Team:Montpellier/Design

Karma

DESIGN

1. The design of our KARMA tool for the proof of concept

In order to ensure the production of a functional KARMA for our proof of concept, we imagined and designed many possible KARMA variants. We selected 3 different TEV proteases to test: two mutants with reduced catalytic activity (TEVPE10 and TEVPH21) and a wild type TEV with full activity (TEVwt). We also chose 4 different linkers to test: each containing from 3 to 6 repetitions of the GGGGS amino acid sequence (linkers nx: 3x, 4x, 5x and 6x). Finally, we designed two directions for protein translation: one with an N terminal TEV protease and a C terminal VHH (Direction 1), and another with a C terminal TEV protease and an N terminal VHH (Direction 2).

We would therefore have 3 x 4 x 2 (=24) different KARMA, with varying combinations of the components and different expression directions. These 24 constructs would constitute a library that we could test by a functional screening, measuring their relative activity by the fluorescence of the GFP reporter. We also planned to do a western blot in order to verify and measure protein expression, thus we added a His tag at the end of the VHH sequence.

In this general KARMA design, the MBP protein is expressed in fusion with TEV, but separated by a TEVcs so that after translation, a TEV cleavage separates MBP from KARMA.

Figure 1 : KARMA constructs in two different directions of expression.

A. First direction of expression. The protein synthesis starts by the MBP, then the TEV, and finally the VHH. A TEV cutting site (TEVcs) separates the MBP from KARMA, so after translation, this site is cleaved. KARMA alone consists of a TEV in N terminal and a VHH in C terminal.

B. Second direction of expression. The protein synthesis starts by the VHH (in N terminal), TEV (in C terminal). The MBP protein fused to KARMA is separated after translation and TEV cleavage.

The production and comparison of all the designed variants would allow us to optimise the production of KARMA. However, for the very limited time we worked in the lab, we were only able to produce and test some constructs, each of them is mentioned in the tables below.

2. The assembly of our constructs

In order to assemble the parts (TEV, VHH, sfGFP, mRFP1, etc…) in the right order and to put them in the right vector, we used Golden Gate assembly and Gibson Assembly.

Expression inducers

Since we need to acquire fluorescence gain data in a controlled manner, we wanted to control the expression of all our constructs. Constitutive expression of the tool and/or the reporter might saturate the fluorescence data acquisition, we, therefore, put the expression of all our constructs under the control of well-known induction systems.

We used the Tet-ON expression system for expressing KARMA and the TEV protease alone. For our fluorescent reporters, we chose an araBAD promoter. The Tet-ON system inducible by atc is known for having very little leakage, and so does the araBAD system inducible by arabinose.

Vectors for expression

We have selected 3 vectors to insert our constructs: pOUT18 for the expression of KARMA and TEV (under atc induction), pBbE8K for the expression of GFP reporters (under arabinose induction) and pBbBK8 for the expression of the RFP reporters (under arabinose induction).

These plasmids were not chosen at random, first of all, we don't want to overexpress our proteins because we were afraid that it would be toxic for the cells, so we had to think about the origin of replication with a relatively low copy number per cell. We have referenced ourselves by this table.
Then it was necessary to take care to choose plasmids compatible in co-transformation according to their replication origin. We have chosen to use Kanamycin and Chloramphenicol resistance for a better selection of our processed strains.

The cloning was performed by replacing the fluorescence genes of these vectors by our designed constructs. Therefore, after successful cloning, the constructs were controlled by the promoters and terminators that were already present in the receiving vector.

Figure 2: Representations of the plasmids used to clone our designed constructs.

A. The pBbE8k plasmid contains the araBAD promoter, inducible by arabinose. As it contains the mRFP1 gene, the replacement by sfGFP reporters can be easily detected by a change of fluorescence from red to green in presence of arabinose.

B. The pBbB8k plasmid also contains the araBAD promoter, inducible by arabinose. It contains GFPuv gene, which makes the replacement by the mRFP1 reporters easy to detect by a change of fluorescence from green to red in presence of arabinose.

C. The pOUT18 plasmid contains the pTET promoter, inducible by atc. It contains the mRFP1 gene, which is replaced by our KARMA and TEV alone constructs.
Source Addgene : Here and Here

Experimental design

We first cloned all our designed constructs into the chosen vector, we did this either by Golden Gate or Gibson Assembly. Then, to perform all our tests, we co-transformed E. coli NEB10B by 2 constructs at a time : one construct producing a reporter and one construct producing the KARMA tool or TEV alone control.

