Team:Marburg/Engineering

Engineering Success

Abstract

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. This composite part allowed us to develop the first fully functional cell-free extract of English Oak chloroplasts, which would not have been possible without our engineering iterations. We hope that future teams could use our Universal Test Construct 7.0 to develop completely new project ideas, involving any plant chassis they could think of and to directly start troubleshooting their extract preparation without worrying about the optimal DNA construct design in the first place.

The Problem

When we started our project, we had to face a typical “chicken or egg” problem: How would we design our first test construct in order to verify that we have a working chloroplast cell-free extract without having a working extract where we can test if our test construct design works. For this we tried to use one of engineering principles, which was proposed in the early days of synthetic biology: Decoupling. Decoupling is the concept of dividing a complex problem into many simpler problems that can be worked on separately[1]. Although this principle is not 100% translatable to some biological engineering problems, as many things are interconnected and influence each other, it still helped us to overcome the “chicken or egg” barrier. The following basic parts are required in order to build and test a functional transcriptional unit for our chloroplast extracts:

Figure 1: SBOL scheme of a transcriptional unit in the chloroplast Gene organization in the chloroplast. Symbols correspond to a 5' connector, Promoter, 5'UTR, coding sequence, 3'UTR and 3' connector respectively
  1. Promoter
  2. 5’untranslated region (UTR)
  3. Reporter gene/codon optimization of the reporter
  4. 3‘untranslated region (UTR)

For the promoter we decided on using the T7 promoter for our Universal Test Construct, which requires the addition of the T7 polymerase to the cell- free reaction. The usage of the T7 transcription system allowed us to reduce the number of possible problems, due to the fact that we have more control over it as we can just titrate the amount of T7 polymerase for our cell-free reaction.

Reporter System

Fluorescence

The next important design decision was the selection of the reporter system we would use. Therefore, we looked into different fluorescent proteins, such as mScarlet and sfGFP, but also into luciferase-based reporters, like the Firefly and Nanoluc luciferase. After our first measurements, it was clear that fluorescent reporters are not suitable for our cell-free measurements in our chloroplast systems, as the signal can not be distinguished from the noise (Figure 1). On the one hand, this can be explained due to the autofluorescence of the remaining chlorophyll molecules, which increase the fluorescence background levels of the extract control. And on the other hand the expression level of our cell-free reaction seemed to be too low, at least for the beginning, when we were still at the troubleshooting phase of our extract preparation.

Figure 1: GFP as a reporter system for our chloroplast cell-free system Fluorescence 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.

Luminescence

Therefore, we went back to the drawing board and decided to further look into luminescence-based reporters. The next experiment we performed was the comparison of the Firefly with the Nanoluc luciferase. Generally, it can be observed that the background of luminescence is very low in the control, compared to fluorescence measurements (Figure 2). Additionally, high luminescence levels can be observed, which allows the detection of successful expression from our chloroplast cell-free extracts. When comparing the two different luciferases, Nanoluc shows a superior signal overall and an improved signal-over-noise ratio. Hence, we selected the Nanoluc luciferase as our reporter system for our Universal Test Construct for all our future experiments.

Figure 2: Comparison of different luciferase reporters in a tobacco chloroplast cell-free system 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.

Proof of Concept Cell-Free Extract

This first version of our Universal Test Construct allowed us to develop, prototype and troubleshoot our first cell-free systems of tobacco and spinach (Figure 3). But we did not want to stop there and went though more rounds of the design-build-test-learn cycle and aimed to optimize this Universal Test Construct further, with the goal in mind to have a construct at hand which can even work for cell-free systems of non-model plant species.

Figure 3: Proof of concept: Developing working chloroplast cell-free extracts of tobacco and spinach 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.

Codon Optimization

As we have now decided for the Nanoluc luciferase as a reporter we wanted to test if different codon optimizations would have an effect on the expression level in our system. Surprisingly, we found that our Nanoluc part, which is optimized for the chloroplast of Chlamydomonas reinhardtii, shows higher expression in a tobacco cell-free system, than a Nanoluc part specifically optimized for the chloroplast of Nicotiana tabacum (Figure 4). There are many possible explanations for that, one example would be the mRNA structure at the beginning of the coding region of the optimized Nanoluc parts. Due to the fact that we were just interested in the optimization of our construct for high expression, we chose the version which is optimized for Chlamydomonas reinhardtii for our next version of our Universal Test Construct.

Figure 4: Comparing different codon optimization of the Nanoluc luciferase 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.

