Design
I – Chlamydomonas
Chlamydomonas reinhardtii is a unicellular green microalga that lives in freshwater.
Microalgae are photosynthetic organisms that have a huge potential for synthetic biology as they can be used to produce human therapeutic proteins (Rasala and Mayfield 2011), biofuels (Scranton et al. 2015) or sugar in a more sustainable way than common heterotrophic chassis. Indeed, as microalgae are photoautotrophic, they need much less inputs than heterotrophic chassis, and they can be grown in marine water or photobioreactors thus not competing with arable land.
Initially our project was to make a proof of concept of an engineered microalgae for sugar production in marine water. The goal was to fix carbon as carbohydrates to feed large scale fermentation with a minimal impact on the environment. We chose the microalgae Chlamydomonas reinhardtii as chassis even so it is a unicellular green microalga that lives in freshwater as Lachapelle and colleagues adapted through long term selection it to marine water (Lachapelle, Bell, and Colegrave 2015).
Moreover, the sequencing of the Chlamydomonas genome is complete since 2007 (Merchant et al. 2007) and several molecular tools are available to perform transformation and edition of the genome (Jinkerson and Jonikas 2015). More recently, a Modular Cloning (MoClo) kit has been generated allowing combinatorial assembly and faster building than other cloning methods (Crozet et al. 2018). The combination of these tools should allow efficient engineering of Chlamydomonas and a great potential to the development of our project
One of the most powerful tools for metabolic engineering is directed evolution. This is usually done by an in vitro library generated by error-prone PCR and cloned by restriction enzyme into a plasmid that is transformed with high efficiency into competent cells. However, “PCR-mediated directed evolution” cannot be done effectively and easily on slow-replicating and non plasmid bearing eukaryotes such as microalgae. To date, no study of directed evolution in vivo in microalgae was ever reported.
Crook et al. developed in vivo directed evolution mediated by retrotransposons (Crook et al., 2016). They synthetically optimized the retrotransposons Ty1 to enable in vivo generation of mutant libraries in the yeast S. cerevisiae. When coupled to growth selection, this approach enabled in vivo continuous evolution (ICE). We chose to adapt this strategy to the microalgae Chlamydomonas reinhardtii. We aim to go further and improve this work (that was also the project on the Vienna BOKU team in 2017) by creating a synthetic retrotransposons completely modular and compatible with the MoClo standard. We hope this work will help spreading in vivo directed evolution in algae among fundamental scientist or for industrial application.
II – Modular Cloning (MoClo)
As we want the retrotransposon to be fully modular, we used a standardized DNA strategy called Modular cloning (referred to hereafter as MoClo). Both BioBricks and MoClo allow the assembly of basic parts such as promoters, CDS or terminators but the MoClo relies on the Golgen Gate cloning strategy (Weber et al. 2011).
Type IIS restriction enzymes
Restriction enzymes are endonucleases which catalyze the hydrolysis of a double stranded DNA resulting in a double strand break at a specific sequence.
The classical restriction enzyme (e.g. EcoRI) cut inside their recognition site. They released a part of their recognition site and have to be ligated with DNA piece cut with the compatible restriction enzyme. After the ligation, the restriction site remains. (Scheme 3)
The type IIS restriction enzyme (BpiI, BsaI) cut downstream of their recognition sites and generate 4 bp 5’-overhangs (Engler et al. 2009) that are not specific of the restriction enzyme (Scheme 3).
One-pot, one-step reaction (Engler et al. 2009)
The golden gate cloning allows a directional assembly of multiple DNA fragments into a one-pot, one-step reaction. Indeed, the restriction enzyme and the ligase are added at the same time as there are no scars of the restriction site after the ligation when type IIS enzymes are used, making the final product stable during the reaction. The two enzymes (digestion and ligation) do not work at the same temperature (37°C for the restriction enzyme and 16°C for the ligase). To increase the cloning efficiency, several cycles (37°C-16°C) are made.
MoClo assembly standard (Scheme 4),
In the MoClo assembly standard (Weber et al. 2011), the smallest units are basic genetic parts such as promoters, CDS, 5’UTR. Each type of basic part have a position defined by fusion sites on both sides. Also, as all the same basic parts have a unique fusion site there are commutable at will. These basic units are cloned into plasmids with BpiI and are called the level 0. A level 0 acceptor plasmid contains BpiI restriction sites flanking a lacZ cassette, a pUC origin of replication, and the spectinomycin resistance gene. Upon the digestion by BpiI, the plasmid releases the lacZ cassette and the part of interest hybridizes to the plasmid through complementary of these flanking 4 bp overhangs, corresponding to the fusion site. This implies that there is a specific acceptor plasmid for each position within the standard. These molecules are then ligated and the final products are used to transform E. coli (LacZdelta). The selection of the transformants carrying level 0 plasmids is made by blue/white selection (pick white) on a LB medium with X-gal and spectinomycin.
