Natural competence

Reintroducing natural competence into our S. elongatus strain was an important goal for us. To make sure that our S. elongatus UTEX 2973 actually holds the point mutation in the pilN gene we thought it has, we sequenced this region - and the results showed the expected mutation

As there were multiple methods at hand that we could use to get our strain naturally competent again, we tried all those at hand, making sure we do everything we can to tackle this issue.
Although not our favorite method, we tried integrating an intact copy of the pilN gene into neutral site II of S. elongatus UTEX 2973 following the example of Li et al., 2018 . We received pSII-trc-pilN, the same plasmid used by Li et al., as a gift from Petra Wurmser from the research group of Prof. Kaldenhoff in Darmstadt and conjugated it into our strain via triparental conjugation.

Beforehand, Petra Wurmser told us that she was not able to successfully reproduce the experiment, motivating us to try it ourselves. We made sure to follow the protocol in the above mentioned paper, transforming pRL443 and pRL623 into E. coli HB101 and pSII-trc-pilN into E. coli HB101. Overnight cultures of these cells were inoculated and grown to OD600≈0.5. After washing and incubating them together for half an hour, they were mixed with a exponentially growing S. elongatus culture and incubated for 30 minutes again. Thereafter the mixture was blotted on sterile filters and incubated on BG11 plates for 24h before being transferred onto BG11 plates containing kanamycin (Li et al., 2018). It is important to add that we also went with another approach for conjugation: Martina Carrillo Camacho from the working group of Prof. Dr. Tobias Erb provided us with the pRK2013 plasmid, mentioning that she uses it for conjugation. So we transformed the plasmid into E. coli DH5ɑ and pSII-trc-pilN into HB101, performing the same procedure with them as stated above.
After a whole week we could actually see growing colonies, though they were only from attempts with the pRK2013 plasmid.

We were still excited, directly starting liquid cultures and running colony PCRs in hope to find our desired result. Although trying a wide variety of primer combinations, we were not able to find any successful integrations, but as the strain was growing on kanamycin and colony PCR does not always work correctly, we wanted to make sure and tried an actual transformation: A YFP construct was transformed into our seemingly competent strain, allowing for easy selection afterwards. Disappointingly we could not transform the strain. This was in accordance to the results of Petra Wurmser, who was also not able to reintroduce natural competence into S. elongatus UTEX 2973 through this method.
In hope of better results we decided to try and revert the point mutation in the pilN gene with a CRISPR/Cas12a system.While still working on the system itself as described in the design section, we used the pSL2680 plasmid (Ungerer and Pakrasi, 2016) to engineer our strain. We followed their protocol (available here on Addgene) to construct a guiding crRNA, leading the Cas12a enzyme to the mutated pilN gene and cloned it together with the enzyme and a repair template to revert the point mutation. Sequencing results of all parts needed were correct, so the final construct could be transformed into our cyanobacteria. In this approach triparental conjugation was performed slightly differently, as we wanted to exactly follow the provided protocol: In contrary to the other method, a HB101 strain harboring pRL623 was again transformed with the plasmid of interest, in this case the fully assembled pSL2680 and the other HB101 strain just contained pRL443. Again, both strains were mixed together and then inoculated before being blotted on filters on BG11 plates. This time, after only 6h the filters were transferred on BG11 plates with kanamycin. During our first run a few colonies were visible after just three days, but getting cultures to grow in liquid media did not work as expected and the colonies did not look as healthy as hoped. As we have had this kind of issue several times during our project and saw that in literature the colonies growing on the plates looked differently, we decided to reach out to experienced researchers working with the same Synechococcus elongatus strain. We contacted Prof. Dr. James W. Golden from the University of California, San Diego to ask for advice and were able to discover a crucial difference in the way we work with our strain: While in our own lab this is not commonly done, the researchers from the Golden lab put Na₂S₂O₃ (sodium thiosulfate) in their BG11 plates and we could find literature to support this advice (Thiel et al., 1989). This might seem like a small addition, but it proved to be an essential factor, as after adding this ingredient to our own plates we were able to grow way better looking colonies, and not just that - the colonies started coming up after just one day in our incubator. The reason is simple: When autoclaving agar that contains phosphates, reactive oxygen species such as H₂O₂ can be formed, which can drastically hinder bacterial growth (Tanaka et al., 2014). The thiosulfate from Na₂S₂O₃ is oxidized as H₂O₂ is reduced to 2 H₂O - so clearly thiosulfate can act as a reducing agent, eliminating reactive oxygen species from the medium. Presumably the missing sodium thiosulfate in our BG11 agar was not the sole reason we did not get the desired results on our plate, but in combination with the toxicity of Cas12a (though less toxic than Cas9, it still exhibits nuclease activity, cutting the DNA of the bacteria) the pressure was too high for most of the cells. With our new found ingredient we were certain to get it right this time and set the whole process in motion, this time with BG11 plates including sodium thiosulfate. In one huge experiment we tried a wide variety of different conjugation approaches, following the instructions of multiple papers ( Ungerer and Pakrasi, 2016; Yu et al., 2015 ; Wendt et al., 2016). We are currently screening for correct edits and are certain to hold a naturally competent strain in our hands in a matter of days.