Fluorescent reporters assembly :

Figure 3. Assembled reporters for GFP expression.

A. This construct produces our reporter sfGFP-TEVcs-ssrA under the control of a BAD promoter. Its is greatly improved thanks to the use of the RiboJ insulator and the Bicistronic Design 2 (BCD2) RBS placed upstream of the codon initiating the transcription. The sfGFP is continuously degraded as long as the proteolysis tag is attached to.

B. This construct produces an sfGFP-TEVcs-ssr. The reporter is under the control of a BAD promoter, sfGFP fluorescence is inhibited by the shadowG-induced quenching phenomenon.

C. This construct produces an sfGFP-TEVcs that simulates a cleavage of the first two reporters (see A and B) by the TEV protease. Even after cleavage of the ENLYFQ/S, a peptide remains molten with the sfGFP that can modify its fluorescence. This construct is as a control of the maximum fluorescence expected by the sfGFP reporters after TEV cleavage.

Figure 4. Assembled reporters for RFP expression.

A. This construct produces an mRFP1-TEVcs-ssrA type reporter. We used this reporter to check the specificity of our tool.

B. This construct serves as a basal control of the reporter's fluorescence maximum, as it simulates a cut by the TEV protease. Even after TEV cleavage, a peptide from the consensus sequence ENLYFQ/S remains attached to the mRFP1, and that can modify its fluorescence.

We did not design a mRFP1-TEVcs-ShadowG because ShadowG only quenches green fluorescence. Therefore specificity can only be tested for the proof of concept strategy that concerns ssrA removal. The quenching strategy only allows us to test the cleavage improvement for sfGFP.

The genes used for the fluorescent reporters

The GFP : For the reporter constructs containing GFP, we have chosen to use a variant of GFP which is the super-folder GFP (sfGFP) [1] [2], a robust variant of GFP that folds better than GFP when expressed as fusions with other proteins or poorly folded polypeptides.

The RFP : We chose the mRFP1 to show that our tool was specific. As the VHH is specific to GFP, we expected to see no difference in enzymatic activity between the KARMA tool and the control (TEV protease only).

Quencher : We chose ShadowG [3] as a quencher because it is very stable when fusionned to other proteins like GFP, it was characterized as a better quencher than REACH.

Unfortunately, the protein product of the construct with the quencher did not behave as we expected, the quenching did not take place. Nevertheless, we were very keen on making this new gene reporter because a similar construction existed in the registry (designed and characterized by the Oxford team in 2017), they also used a quencher but it was a REACH quencher. We could have reused the work of the Oxford team, but we wanted to bring a variant of this reporter in order to enrich the registry. Moreover, according to the authors who described shadowG, it would be a better quencher than REACH. > HERE <

The assembly of KARMA and TEV controls:

The addition of MBP in our construct was highly beneficial and not only by improving the expression. While browsing the literature on the intrinsic property of MBP we saw authors who produced, a TEV cleavage site in the MBP to get rid of MBP once folded. This was very interesting for our tool because in fact the TEV is capable to self cleave [4] and get rid of the fused MBP.

A. This construction expresses our KARMA tool (MBP-TEVcs-TEV-Linker 6x-VHH α-sfGFP), it is the TEV protease fused to a VHH specific to the sfGFP.

B. This construction expresses our KARMA tool with a variant of the TEV (MBP-TEVcs-TEVcs-TEVPE10-Linker 6x-VHH α-sfGFP), it is the TEVPE10 protease from which its Kcat and Km were reduced. it is also fused to a specific VHH from the sfGFP.

C. This construction expresses our KARMA tool with a variant of the TEV (MBP-TEVcs-TEVPH21-Linker 6x-VHH α-sfGFP), it is the TEVPH21 protease from which its Kcat and Km are also reduced. it is also fused to a specific VHHV from the sfGFP.

Figure 6 : Assembled TEV alone control variants

A. This construction expresses our native control (MBP-TEVcs-TEV), which is the TEV protease that is expressed in fusion with MBP for better expression.

