The Bacterial Etch-a-Sketch

Abstract The Etch-a-Sketch project aims to create a lawn of bacteria that can be drawn on with a laser pointer. This seemingly inconsequential task actually presents many interesting engineering challenges. In particular, the bacteria need to be sensitive enough to respond to a short light pulse from a laser, but still “selective” to not respond to ambient lighting. We have designed a novel genetic switch to tackle these problems. If our work proves successful, it will serve as a useful model for future projects that require large signal amplification. For example, researchers creating biosensors may find our work very helpful.

General Overview
Our project can be broken down into three parts: light input, signal amplification, and color output. We used LovTAP to receive the light input; we designed a switch based on Peking 2007’s Bi-stable switch to amplify the signal; and, finally, (after dabbling with some of Cambridge 2009’s E. Chromi colors) we decided to use mRFP1 as a color output.

LovTAP
Introduction In order to draw on a bacterial lawn with a laser pointer, we need bacteria to be able to respond to light. Previous iGEM teams have worked on many different ways to do this each with different advantages and disadvantages. We were struck by the simplicity and beauty of LovTAP. In particular, it functions by itself as a single protein; it is a brilliant example of genetic engineering; and, finally, it does not require any exotic supplements or specific strains of bacteria to function. LovTAP Parts As mentioned previously, LovTAP is a fusion protein. It consists of a light-response domain and a DNA binding domain, each of which are parts of other natural proteins.		Light Response The light-response domain is LOV2, the photoactive domain (i.e. the light responsive part) of AsLOV2 (Avena sativa phototropin 1). AsLOV2 is a protein which allows Avena sativa to respond to 470 nm light. It does this by undergoing a major conformational change upon being struck by a photon with a wavelength near 470 nm. The absorption of the photon leads to the formation of a covalent bond between a flavin mononucleotide (FMN) cofactor and a conserved cysteine residue. This new bond distorts the conformation of the protein, causing the detachment and unfolding of the Ja-helix (see figure 1). In natural AsLOV2, the unfolding of the Ja-helix results in further downstream signalling. However, we will be most interested in the fact that the Ja-helix detaches when LOV2 is hit by blue light.   Figure 1.   DNA Binding The DNA binding domain of LovTAP is the well-known bacterial transcription factor trpR. In the presence of tryptophan, the trpR protein will repress transcription of the E. coli trp operon by binding the operator region in the trp promoter and, thus, blocking RNA polymerase. Figure 2

LovTAP

Overview The two seemingly unrelated parts described above share one crucial feature: an alpha helix. LOV2 binds and unbinds an alpha-helix in response to light, and one domain in the functional trpR structure is “bound” to an alpha-helix. Strickland et al.’s idea was to force them to “fight over” a single alpha helix. Thus, since LOV2 has a higher affinity for the helix in the dark than trpR, there is no trpR activity in the dark. However, when exposed to light, LOV2 releases the alpha helix, allowing trpR to bind it and, thus, result in trpR activity (which of course is repression of the trp promoter).   Figure 3   Details Thirteen successive amino terminal truncations of trpR were ligated, in frame, downstream of the region coding for the Ja-helix in AsLOV2. One construct, referred to as LovTAP (the LOV- and tryptophan-activated protein), showed increased binding affinity to trp operator DNA when illuminated. Further tests showed that, upon light exposure, LovTAP binds DNA in a manner that is characteristic of the trpR domain. Additionally, mutation of the conserved cysteine of the LOV2 domain, which should prevent the conformational change that releases the Ja-helix, was shown to abolish binding to DNA in the presence of light. As this cysteine is crucial in the function of LOV2, this result suggests that the observed light sensitivity of the DNA-binding activity is due to LOV2. Taken from Strickland et al. A. Dark state-DNA dissociated: The shared helix contacts the LOV domain and trpR is in an inactive conformation. B. Photoexcitation occurs at 470nm (represented by the FMN chromophore going from yellow to white) causing the contact between the LOV domain and the shared helix to be disrupted. trpR now binds the helix and assumes its active conformation (represented by the helix going from blue to red). C. The active trpR domain of LovTAP binds the trp operator D. Cessation of light causes the LOV domain to return to its dark state, dissociating LovTAP from the DNA and reinstating contact between the LOV domain and shared helix. Figure 4.