CRISPR gene editing

Although CRISPR/Cas systems have been discussed as incredibly powerful tools in genetic engineering, they have not yet been widely used in cyanobacterial research, which is why we set out to implement such a system, based on CRISPR/Cas12a, into our Green Expansion of the Marburg Collection.
As CRISPR/Cas12a has already been reported to work in S.elongatus UTEX 2973 (Ungerer and Pakrasi, 2016), we were sure that it could be transformed into a Golden Gate Assembly compatible version, allowing for more flexible design considerations. While we started the cloning processes needed to change the existing vector into the phytobrick standard, we tried the vector at hand ourselves, in order to assess its usefulness. Following the given protocols we constructed a CRISPR/Cas12a vector harboring a crRNA and repair template designed to revert the point mutation in the pilN gene of our S.elongatus strain. After a few initial problems we were able to get conjugants and are currently screening for those containing the desired edit - more on this approach can be found in the natural competence section of our results.
In order to modularize this system we built different parts for our genetic toolbox. First of all we created a lvl0 part of the Cas12a protein by amplifying the sequence from the pSL2680 plasmid, including overhangs that enabled us to clone the PCR product into a lvl0 acceptor vector.

Sequencing results proved, that this crucial part was correctly assembled, ready to be used in lvl 1 constructs - which we promptly did, using the following lvl0 parts: pMC0_1_03 + pMC0_2_03 + pMC0_3_07 + pMC0_4_33 + pMC0_5_07 + pMC0_6_17.

The chosen promoter is a rather weak one, so that overproduction of Cas12a is prevented, leading to less toxicity in the cells.

Having built this construct, we continued to build the other missing part: the crRNA. The design of the pSL2680 plasmid was mostly kept the same, but in order to have an easy and cheap selection method we switched the lacZ cassette with a GFP cassette.

We could show the correct assembly of this part - everything was as we planned in our design meaning that we had all the parts in our MoClo standard.

As the whole system is built for modular cloning in the PhytoBrick syntax, it is possible to freely exchange the parts around the Cas12a and crRNA parts. This enables the use of different promoters, allowing for easy screening: Constructs with weaker promoters in front of the Cas12a gene would lead to less gene expression and therefore lower toxicity of the whole system. The free exchange of these promoter parts can consequently be used for the creation of a library in order to look for the perfectly fitting promoter for this system. Successfully creating these invaluable parts, we were able to establish a workflow for faster cloning in S.elongatus. As our system is modularized, it is possible to easily exchange the GFP cassette for the desired crRNA, which can be done in a single reaction, further simplifying the cloning process of CRISPR/Cas12a constructs. As shown before , the cloning process with the pSL2680 can take over a week, is tedious work and is accompanied by another couple of days waiting for colonies. In comparison, our system enables for efficient cloning in only four days: On the first day the construct is assembled in a Golden Gate reaction, which is thereafter transformed into E.coli. The next day colonies can be picked, inoculated and the construct can be extracted in the evening. On the third day it can be transformed into S.elongatus - and on the fourth day colonies can be screened. The missing piece to apply an edit is the repair template. Skimming through literature, we noticed that transformation of linear DNA fragments into S.elongatus is supposedly more efficient than the transformation of whole plasmids (Almeida et al., 2017) - and we were able to verify this fact in our own experiments. This further simplifies our above mentioned workflow, as we are able to simply PCR the needed repair template from a DNA sequence and use the PCR product for transformation into S.elongatus. Our toolbox has a special feature that can be used for exactly this workflow: a NotI cutting site can be found in our constructs, which is used to linearize them, so that they can be more efficiently transformed.
We are more than certain that our modular CRISPR/Cas12a proves to be an invaluable contribution to the tools available in cyanobacterial research, especially for the Golden Gate community, which is growing bigger and bigger every year - also thanks to the iGEM headquarters finally integrating the TypeIIS standard into the competition!