All the level 0 plasmids also contain a BsaI recognition site that allows a second golden gate reaction to create an entire transcription unit (level 1) from several level 0 plasmids (Weber et al. 2011). This is the specificity of the overhangs which define the order of the parts into a transcriptional unit. And we can screen for the level 1 plasmid because they have the resistance gene of the ampicillin (and not spectinomycin as the level 0).
A full transcription unit can be made by only choosing the basic units. It also allows to test a module library in the same transcription unit and thus screen the best module for a specific application (e.g. different promoter for a coding sequence). It is very useful to optimize complex pathways in metabolic engineering.
For Chlamydomonas reinhardtii, a specific MoClo toolkit has just been reported (Crozet et al. 2018), and we used their standard for the construction of the retrotransposon, as well as some parts of this kit
III – MoClo Synthetic Retrotransposon
A- Retrotransposon Generality
Retrotransposons are Mobile Elements of the genome. There are classified in two main categories: class I and class II mobile elements. The latter have a RNA intermediate in opposition to class I that have a DNA intermediate. The retrotransposon uses a “copy and paste” mechanism where the DNA is first transcribed in RNA, then reverse transcribed in DNA and finally inserted at the target site (Piégu et al., 2015). Scheme transposition
B – The choice of the Retrotransposon
There are few LTRs retrotransposons described in green algae. There are REM1 (related to Ty3/gypsy retrotransposon) and TOC1 retrotransposons in Chlamydomonas reinhardtii, and OSSER (related to Ty1/copia retrotransposon) in Volvox carteri (Sormacheva and Blinov 2011). As TOC1 belongs to the DIRS-I family and transposes via a circular intermediate (Kim, Kustu, and Inwood 2006), we chose to focus on REM1 and OSSER retrotransposons that are classical RNA intermediates retrotransposons. The main difference between these two is the order between the reverse transcriptase (RT) and the integrase within the POL gene (Sormacheva and Blinov 2011). REM1 is a functional retrotransposon in Chlamydomonas reinhardtii but it has some structural features not described yet. Indeed, the typical RNAse H and integrase have not been identified, and a third ORF is in the reverse orientation between the 3’LTR and the ORF1 of GAG. Moreover, the copy number present in the genome vary depending on the strains which suggests different regulation mechanisms (Pérez-Alegre, Dubus, and Fernández 2005). OSSER retrotransposon (Scheme 1) is similar to other copia retrotransposons already known, with a central domain coding for GAG and POL flanked by a direct repeat of 197 bp. The only specificity of OSSER is a one bp insertion in the middle of the RT, which is a translational slippage (Lindauer et al. 1993). We haven’t found data about its transposition rate.
These data led us to choose OSSER as retrotransposon of our system, knowing the risks it could not be functional, however Volvox is a very close relative to Chlamydomonas and orthogonality of DNA parts between species is expected.
C – Design of our System
The purpose of our system is the induction of mutations via the error rate of the reverse transcriptase into a gene (named hereafter CARGO) inserted in the redesigned retrotransposon. We were inspired by the article of Crook and colleagues which describes this strategy to perform in vivo evolution in yeast (Crook et al. 2016).
We designed the synthetic retrotransposon to be fully modular. This was quite challenging in term of MoClo design as we had to build a library of the different parts composing the synthetic retrotransposon: pNIT1, 5’UTR of NIT1, 5‘LTR, GAG-POL, pPSAD, 5’UTR of PSAD, paromomycine resistance CDS (AphVIII), the 1st intron of CrRBCS2, 3’UTR of PSAD, and 3’LTR. Instead of using level 1 plasmids to assemble transcription unit, we split the synthetic retrotransposons in two level 1 plasmid pL1-1F and pL1-2R. The inserts of these two plasmids can then be linked to a recipient plasmid forming a final construct (level M) giving the full synthetic retrotransposon.
Our synthetic Osser retrotransposon consists of different main parts and features (Scheme 2A):
- The 5’ LTR: The main difficulty was to identify the wild type transcription regulators inside the 5’LTR. The purpose was to remove it and put Osser under the control of an inducible promoter pNIT. There are three regions: U3 with the TATA box, R with the transcription initiation and U5 (Curcio et al., 2015) . We manually annotated the LTR prior to design. The first feature we used to delimit U3 from R is the Polyadenylation site (PAS). PAS are both in 5’LTR and 3’LTR because those two sequences are strictly identical. Interestingly, the Osser PAS from Volvox carterii UGUAA is the same in Chlamydomonas reinhardtii. We truncated U3, 22 bases before the PAS motif. We linked the inducible promoter pNIT1 with the fusion site B2 (CCAT).
