Orthogonal Pair

The recognition of the amber stop codon requires a tRNA with an anticodon complementary to the amber stop codon and an aaRS specifically loading the tRNA with the nnAA. In order to ensure the nnAA is not incorporated for other codons except the amber stop codon, the tRNA and the aaRS have to be orthogonal to the natural aaRS's and tRNAs. This means the aaRS must not load any other tRNA and the tRNA must not be loaded by any other aaRS. Therefore, Wang et. al originally used the tyrosyl-tRNA and tyrosyl-RS from the methanogenic archaeon Methanocaldococcus jannaschii : The anticodon of the tRNA was replaced by the amber anticodon and the aaRS was optimized for the recognition of OMT in place of tyrosine via directed evolution. Introduced into Escherichia coli, this pair is orthogonal to every natural pair due to the genetic distance between E. coli and M. jannaschii. Nowadays, over 70 different aaRS [3] have been designed, each one capable of incorporating a specific amino acid, many of them with special chemical characteristics, allowing e.g. 'click' chemistry or photoactivation.

In our project, we use an orthogonal pair from the "Expanded Genetic Code Measurement Kit" as template, specifically the one used for incorporation of ONBY (BBa_SomeBrick), and replaced the ORF with an E. coli codon optimized ORF for OMT-RS. Furthermore we placed the OMT-RS coding region behind a RBS (BBa_B0034) and a strong constitutive Anderson promotor (BBa_J23101). A successful expression of the OMT-RS gene in this construct was observed (Fig. 1).

Usage of amber codon

The incorporation of an amber codon causes the complete translation of the respective protein in presence of the nnAA and cancels the translation in absence. In our implementation the amber codon is replacing a codon in the beginning of the ORFs of the Colicin E2 Immunity protein (Y8OMT) and the Zif23-GCN4 repressor (F4OMT). In consequence, both proteins are functionally produced only if the nnAA is available in sufficient concentration in the medium.

The non-natural amino acid

We decided to use O-methyl-l-tyrosine for our nnAA due to its multiple advantageous properties:

• Low costs
• Nontoxic
• Unproblematic import into cells
• No further biochemical activity
• Feasible chemical synthesis
• Stable in water
• Unavailable in nature
• Well documented
• Low interference with protein activity

An institute or company could choose its own specific nnAA with the corresponding orthogonal pair. This enables a reliable protection against corporate espionage or bioterrorism, since the opposing party does normally not know which nnAA is used in the respective application. However, using the same nnAA like OMT in every application should prevent the biological and genetic spread of the respective microorganism in the environment.

The Reporter System

For the detection of a low non-natural amino acid concentration, which in this case is O-methyl-L-tyrosine, we designed a reporter system that includes the reporter protein mVenus. In order to make sure that the expression of mVenus does only start at a low OMT concentration, we use a dimeric repressor. An amber mutation was introduced to the DNA sequence of the repressor. This amber mutation leads to OMT being integrated in the dimeric repressor protein. However, the repression of the mVenus promoter can only be executed if there is a sufficient amount of OMT in the medium. If the OMT concentration drops below a threshold, the expression of mVenus is induced. As a result, we can detect a yellow fluorescence signal.
The reason why we utilize a dimeric repressor was that this kind of repressor binds strongly to the respective promotor. Moreover, this dimeric repressor creates a sigmoidal repression curve (x‑axis = concentration of OMT; y‑axis = repressor molecule concentration). Once the concentration of OMT drops, we get a signal quickly.

To make sure that the repression does not take place even if the concentration of OMT is low, an LVA degradation tag is expressed with the dimeric repressor. To ensure that there is no permanent fluorescent signal caused by mVenus, it is marked with an LVA degradation tag as well. So, both proteins degrade quite fast after their translation. To connect this system to the expression of colicin, we can use different Anderson promoters for test purposes (BBa_J23104, BBa_J23113, BBa_J23107, BBa_J23100 and BBa_J23114). By doing so, we take care that the fluorescent signal of mVenus appears before the expression of the DNase that degrades the DNA and makes the genomic information inaccessible.

mVenus

The fluorescent reporter protein mVenus is a mutant of the green fluorescent protein GFP which is often used for fluorescence assays. Due to mutagenesis (F46L/F64L/M153T/V163A/S175G), the maturation time is decreased compared to GFP. In general, the maturation process can be divided in the folding step and formation of the chromophore. During the maturation process, the chromophore formation is the rate-limiting step. After the folding, a torsional rearrangement effects the formation of the chromophore. This results from the involved residues being in close proximity. After cyclization of two amino acids has taken place, oxidation is the final step. Molecular oxygen is necessary for the reaction that generates the delocalized pi electron system, resulting in the fluorophore being maturated and fluorescent. It is protected by the Beta-barrel protein from interfering influences. All the processes are influenced by the general cell- and cell-cycle processes and can be delayed or accelerated. In vitro, the maturation time of mVenus is in average 40 min (Lizuka et al., 2011). Another effect of the mutation F46L is the lowered sensitivity to the pH and chloride ion concentration which is one of the drawbacks of wild‑type GFP.

