Team:Paris Bettencourt/Proteins Caging



Photoreceptors are valuable optogenetic tools which, upon coupling with other proteins, activate certain functions in a controlled spatial and temporal manner when exposed to the appropriate wavelength of light. However, the usage of photoreceptors suffers from many drawbacks including the toxicity of the light to the cells, photobleaching of the receptors and the delay in the response i.e. the time needed for transcription and translation of the target protein to be controlled. Also they always require cells, which is not good for safety. The emergence of Fluorescent light-inducible proteins is an attractive alternative that doesn’t suffer from these drawbacks.

Dronpa is one of the reversible photoswitchable fluorescent proteins (RSFPs), these are proteins that are switched on and off reversibly by specific wavelengths. Dronpa is switched on by default “fluorescent” and is switched off when illuminated by cyan light (~500nm). Dronpa Fluorescence is recovered by shining violet light (~400nm)

Figure 1: an illustration of the on/off switching of dronpa and the associated alternation between the monomer/dimer structures

The conformational changes that are associated with the on/off switching of Dronpa Lys145Asn have been used in a design that facilitates the optical control of protein activities. When Dronpa domains are fused to both termini of an enzyme of interest, the Dronpa domains form a tetramer and cage the enzyme leading to its inactivation. By Shining cyan light, Dronpa is switched off and the tetramer dissociates into monomers, as a result, the caged enzyme is activated (1) (3). However, traditional methods only turn the monomer enzymes off. Here we show there are much more possibilities of the design principles.

Figure 2:A fluorescent light-inducible protein design based on Dronpa Lys145Asn- From Zhou, X.X. and Lin, M.Z., 2013.


1) Design of protein Caging:

In our design, we had two copies of Dronpa Fluorescent Protein with two BsaI cutting site in between to allow the insertion of our genes of interest .

Figure 3: construct design for 2 Dronpa domains

2) Molecular mechanism:

The enzyme of interest is placed between 2 copies of Dronpa fluorescent protein via a linker. By switching on Dronpa (violet light) the protein of interest is caged as the two copies would dimerize. And by switching off Dronpa (cyan light), the two copies would dissociate which will result in activation of the protein of interest.

Figure 4: Shining cyan and violet light using LEDS.

3) MutDronpa:

Using error prone PCR we have obtained several versions of Dronpa with several mutations from which we picked one version that interested us the most, as it had an R149H mutation which is located in the dimer interface. We proceeded in our experiments with 2 variants of Dronpa. The original version was annotated as wtDronpa and the mutated version was annotated as mutDronpa which has 2 mutations I4V and R149H in the first dronpa domain and an F78S mutation in the second domain.

Strategy I: Caging increases -instead of blocks- the activity of caged proteins:

Repressors bind DNA and setback transcription. In our project, we developed a logic gate at the promoter level by creating dually repressed promoters using different combinations of the operators for TetR, P22 c2, and HKCI and it was interesting for us to test if these repressors can be controlled by light thus creating a light-inducible library of transcription factors. Dronpa hasn’t been used to control transcription factors, so we developed the following constructs to test if Dronpa can control proteins that function by dimerization.

Figure5: A diagram of the design of Dronpa-Repressors constructs.

By running a structure prediction test [figure 6A], we could obtain some information regarding the expected behavior of the output of our design. In our design, the two Dronpa domains are connected to the repressors via long flexible linkers, which indicates that homodimers of repressors are likely to be formed even if caged by dronpa. The folding of the repressors that we are testing will be fast due to their small size and simple structure, which will make the dimerization of 2 copies of the repressors -before the complete folding of the second dronpa domains and caging the repressor- very likely. We expected in our design that the repressor homodimer will bind to its operator and once the second domain of Dronpa is fully folded, it will cage the already bound repressor to the DNA thus further stabilizing the repression when the violet light is shone [figure 6B].

Figure 6A: Structure prediction of A) P22C2-dronpa fusion B)TetR-dronpa fusion C)HKCI-dronpa Fusion Figure 6A: Dimerizing Dronpa might be locking the proteins around their substrates

To validate our design, we tested the three constructs in [figure 5] for both wtDronpa and mutDronpa with the logic gate of the dually repressed promoters in a cell-free system giving us 96 combinations[figure 7 A].

Figure 7A : the experiment conducted with the repressors caged with Dronpa Figure 7B : An over view of the results of the cell-free experiment. Each promoter was tested with its cognate repressors. Our results show that in 90% of our constructs caging the repressors with Dronpa has increased the repression strength. We have also obtained a wide linear range of repression strength-indicated by the red slope-. Figure 7C: detailed Results of the cell-free experiment. Each promoter was tested with its cognate repressors. Top: Testing with TetR caged with either wt-Dronpa (BBa_K2510108) or a mutated version(BBa_K2510109)Middle: Testing with P22c2 caged with either wt-Dronpa (BBa_K2510112) or a mutated version(BBa_K2510113).Bottom: Testing with HKcI caged with either wt-Dronpa (BBa_K2510110) or a mutated version(BBa_K2510111) Figure 7D: the wtDronpa has a better control of HKC1 and C2P22 repressors while the mutDronpa has a better control of TetR repressors

By illuminating our system with violet light, caging increase the repression strength while uncaging with cyan light decreases the repression strength. Our results also show that the mutDronpa can better control the activity of TetR than the wtDronpa. However, HKC1 and C2P22 are better controlled by the wtDronpa [Figure 7].