-The GAG-POL part contains all the activities necessary for the retrotransposition (RNA binding protein for gag, reverse transcriptase, integrase and RNAseH for pol). We did not change anything in this sequence. One of the project’s perspective will be to add all the GAGPOL element as separated parts in the MoClo standard to increase modularity.
- Based on Crook et al design we inserted in reverse a whole transcription unit between the end of GAGPOL and before the 3’LTR. This Cargo contains a resistance gene under the control of a constitutive promoter (pPSAD) and an intron (RBCS2i1) in the reverse orientation compared to the cargo. The orientation of the intron inserted at the beginning of the sequence of the resistance gene will avoid expression of the cargo without retrotransposition. Indeed, it is only after RNA splicing that the cargo will have an ORF that can be translated. We chose to use the B3 fusion site which contains the ATG codon initiator. In order for the paromomycine resistance gene (or any other CDS) to be in frame with this ATG, once the intron is spliced we added 2 nucleotides after the B4 site. Moreover, it is important to note that the PSAD 3’ UTR and terminator part should not have bidirectional terminator activity.
-To allow selection of transformants of the transposon, another resistance gene will be coupled to these TUs, the hygromycin resistance module. This module expresses under the control of the pAR strong and constitutive promoter and the 3’UTR/terminator of RbcS2 (Schroda et al Plant Journal, 2002), the AphVII CDS coding for an aminoglycoside phosphotransferase (Berthold et al, Protsist, 2012).
-We inverted the 3’LTR in pL1-R so it will be in 5’-3’ direction in the final construction.
D – Proof of Concept
Once the construction will be obtained, we want to test its activity with our reporter system which relies on the paromomycine resistance gene (controlled by the pPSAD constitutive promoter) with the intron in reverse that disrupt the transcription frame. The main steps are: 1/ Transformation of WT Chlamydomonas reinhardtti, 2/ selection of hygromycine resistant clones, 2/ culture in TAP with ammonium until early log phase (106 cells/mL); 3/ Induction of the pNIT promoter by changing the nitrogen source (TAP with nitrate); plating the microalgae on TAP + paromomycine medium. The number of resistant colony will depend of the transposition rate.
Highlight of the reporter system to test the synthetic transposon activity:
Scheme 2B:
1 – The pPSAD promoter is in the reverse orientation than the intron. When the resistance gene is transcribed, the protein cannot be produced because the intron cannot be spliced.
2 – If the pNIT1 promoter controlling GAG-POL expression is active (nitrate as nitrogen source), GAG-POL and the cargo are transcribed as a single mRNA allowing the intron to be spliced. However, the resistance gene is still untranslated because it is not in the correct orientation.
3 – If the transposition works and the resistance gene (which has undergone splicing and reverse transcription) is reinserted in the genome, the pPSAD promoter will allow constitutive expression of the paromomycine resistance, now without intron and thus giving a fully functional protein.
To check if the transposition is working well, we designed a cargo with the intron before the paromomycin resistance CDS. Transposon-activated (through pNIT1) lines selected from the first screen based on hygromycin resistance will be grown in a medium supplemented with paromomycin. Thereby, the only colonies that should be able to grow would be the one having undergone the transposition. We plan to sequence the locus of insertion and verify the integrity of the transgene.
Once the proof of concept will be done on the retrotransposon, we want to engineer a simple metabolic pathway composed of two enzymes producing trehalose (otsA and otsB, see IV – trehalose and biosensor) instead of the paromomycin resistance in the cargo. This way, in vivo directed evolution can be performed on these enzymes.
IV – Trehalose and biosensor
To produce trehalose in Chlamydomas, we chose to use the otsA and otsB genes from E. coli. Indeed, they have already been described to be functional in land plants such as Arabidopsis thaliana (Schluepmann et al. 2003; Wingler et al. 2012). We designed them to be in a unique transcription unit, expressed as a cistron-like thanks to the 2A peptide from the foot-and-mouth-disease virus (FMDV) linking the 2 coding sequences. This way, they will be under the control of a unique promoter and co-expressed, in the same amount (Wang et al. 2015). Upstream and downstream in the ORF, we added tags to easily detect the production of each proteins.
To be able to detect the production of trehalose and thus the functionality of this metabolic pathway, a trehalose biosensor was also adapted to C. reinhardtii. This biosensor is a modified GFP coupled with a protein that can bind specifically to the trehalose. As the protein binds it, the GFP conformation is modified allowing it to be functional (Nadler et al. 2016).