Regulatory Elements

5'UTR

The next element we aimed to optimize was the selection of the 5’UTR. For this we looked into literature again and compared two different 5’UTRs, which have been reported to enable high expression in vivo in the chloroplast of tobacco [2, 3]. When comparing these two 5’UTRs it can be observed that the 5’UTR of gene10, which originates from the E. coli T7 phage, shows far higher expression than the endogenous 5’UTR of rbcL (the large subunit of the RuBisCo). With that knowledge in mind we designed the next version of the Universal Test Construct with the aforementioned phage part.

Figure 5: Comparing different 5'UTR of the 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.

3'UTR

The last part of the construct we wanted to optimize was the 3’UTR. For this we even went through two rounds of engineering and optimization. In the first experiment we compared the commonly used 3’UTR rrnB, which originates from E.coli, with the psbA 3’UTR (photosystem II D1 protein) of Nicotiana tabacum. In this experiment it can be observed that our construct can already be significantly improved by exchanging the rrnB 3’UTR to the psbA 3’UTR.

Figure 6: Comparing different 5'UTR of the 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.

Viral 3'UTR

At that point we thought we had the best possible Universal Test Construct for future screening of non-model cell-free extracts. But when we did the extensive part characterization of different 5’ and 3’ UTR (Results), we were able to identify another even better 3’UTR (Figure 7), which again comes from a viral background, but in this case from the tobacco mosaic virus. We were very surprised by the results, as there was no experimental evidence in literature that this part would be functional at all. This again highlights the big advantage of our high-throughput part characterization approach.

Figure 7: Comparing different 3’UTRs 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.

With that data at hand we built our final version of our Universal Test Construct 7.0 which went through 7 rounds of the design build test learn cycle for the most optimal construct for chloroplast cell-free extract prototyping. This story teaches us how important the engineering principles like standardisation and decoupling are, they allowed us to treat every part position in our test device as a seperate engineering problem.

When we finally had our Universal Test Construct 7.0 at hand, we wanted to use it for something really ambitious, an idea we had from the very beginning: how exciting would it be, if we could develop a cell-free system of chloroplasts of tree species. Climate change is affecting our forests drastically and is causing forest dieback. Designing, building or even testing BioBricks for a tree seems impossible in the time frame of an iGEM project. But imagine if it would be possible! That is why we made it our duty to establish a cell-free system of the English Oak, knowing that this bears many more challenges than common model plants, as none of the protocols for the extract preparation are optimized for these species.

English Oak Cell-Free Extract

We were super excited about our results. After all that optimization of our Universal Test Construct, we were able to develop a fully functional chloroplast extract of the English Oak tree and were even able to test other parts in this system. This would not have been possible without all of these iterations of the design-build-test-learn cycle, as even our version 5.0 of the test device showed far less expression.

Figure 8: Successful Expression using the Universal Test Construct 7.0 in a Cell-Free Extract of Quercus robur 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 160µl. Negative controls using only the plasmid DNA or the crude chloroplast extracts have been included respectively in order to verify the expression. This measurment has been performed using our best composite part Universal Test Construct 7.0 and comparing it to the version Universal Test Construct 5.0

Conclusion

All in all, we hope that future teams could use our knowledge, but especially our latest version of our Universal test construct 7.0 to develop completely new project ideas, involving any plant chassis they could think of. By using our best new composite part, future teams can already skip the “chicken or egg” problem, which we described at the beginning, and start troubleshooting their extract preparation without worrying about designing-building and testing their own test device in the first place.

Sources

  1. Endy, D. (2005). Foundations for engineering biology. In Nature (Vol. 438, Issue 7067, pp. 449–453). Springer Science and Business Media LLC. https://doi.org/10.1038/nature04342
  2. Kuroda, H. (2001). Complementarity of the 16S rRNA penultimate stem with sequences downstream of the AUG destabilizes the plastid mRNAs. In Nucleic Acids Research (Vol. 29, Issue 4, pp. 970–975). Oxford University Press (OUP). https://doi.org/10.1093/nar/29.4.970
  3. Eibl, C., Zou, Z., Beck, andreas, Kim, M., Mullet, J., & Koop, H.-U. (1999). In vivo analysis of plastid psbA, rbcL and rpl32 UTR elements by chloroplast transformation: tobacco plastid gene expression is controlled by modulation of transcript levels and translation efficiency. In The Plant Journal (Vol. 19, Issue 3, pp. 333–345). Wiley. https://doi.org/10.1046/j.1365-313x.1999.00543.x