UCSF iGEM 2005: Bacterial Thermometer
Our goal is to build a genetic circuit that enables a lawn of bacteria to respond to a temperature gradient. Inspiration came in the form of those early 1990's T-shirts with the thermosensitive dye that change color that can record palm prints.
1. Determine the transfer function of a set of heat and cold shock promoters, where the input is temperature and the output is a steady-state (!) expression of green fluorescent protein (gfp). In essence, this will provide new parts, in the form of promoters, that respond to different temperatures.
2. Use existing transcriptional inverters to process the signal from the T-sensitive promoters (obtained from goal 1). For example, if we discover a cold shock promoter with a suitable transfer function, this can be 'inverted' into a heat-shock promoter.
3. Build new genetic amplifiers to increase the signal from promoters with weak or adaptive responses. This is a particular problem with heat-shock promoters, which classically produce a pulse-like response when unfolded proteins are present. Even those that have been identified as reaching a T-dependent steady-state produce a very weak signal. To boost this signal, we will construct a series of new genetic amplifiers.
4. Link genetic devices that respond to high and low temperatures using an OR-gate. This will result in an output that is on at both extreme high and low temperatures. Further, this output can be inverted such that it responds in a narrow range of temperatures.
5. Determine if the T-dependent genetic device can function on plates to record an applied temperature gradient.
As it turns out, identifying suitable temperature-dependent promoters was a non-trivial task. This was particularly surprising because of the number of known heat and cold shock promoters that are available in the literature. In the end, over 40 promoters were tested. These promoters were identified using micrarray data and the known architecture of the bacterial networks that respond to heat and cold.
Each promoter was screened to determine its transfer function. The input of the transfer function is temperature and the output is the steady-state activity of the promoter. The promoters were put into a screening vector, which has a high-copy (pUC) origin and gfpmut3 as an output. It took some tweaking to obtain growth conditions that produced accurate, robust, and reproducible transfer functions. The following protocol is the one that we settled on. First, the strains are grown overnight in 5ml LB with the appropriate antibiotics. The overnights are diluted first 1:1000 into 5ml fresh LB. After reaching OD ~0.4, they are rediluted 1:100 and grown to OD 0.8. Then, aliquotes are placed at different temperature (no dilutions) and grown at different temperatuers for 4 hours. The samples are then assayed for fluorescence using flow cytometry. In practice, this protocol works because it reduces variation due to different growth rates at different temperatures.
Using this screen, the vast majority of promoters were eliminated. Eliminations occured for several reasons. First, the reporter did not show any change in activity as a function of temperature. Second, the activity of the promoter was significantly growth phase dependendent. This is particularly bad because of the variation of growth rates due to temperature. To make sure that the T-dependence was not due to changes in growth rate, we had to determine the OD-dependence of the reporter for each of the promoters. This was a lot of work! Third, the promoter produced some sort of toxic affect. This could usually be traced to a leader sequence on the mRNA. To try to bypass this problem, we also constructed promoters where this leader was removed and this had some success.
Only four promoters (barely) survived the screen. By far, the best promoter is hybB, which controlls the hydrogenase II operon. It is clearly active at temperatures lower than 30oC and is off at temperatures higher than 30oC. Two other cold-shock promoters also showed a T-dependent response: ansB and cspA_x. AnsB controls an asparginase operon. CspA is part of the classical cold shock response. It's mRNA has a toxic leader, which is also supposed to participate in adaptation. We removed this sequence, which we denote with an 'x.' Finally, we have had mixed success with the heat-shock htpG promoter, which is part of the classical heat-shock response. It is known to produce a pulse first, but it is unique in that it (in the literature) comes to a T-dependent steady-state.
We have connected the most promising cold-shock promoter (hybB) to various inverters, which we obtained from the MIT Registry of Standard Biological Parts. We connected it to both the lambda- and tetR- based inverters (parts BBa_ and BBa_ , respectively). Neither properly inverted the signal. Both produced a T-independent background fluorescence. We could not figure out if this background was 'ON' or 'OFF,' which is where the tetR-based inverter became useful. Using aTc to bypass this circuit, we determined that the phenotype was constituatively off. HybB is most activate at 30 and least active at 37. We could grow the inverter at these temperatures and determine the dose-response curve using aTc. Indeed, there is an intermediate region of aTc concentrations, where the circuit is ON at 37 and OFF at 30 (propertly inverted). This means that weakening TetR expression should produce an inverter that works. We have taken two approaches to this problem. First, we built a set of three additional inverters which have different ribosome binding sites controlling TetR translation. These all have been shown previously to weaken the ribosome binding site to different degrees (parts BBa_, BBa_, and BBa_). Second, we constructed a library where random mutagenesis was applied to the rbs.
In an attempt to boost the signal from heat-shock promoters, a series of four amplifiers was constructed. The promoter being amplified drives the expression of a transcriptional activator Spo0A from B. subtilis. Then, four different promoters that are activated by Spo0A are used that have different output strengths: spoIIE > sin-P1 > spoIIGE > spoIIG. The spread in outputs is approximately 10-fold difference between each promoter (10,000 fold overall). The spoIIGE promoter is a hybrid of the spoIIG and spoIIE promoters. Right now, each of the promoters drives the expression of gfp as an output. The SpoIIE varient is so strong that we use the fast degrading gfp with an LVA tag. To compare the amplifiers, the tetR promoter is placed in front of spo0A for each of the constructs.
We then tested the ability for the amplifiers to increase the signal of the cold-shock promoter hybB and the heat-shock promoter htpG. Without the amplifier, activity from htpG can barely be measured.
- New Parts
hybB promoter specs
ansB promoter specs
cspA_x promoter specs
htpG promoter specs
spo0A_CTD transcriptional activator
- New Devices
amplifier 1 (IIE)
amplifier 2 (P1)
amplifier 3 (IIG)
amplifier 4 (IIGE)
- Team Members
Liz Clarke (Voigt Lab)
Matt Eames (Kortemme Lab)
High School Students
Xiao Yan (Apple) Liu (SFUSD Mission High School)
Nessa Ramos (SFUSD Mission High School)