endosymbiotic event that took place approximately 1.7 billion years ago [5]. After characterizing parts in chloroplast extracts of two different species, we were interested in how transferable chloroplast parts are and how they behave in non-native cell-free extracts. We compared expression of numerous parts in chloroplast cell-free systems of tobacco and spinach. We observed similar behaviour of the parts in both systems, showing that a sequence leading to strong expression in one extract, also resulted in comparably high expression in the other (Figure 17 and 18). Although Spinach chloroplast extracts showed lower ratios of Nluc to Fluc expression, we still see the same trends across both CFS.

Figure 17: Part characterization of different 5’UTRs via via a ratiometric luminescence measurements in two different plant species
Luminescence values are given as a normalized ratio between the Nanoluc and the firefly luciferase The reaction was set up in a total volume of 10µl. Negative controls using only the plasmid DNA or the crude chloroplast extracts can be found in the non normalized data set.
Figure 18: Part characterization of different 3’UTRs via via a ratiometric luminescence measurements in two different plant species
Luminescence values are given as a normalized ratio between the Nanoluc and the firefly luciferase The reaction was set up in a total volume of 10µl. Negative controls using only the plasmid DNA or the crude chloroplast extracts can be found in the non normalized data set.
Figure 19: Heatmap to compare the log normalized rations between tobacco and spinach for the 5’UTR part characterization measurement
Figure 20: Heatmap to compare the log normalized rations between tobacco and spinach for the 3’UTR part characterization measurements

When plotting the ratios in a heatmap, the trends of well-functioning parts become more obvious (Figure 19 and 20). The heatmaps show log normalized ratios of the expression of various constructs in tobacco and spinach extracts. This is done for the 5’ UTR as well as for the 3’ UTR measurements. Darker colours show higher ratios. By simply comparing the colour variations of different constructs, it becomes evident that parts behave similarly in both extracts. The TMV 3’UTR leads to the strongest expression in both the tobacco, as well as the spinach chloroplast extracts (Figure 20). Similarly, the gene10 5’UTR has been shown result in the highest expression in extracts of both species (Figure 19).

Figure 21: Linear regression line to correlate the expression of Spinach and Tobacco

To further investigate the correlation between cell-free extracts from different organisms, we plotted a linear regression line from our data set and could further highlight the similarity between the two results (Figure 21). This regression shows a high correlation between the expression ratios of the two extracts (R² = 0.81). In conclusion, we were positively surprised by the transferability of the genetic regulatory parts even beyond the boundaries of plant species.

This can prove very important when engineering chloroplasts of different species, as existing parts can quickly be screened in a cell-free extract to assess whether they are also functioning in the chloroplast of another plant. If this transferability proves to be more universal and applicable to other organisms, this will drastically expand the range of available genetic part for chloroplasts and could allow future iGEM teams to use any plant chassis they could dream of by utilizing our highly characterized part collection.

Endogenous Transcription

In an inspiring interview with researchers from Northwestern University, we got the feedback from experts that endogenous transcription is key to broad applicability of chloroplast CFS in synthetic biology. Thus, we invested a lot of time optimizing our systems to get endogenous transcription to work.

On our design page we describe several aspects we considered while working towards this goal, as well as how transcriptional regulation works in general within the chloroplast. Ultimately, our efforts were successful, and we are now able to use endogenous promoters from plant chloroplasts to drive transcription in our cell-free extracts (fig. 23)!

After our first proof of concept endogenous transcription cell-free experiments, we kept optimizing and tested out different magnesium concentrations for our Best basic part, the 16s promoter of tobacco in two different extracts from spinach and tobacco (Figure 22). This graph also showed that the 16s promoter seems to behave very similarly in two different chloroplast cell-free extracts.

Additional to the 16s promoter we were able to characterize several other chloroplast-derived promoters in extracts of spinach chloroplasts. Not having to rely on the T7 promoter for transcription is a crucial step in the development of our systems, as this allows us to now design, build, and test whole genetic networks consisting of native regulatory elements. Only using the same, the T7, promoter, this would not be possible and furthermore not be desirable when prototyping for an in vivo project.