Strategy II: Caging controls the protein activity by causing conformational changes:

T7 RNA polymerase (RNAP) transcribes only DNA downstream of a T7 promoter and it has a low error rate. Putting into account that T7 RNAP flexibility as it undergoes dramatic conformational changes during the transition from an initiation complex to an elongation complex. We hypothesized that adding two copies of Dronpa might control the activity of T7 RNAP by altering its conformation. To test this hypothesis we made the following construct:

Figure8: A diagram of the design of Dronpa-T7RNAP constructs.

By running a structure prediction test [figure 9], the prediction indicates the that the addition of the two dronpa domains is causing drastic changes in the structure of T7 RNAP. In the original T7 RNAP structure the C and N termini are adjacent, so fusing the 2 dronpa domains to both termini is changing its conformation.

Figure 9: A) The structure of T7 RNAP binding DNA (PDB: 1t7p) B)The prediction of the structure of Dronpa caging of T7 RNAP, in red circles the C and N termini of the protein.

The inactivation of T7RNAP was confirmed by using a reporter mrfp construct [figure 10] to indicate the activity of T7 RNAP as fluorescence output. By plating the cells overnight in dark and cyan conditions, no RFP was generated.

Strategy III: Oligomerization of Dronpa competes with the assembly of other enzymes

Many interesting proteins function as tetramers, for example, glutathione S-transferase, beta-glucuronidase, magnesium ion transporters such as CorA and export factors such as SecB from E Coli. Developing a strategy to optically control tetrameric proteins has remained challenging, due to their huge structure. In our project, we found it would be of great interest to develop a principle for controlling the activity of these proteins by light. The interaction between subunits forming a tetramer is determined by their rate of association and dissociation. By switching on/off Dronpa it’s altered between multimer and monomer structures. We hypothesized that by fusing 2 domains of dronpa in each subunit of a tetramer, the oligomerization of Dronpa would interfere with the assembly of the 4 subunits, thus allow the control of the tetrameric proteins [fig 11]. Figure11: contrelling enzymes might be possible by disrupting their oligomerization To have a proof of concept we chose to work with β-galactosidase that functions as a homotetramer and its activity can be measured visually by X-Gal which produces a blue color that can easily be detected visually upon β-galactosidase activity. We made the following construct and we put it under testing.

Figure12: construct of lacz-Dronpa fusion to test for β-galactosidase activity.

BWe tested the protein activity in vitro by the experiment indicated in [figure 13]. Our preliminary results [figure 14] suggest that β-galactosidase activity can be controlled by light, although there exists a background that indicates a leaky control.

Figure 13: An overview of the experiment done to evaluate the activity of β-galactosidase-Dronpa fusion. Figure 14: Top: X-Gal grayscale picture, testing the activity of β-galactosidase fusion with both wtDronpa and mutDronpa, indicating that β-galactosidase-mutDronpa fusion is more responsive to cyan light than the β-galactosidase-wtDronp. Down:90 fold difference in the activity between the MutDronpa caged beta-gal open and closed state after 4 hours of incubation


The usage of fluorescent protein domains to control protein activity has many advantages over the commonly used transmembranes photoreceptors. As the heavy circuits of transmembrane proteins are replaced by a single construct in which the protein of interest is fused with the fluorescent protein domains to be controlled directly without intermediates and the delay in response resulting from transcription and translation processes is replaced by a fast response once the already existing proteins are activated by light. A major disadvantage of the transmembrane photoreceptors is their unsuitability to cell-free systems, a problem that photoswitchable proteins don’t face as they don’t require membranes or co-factors. In our project, we tried to expand the usage of Dronpa fluorescent protein domains by exploring the possibility of controlling proteins that haven’t been controlled before by hypothesizing different strategies for protein control and by obtaining an interesting mutant version of Dronpa that better suited the control of some proteins than the original Dronpa. Fluorescent protein domains are being proved as promising tools that have many possibilities to control different types of proteins that are yet to be explored. Utilizing other RSFPs that are on/off-switched at various wavelengths in building similar light-controllable system would expand the fluorescent inducible reversible lights toolbox, which in turn might lead to many new applications. Where the activity of multiple proteins can be controlled independently in the same cell. Or to construct logic gates where protein activity is determined by multiple inputs- wavelengths-.


Zhou, X.X., Chung, H.K., Lam, A.J. and Lin, M.Z., 2012. Optical control of protein activity by fluorescent protein domains. Science, 338(6108), pp.810-814.

Zhou, X.X. and Lin, M.Z., 2013. Photoswitchable fluorescent proteins: ten years of colorful chemistry and exciting applications. Current opinion in chemical biology, 17(4), pp.682-690.

Zhou, X.X., Fan, L.Z., Li, P., Shen, K. and Lin, M.Z., 2017. Optical control of cell signaling by single-chain photoswitchable kinases. Science, 355(6327), pp.836-842.

Centre for Research and Interdisciplinarity (CRI)
Faculty of Medicine Cochin Port-Royal, South wing, 2nd floor
Paris Descartes University
24, rue du Faubourg Saint Jacques
75014 Paris, France