B. This construction expresses our mutant native control (MBP-TEVcs-TEVPE10), it is the protease TEVPE10 with a kcat and km lower than a wt TEV . It is expressed in fusion with MBP for better expression.

C. This construction expresses our mutant native control (MBP-TEVcs-TEVPH21), it is the protease TEVPH21 with a kcat and km lower than a wt TEV. It is expressed in fusion with MBP for better expression.

3. The methods used for assembly

Golden Gate assembly

First, we chose Golden Gate to assemble all our biobricks. This technique is particularly adapted to the construction of our library as it allows us to establish the order of the components in a very precise way. But more importantly, Golden Gate allows fast production of many random variants in the right order.
Unfortunately the assembly by Golden Gate was much more difficult than we thought and we encountered several problems to obtain well-cloned bacteria. Still, we were able to produce many of the desired variants with this technique, we then sequenced them for verification, but the results were not satisfying. The majority of our cloned constructions had point mutations, deletions, and insertions. Solving these problems was difficult and took us a lot of time; we no longer had the time to make the functional screen as originally planned.
Therefore, the constructs variants produced by Golden Gate were not characterised.

KARMA variants	Assembled parts	Designed	Cloned	Characterised
KARMA-TEV Direction 1	MBP-TEV-Linker6x-VHH	YES	YES	NO
	MBP-TEV-Linker5x-VHH	YES	YES	NO
	MBP-TEV-Linker4x-VHH	YES	YES	NO
	MBP-TEV-Linker3x-VHH	YES	YES	NO
KARMA-TEV Direction 2	VHH-Linker6x-TEV-MBP	YES	YES	NO
	VHH-Linker5x-TEV-MBP	YES	YES	NO
	VHH-Linker4x-TEV-MBP	YES	YES	NO
	VHH-Linker3x-TEV-MBP	YES	YES	NO
KARMA-TEVPE10 Direction 1	MBP-TEVPE10-Linker6x-VHH	YES	YES	NO
	MBP-TEVPE10-Linker5x-VHH	YES	YES	NO
	MBP-TEVPE10-Linker4x-VHH	YES	YES	NO
	MBP-TEVPE10-Linker3x-VHH	YES	YES	NO
KARMA-TEVPE10 Direction 2	VHH-Linker6x-TEVPE10-MBP	YES	YES	NO
	VHH-Linker5x-TEVPE10-MBP	YES	YES	NO
	VHH-Linker4x-TEVPE10-MBP	YES	YES	NO
	VHH-Linker3x-TEVPE10-MBP	YES	YES	NO
KARMA-TEVPE21 Direction 1	MBP-TEVPH21-Linker6x-VHH	YES	YES	NO
	MBP-TEVPH21-Linker5x-VHH	YES	YES	NO
	MBP-TEVPH21-Linker4x-VHH	YES	YES	NO
	MBP-TEVPH21-Linker3x-VHH	YES	YES	NO
KARMA-TEVPE21 Direction 2	VHH-Linker6x-TEVPH21-MBP	YES	YES	NO
	VHH-Linker5x-TEVPH21-MBP	YES	YES	NO
	VHH-Linker4x-TEVPH21-MBP	YES	YES	NO
	VHH-Linker3x-TEVPH21-MBP	YES	YES	NO
Control constructs	MBP-TEV	YES	NO	NO
	MBP-TEVPE10	YES	NO	NO
	MBP-TEVPH21	YES	NO	NO

Figure 7 : Summary table of all the KARMA variants and controls designed for Golden Gate assembly. All KARMA variants were designed and cloned in E.coli, but none of them were characterised. None of the control constructs (TEV alone) was cloned by Golden Gate, nor characterised.

Reporter constructs	Assembled parts	Designed	Cloned	Characterised
Constitutive Reporters	sfGFP-TEVcs-SSRA	NO	YES	YES
Constitutive Reporters	sfGFP-TEVcs	NO	YES	YES
Inducible Reporters	Pbad-sfGFP-TEVcs-SSRA	YES	NO	NO
	Pbad-sfGFP-TEVcs	YES	NO	NO
	Pbad-mRFP1-TEVcs-SSRA	YES	NO	NO
	Pbad-mRFP1-TEVcs	YES	NO	NO
	Pbad-sfGFP-TEVcs-ShadowG	YES	NO	NO

Figure 8 : Summary table of all the reporter constructs designed for Golden Gate assembly. We were not able to produce nor test any of the reporter constructs we designed for Golden Gate assembly. The constitutive reporters sfGFP-TEVcs-SSRA and sfGFP-TEVcs were obtained from Hung-Ju Chang previous work.