Figure 22: Optimization of the magnesium concentration for endogenous transcription
Luminescence values are given as arbitrary units and the data is presented on a linear scale. The reaction was set up in a total volume of 10µl. Negative controls using only the plasmid DNA or the crude chloroplast extracts have been included respectively in order to verify the expression
Figure 23: Characterization of endogenous promoter
Luminescence values are given as arbitrary units and the data is presented on a linear scale. The reaction was set up in a total volume of 10µl. Negative controls using only the plasmid DNA or the crude chloroplast extracts have been included respectively in order to verify the expression

Our Best Composite Part Optimization

We proudly present our best new composite part: The Universal test construct 7.0. After seven rounds of the design-build-test-learn cycle, we have identified the most efficient combination of regulatory parts and reporter genes for the expression in chloroplast cell-free extracts. We hope that future teams could use our Universal test construct 7.0 to develop complete new project ideas, involving any plant chassis they could think of and directly start troubleshooting their extract (Figure 24) preparation without worrying about the optimal DNA construct design in the first place.

Figure 24: Figure 7: Comparing different 3’UTRs in our T7 Universal Test Construct
Luminescence values are given as arbitrary units and the data is presented on a logarithmic scale. The reaction was set up with a total volume of 10µl. Negative controls using only the plasmid DNA or the crude chloroplast extracts have been included respectively in order to verify the expression. These data were extracted from the larger data set of our large scale part characterisation experiment and the other parts were left out to present the data more clearly.

Oak Proof of Concept

After we were able to produce the first chloroplast cell-free extract from Wheat, we highly optimistic in our ability to further implement this technology for chloroplasts of plant chassis that no one would have ever dreamt of being able to use during an iGEM season. We ventured out and gathered branches from the English Oak Tree and tried our luck. To our pleasant surprise, our protocols worked well enough so that we were successful in creating the world's first chloroplast cell-free system from a tree species! In order to prototype our oak chloroplast cell-free extracts, we also just our highly optimized Universal test construct 7.0.

This first proof of concept (Figure 25) was still done using the t7 promoter and two different 3’ UTRs. We could observe that also in this case, the TMV 3’ UTR performed best (as can also be seen in Figure 18), giving us reason to believe that our oak extracts can effectively be used for part characterization, showing similar behaviour as the other extracts created.

Figure 25: Proof of concept: Working chloroplast cell-free extracts of an English Oak Tree
Luminescence values are given as arbitrary units and the data is presented on a linear scale. The reaction was set up in a total volume of 10µl. Negative controls using only the plasmid DNA or the crude chloroplast extracts have been included respectively in order to verify the expression
Figure 26: Proof of concept: Characterizing endogenous English Oak Tree parts chloroplast cell-free extracts of an English Oak Tree
Luminescence values are given as arbitrary units and the data is presented on a linear scale. The reaction was set up in a total volume of 10µl. Negative controls using only the plasmid DNA or the crude chloroplast extracts have been included respectively in order to verify the expression

Because of these promising results, we decided to order a full transcriptional unit as DNA synthesis, in which we used endogenous regulatory sequences from the chloroplast of the English Oak Tree. Using our above-mentioned method of directly using the DNA synthesis as a template in our cell-free reaction, we were able to perform the experiment on the same day the DNA arrived.

It was to be expected that the signal of this genetic design will be a lower compared to our best composite part (T7 promoter), but nonetheless we were able to show expression in our oak chloroplast cell-free system using endogenous regulatory elements from the oak chloroplast (Figure 26).