mVenus is expressed with a LVA degradation tag to decrease the protein half‑life. Moreover, the reporter is not regulated by any proteins, cofactors or substrates. The lack of disulfide bonds supports the choice of mVenus in our model microorganism E. coli. Its absorption maximum is at 512 nm and its emission maximum at 528 nm. The atomic mass is approximately 27 kDa.

The figure shows the mVenus reporter protein (without LVA degradation tag). The typical Beta-barrel fold is highlighted in yellow. The fluorophore is hidden inside the barrel structure. PDB ID 1MYW, created with Pymol.

Rational Design of the Amber Mutant of the Dimeric Zif23-GCN4 Repressor

The regulation of the reporter protein mVenus is carried out by a dimeric zinc finger protein. It binds cooperatively to DNA (a specific promoter region), connecting with the major groove of the DNA. The dimeric Cys2His2 zinc finger protein is the DNA binding domain and attached to a leucine zipper dimerization domain. Therefore, the targeted gene is controlled by the specific DNA binding. The monomers bind the DNA specifically and dimerization happens upon binding.

In order to control expression of the repressor on a translational level, an amber stop codon is introduced to the sequence of the repressor. First, the mutation site had to be determined. A position was chosen in which the non-natural amino acid should not interfere with the protein structure. A localization close to the N-terminus was selected as the protein expression will stop early once the non-natural amino acid concentration decreases. Phenylalanine was replaced by O-methyl-L-tyrosine (F4OMT) in order to retain stacking interactions. All nearby side chains as well as the helix (starting from R15) were considered and destabilizing mutations were avoided. Additionally, it is important to choose a residue that is not involved in DNA binding. Otherwise, the repressor may lose its function. The residue of the amber mutation is highlighted in yellow in the picture.

Overview of the amber mutation site in the repressor protein that binds DNA (shown in black). The phenylalanine residue is mutated to O-methyl-L-tyrosine (F4OMT). The residue is located close to the N-terminus of the repressor protein in order to interrupt protein expression early when the non-natural amino acid concentration decreases. Created with Pymol software, PDB ID 1LLM

[...]

Metabolic burden

.......

Genomic integration

The λ‑integrase, originally derived from the λ‑Phage, catalyzes in combination with several assisting proteins the excessive and integrative recombination of the phage's genome with the chromosomal genome of a host. For this, two attachment sites are needed: one located on the bacterial genome (attB) and the other located on the λ‑genome, which also contains several binding sites for regulatory proteins. The attachment sites contain homologous recognition sequences, called BOB' Region (attB) and COC' Region (attP). These can be connected by the λ‑integrase and the bacterial integration host factor (IHF) via Holliday junction forming an intasome, a DNA‑protein‑complex, producing hybrid attachment sites attL and attR.
For the integration of a gene of interest (GOI) into the chromosomal genome of E. coli there are two plasmids needed. One, called integration plasmid, contains the constitutively expressed GOI GFP with a LVA degradation tag, which, as previously mentioned, is also the reporter that is necessary for the measurement of the metabolic burden and should be integrated into the E. coli genome. It also contains the attP site that enables the integration. There are two bidirectional terminators located on each side of the attP to protect the GFP Operon from the transcription of the other neighbouring genes. The antibiotic resistance will also be integrated into the genome if the genomical integration succeeds, so we decided to use a Kanamycin resistance, as it is less commonly used in iGEM than Ampicillin or Chloramphenicol. Therefore, we chose the backbone pSB3K3, which also possesses a low copy ori and eases the later performed plasmid curing. The second plasmid is a helper plasmid and is necessary for transposing the GFP into the chromosomal genome as it contains the protein λ‑integrase with a ribosomal binding site (RBS).
To verify whether the recombination was successful one can perform a PCR with primers binding to the attB site of the E. coli and the VR Primer, which binds on every BioBrick compliant plasmid. As the one primer binds on the genome and the other on a plasmid, there can only be a PCR amplicon if the integration has succeeded.

Integration strains

A suitable genomic integration strain needs to carry the attB sequence needed for λ‑integrase mediated recombination, which can be troublesome because many commonly used E. Coli strains already have the λ‑phage integrated into its genome. Also, the attB site needed for the integration is blocked in λ (DE3) phage.
For our integration strain we chose the E. Coli JM109 strain because it matched all our demands and was also freely and easily available to us.

[...]