Where iGEM and LOVTAP Meet
EPFL LovTAP was initially cloned into a BioBrick and characterized by the EPF-Lausanne 2009 team. Strickland et al. measured only a 5-fold change in DNA-binding affinity from dark state to light state in LovTAP. Though this change is relatively small (in biological scales), the EPF-Lausanne team found that there is a significant change in transcriptional output in response to 470nm light. Using molecular dynamics simulations, the EPF-Lausanne team found two mutations that were predicted to improve LovTAP’s function by increasing the stability of the light state and decreasing the time it takes to flip from dark state to light state. UNAM A new LovTAP part was synthesized by the UNAM 2010 team via the alteration of EPF-Lausanne’s LovTAP. Primary changes include the removal of two 2 PstI restriction sites from the coding region of LovTAP and the addition of one of the point mutations that was proposed by the EPF-Lausanne team.

Etch-a-Sketch's LOVTAP
Our Project For our circuit we will be using UNAM’s modified LovTAP, K360127. For the trp promoter and operator, we will be using our own ptrp, based on ptrpL K360023. Possible issues The UNAM team appeared to not have nearly as much success with LovTAP as the EPF-Lausanne team. Whether this was due to the mutation that was introduced or some other factor is unclear. We hoped to study UNAM’s LovTAP and possibly introduce other mutations to improve LovTAP’s function. For example, Strickland et al. identified mutations that increased LovTAP’s dynamic range (i.e. the difference in repression in the dark state vs. the light state) from 5 to 70. Another issue is that the conformational change of LovTAP in the presence of light is relatively unstable. Upon removal of the light source, LovTAP quickly returns to a conformation in which the Jα-helix binds to the AsLOV2 domain, inactivating the trpR DNA-binding activity. Thus, long exposure times might be necessary for the circuit to react to the light, which would be undesirable for drawing with a laser pointer. We may have to search for mutations that stabilize the light state, and thus allow LovTAP to remain in the light state for a longer time after being activated. Figure 5. LovTAP Switch in action.

Peking Bi-stable Switch
Overview The Peking 2007 iGEM team designed a genetic switch that can be flipped between two stable states (on and off), a so-called bi-stable switch. Such a switch would be desirable in the Etch-a-Sketch circuit because it would allow the bacteria to “remember” if it had been exposed to light, so that only a short exposure to light would be necessary to draw on our bacteria (rather than keeping the light on the bacteria until color appeared). However, the switch design was not ideal because one state was not completely stable; once the switch was flipped on, it slowly decayed back to the off state. We would like our drawings to be permanent (to some degree), so we would need the activated state to be stable. We therefore redesigned the Peking bi-stable switch for this purpose. Original Design The Peking 2007 switch uses the pR and pRM promoters and the cI and cI434 transcriptional regulatory proteins. The cI protein activates pRM and represses pR; cI434 represses pRM; pR has high basal transcription; and pRM has low basal transcription. Deactivated State Figure 6.   In the first state, cI434 levels are high and cI levels are low. This leads to high transcription from pR and low transcription from pRM, which, in this particular circuit, results in GFP production. The switch can be flipped by increasing cI levels.   Activated State Figure 7.   In the second state, cI levels are high and cI434 levels are low. This leads to high transcription from pR and low transcription from pRM, and therefore RFP is produced. Empirical evidence shows the switch will spontaneously decay back to the deactivated state. This is probably due to the high basal transcription levels of pR.

Our Design: Locking Switch
For our switch, we desired a simple circuit with a stable on state. The parts we used are ptrp, cI434, pRM, trpR, and cI. As in the Peking switch, cI activates pRM and cI434 represses pRM. We cloned two additional parts for our switch: ptrp and trpR. ptrp is the well-known trp promoter and operator sequence. trpR is the corresponding repressor which is induced by tryptophan. In our system, we will assume there are constantly high levels of tryptophan so that trpR will always act as a repressor. Deactivated State Figure 8.   A new LovTAP part was synthesized by the UNAM 2010 team via the alteration of EPF-Lausanne’s LovTAP. Primary changes include the removal of two 2 PstI restriction sites from the coding region of LovTAP and the addition of one of the point mutations that was proposed by the EPF-Lausanne team.   Activated State Figure 9.   In the active state, cI434 levels are low, trpR levels are high, and cI levels are high. This results in low transcription from ptrp, high transcription from pRM, and thus high output signal. Since trpR is a strong repressor for ptrp, we expect state 2 to be stable.