In Vivo

We stated our goal to be the creation of chloroplast cell-free systems that can be used for prototyping of chloroplast SynBio projects. This builds on the premise that the prototyping performed in our in vitro system is then being transferred into an actual in vivo project.
Thus, we wanted to show that measurements performed in our cell-free extracts are actually comparable to data gathered in vivo. Although performing chloroplast transformation is a tedious process, we nonetheless set out to transform 5 different constructs into the chloroplast of N. tabacum. Due to time constraints and several hurdles, we were not able to obtain full homoplasmic plants that we could use for a proper comparison to our cell-free extracts.
Nonetheless, we managed to obtain calli (Figure 27) that show resistance to the selective antibiotics and will be further cultivated into adult plants!

The first callus we obtained was used for some preliminary experiments. In our fluorescence microscopy analysis, we could verify GFP expression in the callus cells (Figure 28, upper right) co-localising with Fluorescence signal, giving us a first indication that the chloroplast transformation has been successful!

Part of this transformed callus was used to perform a first measurement of the transformed construct harbouring our NanoLuc reporter (Figure 29). Although we cannot show the desired comparison we planned, we are happy to verify NanoLuc expression in our first chloroplast-transformed callus!

In the next we would like to continue the selection process in order to reach homoplasmicity and to regenerate transplastomic plants.

Figure 27: Putative transplastomic calli transformed with our psbB 5’UTR test construct Transformed Calli were selected on Spectinomycin. Positive transformation would result in green Callus.

Fluorescence Microscopy

Figure 28: Fluorescence microscopy pictures of tissue of the putative transplastomic calli transformed with our psbB 5’UTR test construct
Transformed Callus was checked via Fluorescence microscopy. Upper Left Autofluorescence channel of Chloroplast, Upper right GFP-channel, Lower left Transmitted light, Lower right Merged Channels

NanoLuc Measurement

Figure 29: Successful NanoLuc gene expression measurement from transformed calli
The NanoLuc measurement cassette was measured using transformed Calli. Luminescence values are given as arbitrary units and the data is presented on a linear scale. Calli was lysed and incubated in Nano-Glo® Buffer. Tabak leave was used as Negative Control

The road to obtain these first results was long and tedious and filled with complications and setbacks. This truly showed us the value of our project, which will allow others to only bring their projects into the living system once a major part of it has been prototyped in our systems!
Read more about our in vivo journey in our diary!

Antibiotic Sensitivity Assay

In order to test different antibiotics as possible selection markers, we performed an experiment in which different antibiotics were used to inhibit the protein biosynthesis in the cell-free systems, whereby strong inhibition of protein biosynthesis would indicate a suitable selection marker. As displayed in the graph, we screened for different selection markers in N. tabacum and T. aestivum (Figure 30).
Most antibiotics inhibited protein biosynthesis in both tobacco and wheat. In this respect, these antibiotics could be tested with a corresponding resistance in the respective plants.

Ampicillin has an inhibitory effect on cell wall synthesis in gram-positive bacteria, but since chloroplasts do not have a cell wall, the extracts should still be able to carry out protein biosynthesis and luminescence should be seen.
This was exactly what was observed in the experiment. Furthermore, high expression can also be observed in wheat when spectinomycin is added. T. aestivum has a resistance to spectinomycin, contrary to tobacco, validating our data. These results demonstrate the possibility of inhibiting protein biosynthesis in the cell-free extracts using antibiotics and, together with resistance marker genes, can be used for targeted selection.

If successfully established, novel selection markers could therefore enable the selection of homoplasmic tissue cultures in a wider variety of plant species, making the process of engineering them much easier to achieve.

If you want to read more about our IHP effort on implementing the feedback on testing for antibiotic sensitivity, you can read more here.

NanoLuc Measurement

Figure 30: High-throughput antibiotic sensitivity screening using our cell-free systems
Luminescence values are given as arbitrary units and the data is presented on a linear scale. The reaction was set up in a total volume of 10µl. Negative controls using only the plasmid DNA or the crude chloroplast extracts have been included respectively in order to verify the expression

Outlook

Our project ended up to be very successful, as we managed to develop chloroplast cell-free systems from tobacco, spinach, wheat, as well as oak. At the beginning, we would not have hoped to achieve that much and are very happy with the outcome! The results presented here convinced us that our cell-free systems are a very promising tool for many intriguing applications.