Cyanobacterial shuttle vectors

As we have already clarified in the description part, self replicating shuttle vectors are essential for many workflows, as the gene expression levels are higher and non of the tedious selection processes that come with genomic integrations have to be done.

On our road to the modular vector we were seeking, we firstly cured our own S. elongatus UTEX 2973 strain of its pANS plasmid. This was done by transforming the pAM4787 vector, which holds a spectinomycin resistance as well as a YFP cassette (Chen et al., 2016). Due to plasmid incompatibility - explained here in our design section - and because antibiotic pressure is applied, the pANS plasmid was over time cured from the strain, which then just kept the pAM4787 plasmid. Transformation was done by conjugation with the pRK2013 plasmid in DH5ɑ and the pAM4787 in HB101. Both were grown to an OD600≈0.5, washed in LB and mixed with S. elongatus which was grown to late exponential phase and then washed in BG11. We could clearly show, that the conjugant strain bears the pAM4787 plasmid if selective pressure is held up.

This was followed by us starting to culture the pAM4787 bearing strain without antibiotics again, slowly removing selective pressure from the cells. As the plasmid does not give them any other advantage and is probably just more metabolic burden due to the constantly produced YFP proteins it is slowly being lost. We could prove this in multiple setups: with the flow cytometry device we were kindly granted access to we could clearly show the missing YFP signal in the cured S. elongatus strain and logically this could also be observed over our UV table.

Furthermore we performed colony PCRs as a test. We sent our plasmid-free strain to Next Generation Sequencing in order to ensure that the strain really has lost the pANS plasmid.

Our next step was the characterization of the cyanobacterial shuttle vector mentioned in our design section. In an extensive flow cytometry experiment we assessed the fluorescence of a transformed YFP-construct in our cured strain, showing that the shuttle vector with the minimal replication element can be maintained in S. elongatus UTEX 2973.

After another four weeks of cultivation we looked at our cultures again on the UV table to check if fluorescence was still present and the high intensity of fluorescence proved to us, that the plasmid is still stably replicated in our strain, showing us, that the minimal replication element does indeed work in our strain. For further analysis we performed qPCR with this transformed strain, in order to check the copy number of the vector. We used the copy number of pANL as a reference, which is supposedly at ~2,6 copies per chromosome (Chen et al., 2016). Our data shows a ~4,5 times higher copy number relative to pANL, meaning that the construct is maintained with approximately 11,7 copies per chromosome.

For further analysis of this part (BBa_K3228069) we performed a Quantitative-Polymerase-Chain-Reaction (qPCR) with this transformed strain, in order to check the copy number of the vector. This proved to be difficult in Cyanobacteria, due to variation in genome copy number (Griese et al., 2011; Chen et al., 2012). To overcome this problem, we noticed that the copynumber of pANL stays rather stable with 2,6 copies per chromosome (Chen et al., 2016). Therefore we performed the qPCR with pANL as an additional target. For each target, we chose three different primer pairs, with known efficiency. Five technical and two biological replicates were used. The samples were normalised to the chromosome and pANL was used to identify the total copy number per chromosome (figure 14). Our data shows a ~4,5 times higher copy number relative to pANL, meaning that the construct is maintained with approximately 11,7 copies per chromosome. This is comparable with the copynumber of pANS in Chen et al. 2016.

Additionally we measured the fluorescence signals in a plate reader at different optical densities and could again confirm high fluorescence signals, indicating strong gene expression in constructs built around this replication element.

All this data confirms that the construct actually works and can be reliably used as a cyanobacterial shuttle vector, proving that BBa_K3228069 works as intended, thus functioning as our validated part. This assumption is solidified by all our sequence data, showing that the shuttle vectors were completely assembled as planned in our design section.

Overview over the expansion of the Marburg Collection:

We added 55 new parts to the Marburg Collection, adding several new features such as the Green expansion, including a kit for the Modularized Engineering of Genome Areas (M.E.G.A.) and the first MoClo compatible shuttle vector for cyanobacteria. Additionally, we offer a set of reporters suitable for characterization of BioBricks in cyanobacteria and ribozymes for a more stable and species independent transcription. We also provide standardized measurement vectors that were generated using our designed placeholders.