mRFP1
Color production For our output signal, we decided to use mRFP1. mRFP1 is a red fluorescent protein based on DsRed. Natural DsRed requires tetramerization in order to fluoresce. mRFP1 contains 33 mutations that, one, allow it to function as a monomer and, two, allow it to fold more quickly. It has many attractive qualities: the color is visible in normal lighting; no special supplements are necessary for the color to develop; the coding sequence for the protein is short (as compared to other pigment generating proteins); finally, the protein itself generates the color, which hopefully will result in faster image development than a protein which catalyzes a reaction that generates color.

Engineering Problems
Light, Sensitivity, Selectivity, Speed, and Noise We have identified five major engineering problems with Etch-a-Sketch that we have called: Light, Sensitivity, Selectivity, Speed, and Noise. Light - In order to draw on bacteria with a laser pointer, we need the bacteria to respond to light. Sensitivity - In order to draw naturally, we need the bacteria to respond to extremely quick pulses of light; otherwise, we would have to make very slow strokes. Selectivity - In order to use the bacteria in ambient lighting, we need to the bacteria to be able to selectively respond to just the laser pointer. Speed - After we make our strokes with the laser pointer, we would like to see color quickly. Noise - We do not want random spots to appear on our canvas. We have prioritized Light and Sensitivity. This hopefully would result in something that works like an old camera; instant image capture but overnight image development.

The Bacterial Etch-a-Sketch Circuit

Overview We have broken down the circuit into four states: Deactivated, Activation 1, Activation 2, and Activated. To recap the parts of our circuit: LovTAP represses ptrp in the light; trpR represses ptrp; cI434 repressed pRM; cI activates pRM; T7 polymerase (T7 P) transcribes at the T7 promoter; and mRFP1 generates a red color. Deactivated Figure 10.   In the dark, LovTAP is inactive; transcription levels from ptrp should be high, creating a lot of cI434 which should shut down transcription from pRM.   Activation 1 Figure 11. After light is applied, LovTAP is activated, repressing transcription from ptrp. cI434 levels should begin to fall.   Activation 2 Figure 12. Once cI434 levels have dropped enough so that transcription from pRM can begin, the circuit enters an irreversible positive feedback loop. trpR levels will begin to increase, preventing the creation of cI434. cI levels will increase, further increasing transcription from pRM. Finally, T7 polymerase levels will increase, allowing the creation of mRFP1.   Activated Figure 13. After entering the Activation 2 state, the light may be removed, and the cell will continue to produce mRFP1.

Mapping the circuit back to the engineering problems
Light, Sensitivity, Selectivity, Speed, and Noise Light - Our light response depends entirely on LovTAP. If it proves ineffectual, we will either need to introduce mutations that will help its function, or switch to a different system. Sensitivity - The amount of time necessary to keep the laser pointer on the bacteria should be related to the time it takes for the locking switch to flip to the activated state and the length of time LovTAP remains in its activated state once illuminated. The former should be small by design of the switch. The latter may be improved by mutations or changing to a different light response system, if necessary. Selectivity - The circuit should respond to only to wavelengths and intensities of light that can flip LovTAP from the dark state to the light state. Since LovTAP responds primarily to blue light, we may need to work in red ambient lighting. Speed - The delay from the moment the locking switch is activated to the appearance of color will be related to the time it takes for the bacteria to synthesize mRFP1. This will hopefully be minimized as transcription from T7 promoters is very high when T7 polymerase is present. Noise - Unwanted splotches will be created if there is unwanted transcription from the T7 promoter or if the locking switch is accidentally turned on by unwanted transcription from the pRM promoter. The former should be minimal as the T7 promoter is very specific to T7 polymerase. The latter should be reduced by placing the locking switch on a low-copy plasmid.

Results
Overview Unfortunately, we were unable to finish constructing our circuit in time for the wiki freeze.