First, we hope this technology will be expanded to even more plant species. Obtaining chloroplast extracts of banana, bamboo, sugar cane, as well as sorghum could provide vital testbeds for important engineering efforts. These species, as well as many others, are crucial global crops and some of them are facing dire challenges. Bananas for example, are threatened by several pathogens globally, as most of the bananas traded worldwide are the exact same species stemming from clonal cultures, making them prone to diseases. Providing a cell-free platform for prototyping could allow more rapid research on this important crop. Saving cultivated bananas from going extinct.

Another aspect is the development of novel parts. This should not just include interesting parts to be used in already prominent species (as tobacco), but rather focus on developing new tools for species that are currently rarely used in chloroplast SynBio. One example is rice. So far, there is only one rice promoter being used for rice chloroplast transformation. It is not sufficiently strong to enable proper expression in the rice chloroplast, which is the main hurdle in establishing effective chloroplast transformation in rice [6]. This could be tackled using our cell-free systems, as parts could be characterized before selecting the ideal constructs to use in vivo.

Of course, we would also be delighted to see sophisticated SynBio projects using our system. This can range from the development and testing of extensive genetic circuits to the implementation of metabolic engineering projects first in vitro and then in vivo.

In conclusion, we not only managed to establish chloroplast cell-free systems of four different plants, but moreover constructed 157 parts specifically designed to be used in the plant chloroplast. This will allow others to speed up the work on chloroplast Synthetic Biology, giving them a solid foundation to design, build, and test their own constructs and in cell-free systems.

Sources

  1. USDA Foreign Agricultural Service (2021) Grain: World Markets and Trade https://apps.fas.usda.gov/psdonline/circulars/grain.pdf
  2. Borkowski, O., Koch, M., Zettor, A., Pandi, A., Batista, A. C., Soudier, P., & Faulon, J.-L. (2020). Large scale active-learning-guided exploration for in vitro protein production optimization. In Nature Communications (Vol. 11, Issue 1). Springer Science and Business Media LLC. https://doi.org/10.1038/s41467-020-15798-5
  3. Halleran, A. D., Swaminathan, A., & Murray, R. M. (2018). Single Day Construction of Multigene Circuits with 3G Assembly. In ACS Synthetic Biology (Vol. 7, Issue 5, pp. 1477–1480). American Chemical Society (ACS). https://doi.org/10.1021/acssynbio.8b00060
  4. Rascón-Cruz, Q., González-Barriga, C. D., Iglesias-Figueroa, B. F., Trejo-Muñoz, J. C., Siqueiros-Cendón, T., Sinagawa-García, S. R., Arévalo-Gallegos, S., & Espinoza-Sánchez, E. A. (2021). Plastid transformation: Advances and challenges for its implementation in agricultural crops. In Electronic Journal of Biotechnology (Vol. 51, pp. 95–109). Elsevier BV. https://doi.org/10.1016/j.ejbt.2021.03.005
  5. Douzery, E. J. P., Snell, E. A., Bapteste, E., Delsuc, F., & Philippe, H. (2004). The timing of eukaryotic evolution: Does a relaxed molecular clock reconcile proteins and fossils? In Proceedings of the National Academy of Sciences (Vol. 101, Issue 43, pp. 15386–15391). Proceedings of the National Academy of Sciences. https://doi.org/10.1073/pnas.0403984101
  6. Lee SM, Kang K, Chung H, Yoo SH, Xu XM, Lee SB, Cheong JJ, Daniell H, Kim M. Plastid transformation in the monocotyledonous cereal crop, rice (Oryza sativa) and transmission of transgenes to their progeny. Mol Cells. 2006 Jun 30;21(3):401-10. PMID: 16819304; PMCID: PMC3481850.