Gibson assembly

Knowing that we had little time to assemble and produce our constructs, we started a plan B in parallel to Golden Gate assembly troubleshooting. We designed primers for Gibson assembly. This technique was simpler to perform but the design only allowed us to assemble the constructs one by one. We then prioritized the production of some constructs rather than others.

We made a bet by choosing to assemble everything only in the first direction of protein expression. We assembled KARMA variants of all the linker sizes but we characterised only the ones with 6 GGGGS repetitions. We chose to test the largest linker size in order to avoid the steric hindrance between the VHH and the TEV protease.

We were also able to produce and compare the KARMA variants with 3 different TEV and we made sure to produce and characterise all the reporters needed to perform the complete proof of concept.

KARMA variants	Assembled parts	Designed	Cloned	Characterised
KARMA-TEV	MBP-TEV-Linker6x-VHH	YES	YES	YES
	MBP-TEV-Linker5x-VHH	YES	YES	NO
	MBP-TEV-Linker4x-VHH	YES	YES	NO
	MBP-TEV-Linker3x-VHH	YES	YES	NO
KARMA-TEVPE10	MBP-TEVPE10-Linker6x-VHH	YES	YES	YES
	MBP-TEVPE10-Linker5x-VHH	YES	YES	NO
	MBP-TEVPE10-Linker4x-VHH	YES	YES	NO
	MBP-TEVPE10-Linker3x-VHH	YES	YES	NO
KARMA-TEVPE21	MBP-TEVPH21-Linker6x-VHH	YES	YES	YES
	MBP-TEVPH21-Linker5x-VHH	YES	YES	NO
	MBP-TEVPH21-Linker4x-VHH	YES	YES	NO
	MBP-TEVPH21-Linker3x-VHH	YES	YES	NO
Control constructs	MBP-TEV	YES	YES	YES
	MBP-TEVPE10	YES	YES	YES
	MBP-TEVPH21	YES	YES	YES

Figure 9 : Summary table of all the KARMA variants and controls designed for Gibson assembly. All KARMA variants in direction 1 were designed and cloned in E.coli, but only the ones with the 6x linker were characterised.

Reporter constructs	Assembled parts	Designed	Cloned	Characterised
Inducible Reporters	Pbad-sfGFP-TEVcs-SSRA	YES	YES	YES
	Pbad-sfGFP-TEVcs	YES	YES	YES
	Pbad-mRFP1-TEVcs-SSRA	YES	YES	YES
	Pbad-mRFP1-TEVcs	YES	YES	YES
	Pbad-sfGFP-TEVcs-ShadowG	YES	YES	YES

Figure 10 : Summary table of all the reporter constructs designed for Gibson assembly. We were able to produce and characterize all the reporter constructs we designed for Gibson assembly. However, during the characterization of fluorescence, the quencher of the first strategy did not actually quench and we did not have enough time to improve the corresponding reporter construct. We thus focused on the second strategy with the ssrA tag.

[1] Raran-Kurussi, Sreejith, et David S. Waugh. 2012. « The Ability to Enhance the Solubility of Its Fusion Partners Is an Intrinsic Property of Maltose-Binding Protein but Their Folding Is Either Spontaneous or Chaperone-Mediated » éd. Bostjan Kobe. PLoS ONE 7(11): e49589.

[2] Overkamp, W. et al. (2013) Benchmarking various green fluorescent protein variants in Bacillus subtilis, Streptococcus pneumoniae, and Lactococcus lactis for live cell imaging. Appl. Environ. Microbiol. 79: 6481-6490 https://www.nature.com/articles/nbt1172

[3] Murakoshi, Hideji, Akihiro C. E. Shibata, Yoshihisa Nakahata, et Junichi Nabekura. 2015. « A Dark Green Fluorescent Protein as an Acceptor for Measurement of Förster Resonance Energy Transfer ». Scientific Reports 5(1): 15334.

[4] Shih, Y.-P. 2005. « Self-Cleavage of Fusion Protein in Vivo Using TEV Protease to Yield Native Protein ». Protein Science 14(4): 936‑41.