Overview over the different expansions in the Marburg Collection 2.0

To give a better overview we show here the different expansions we added to the Marburg Collection:

Parts

K3228000 (aNSo1 integration up) K3228001 (aNSo2 integration up) K3228020 (NS1 up) K3228021 (NS2 up) K3228022 (NS3 up)	K3228004 (shortTB0010) K3228005 (shortTB0015) K3228006 (shortTB1002) K3228007 (shortTB1003) K3228008 (shortTB1004) K3228009 (shortTB1005) K3228010 (shortTB1006) K3228011 (shortTB1007) K3228012 (shortTB1008) K3228013 (shortTB1009) K3228014 (shortTB1010) K3228015 (shortDummy)	K3228016 (SpecRes_short) K3228017 (CmlRes_short) K3228018 (TetRes_short) K3228019 (KanRes_short) K3228027 (GenRes_short)	K3228002 (aNSo1 integration down) K3228003 (aNSo2 integration down) K3228023 (NS1 down) K3228024 (NS2 down) K3228025 (NS3 down)	K3228026 (oriT integration) K3228069 (panS SpecRes LVL1) K3228089 (panS KanRes LVL2)	K3228051 (Limonene Synthase) K3228052 (Farnesen Synthase)

K3228054 (riboJ B0034) K3228055 (riboJ B0034 noScar) K3228056 (SarJ B0034) K3228057 (PlmJ B0034)	K3228060 (PRO placeholder) K3228061 (RBS placeholder) K3228062 (CDS placeholder) K3228063 (TER placeholder)	K3228073 (Promotor entry vector) K3228074 (BS entry vector) K3228075 (CDS entry vector) K3228090 (Terminator entry vector)	K3228100 (cpf1) K3228101 (crRNA-GFP dropout) K3228103 (cpf1+crRNA-GFP dropout)	K3228040 (NanoLuc) K3228041 (NanoLuc S.e.) K3228042 (TeLuc (S.e.)) K3228043 (Antares S.e.)	K3228044 (eYFP S.e.) K3228045 (sYFP2 S.e.) K3228046 (phluorin2 S.e.) K3228047 (rxYFP S.e.) K3228048 (sfGFP S.e.) K3228049 (mTurquoise S.e.)

Sequencing results of the lvl0 parts

We built and validated 55 new BioBricks this year. They are all listed in the Registry of Standard Biological Parts (Part range BBa_3228000 to BBa_32280103). All lvl0 Parts were validated by complete sequencing.

Building constructs to test the lethality of origin of transfer

If plasmids reach a certain size normal transformation protocols are not feasible anymore to bring the plasmid into the host.
For the transformation of such huge megaplasmids we designed an “origin of transfer” BioBrick that makes it possible to directly transport plasmids of any size from one species to another. To test if this sequence would result in any toxicity in a genomic context (source things where genome parts can be exchanged by integrating such sequences) we built it into an integration vector.

Sequencing results of the lvl1 parts for modularized genome integrations

We successfully build two integration cassettes from our rationally designed artificial neutral integration sites (a.N.S.o. 1 and 2) and verified them by sequencing. These parts contained the “origin of transfer” to test their lethality in the aforementioned experiment.

Workflow to integrate a modularized integration cassette

We established a workflow on how to integrate a cassette - from lvl0 Parts to a finished change in genome. With UTEX 2973 this is possible in less than five days, while in PCC7942 the same integration would take a whole month.

Using the placeholder to build standard measurement vectors

We successfully used our placeholders to build and validate the standardized measurement vectors for promoters, ribosomal binding sites and coding sequences. We evaluated the cost and time savings from a library assembly with a sample size of 25.
Through our design decision to build placeholders we managed to cut the workload for a high throughput assembly by around 72% and the invested financial resources by 40 % with just a sample size of 25 assemblies.

Construction of a promoter library with standard measuring vectors

We built a promoter library using our standard promoter measurement vector and 25 BioBricks. Here we show a list of all the BioBricks we used.

Workflow for the screen of a BioBrick library

We designed a workflow to build a library, introduce it into UTEX2973 and measure its characteristics.

Testing the reproducibility and standard deviation of the screening workflow

We tested how reproducible results from our library screening workflow are with a fluorescence reporter.

Application note for the characterization of BioBricks in our chassis

After calibrating our screening procedure, we decided to share our practical knowledge with other end users.
Application Note