Team:ETH Zurich/Experiments/Heat Sensor

Heat Sensor Experiments

Introduction

We incorporated a module into our system that would allow our engineered bacteria to sense the external toxin release signal of the doctor. The signal is produced with focused ultrasound and increases the temperature in the tumor region to 45 °C. To detect this signal, we needed to fit a naturally occurring heat sensor from Salmonella into our novel genetic circuit. If the heat sensor detects the temperature increase, it should activate the next step (cell lysis), by promoting transcription of protein E.

Figure 1. The genetic mechanism of our TlpA heat sensor. TlpA dimers bind in the PTlpA region and repress transcription of the downstream gene. A temperature of 45 °C shifts the equilibrium of dimerization towards monomers, which don't bind in the PTlpA region. Transcription can therefore happen at 45 °C.

For more details about the mechanism, go to Heat Sensor.

Phase I: Initial System Design

The TlpA heat sensor consists of two parts: the TlpA coding sequence and the PTlpA promotor sequence. We copied the digital sequences from a plasmid used by Piraner et al. and introduced 8 silent mutations to remove forbidden restriction sites inside the coding sequence. We designed test-plasmids to ensure proper function of the system. The initial design of a test system constists of two plasmids, one containing the repressor and the other the promotor and a reporter protein. The first plasmid was designed to have an Anderson Promotor with relative strength of 0.71 that promotes transcription of an RBS (designed with Salis Lab RBS calculator) with a calculated translation initiation rate of 5000 and the TlpA coding sequence.

Figure 2. The heat sensor test plamids. The TlpA coding sequence is placed on the piG17-2-002 (pSEVA361) and the heat inducible gfp is placed on piG17-1-005 (pSEVA291). PF and SF are abbreviations for BioBrick Prefix and BioBrick Suffix restriction sites. RS1-RS4 are restriction sites that we introduced for later cloning.
The optimal amount of the TlpA repressor protein in the cytoplasm was not known to us from the beginning, that's why we decided for a medium amount of TlpA expression. The physical DNA sequences were ordered as gBlocks from IDT and inserted to our test plasmids pSEVA361 and pSEVA291 via Ligase Cycling Reactions.

KEY QUESTIONS TO ANSWER IN FIRST EXPERIMENTS

  • Does the heat sensor work in our Lab with our experimental setup?
  • Does the heat sensor work despite the changes we made to the coding sequence, the calculated ribosome binding site and the BioBrick promotor?
  • Which temperature is needed to activate the heat sensor?
  • How long does the heat sensor have to be induced for decent expression of the regulated gene?

Phase III a): Function test with reporter gene

A sequence of experiments was performed to find optimal induction times and experimental setup. E. Coli Top10 chemical competent cells or Nissle electrocompetent cells were used. Single colonies of double transformants were inoculated to 12 mL round bottom culture tubes in 5 mL LB and grown for 16 h at 37 °C shaking 230 rpm. After growth to stationary phase, they were diluted to OD 0.1 and grown to exponential phase (OD 0.4). At this point the induction procedure was initiated in different formats, for different times and temperatures. Findings were:

  • Induction times of 1 to 15 minutes don't induce the reporter gene strong enough, even though temperatures above 42 °C lead to slightly higher fluorescence after 15 min induction
  • Strong induction takes place in a timescale of 1-5 hours. (more would not be feasible for our application)
  • Growth in 12 mL culture tubes is best suited and the samples should be diluted before plate reader measurement.
Figure 3. Fluorescence readouts of the TlpA Heat sensor after induction times from 0.5 to 5 h. Biological triplicates were measured and the standard deviation is depicted. Longer induction times lead to higher fluorescence values. Even though induction is significant, the fold change is very low. Induction was performed in 60 uL cell culture in 100 uL PCR tubes in thermocyclers for the time and temperature indicated and after induction stored at 30 °C until start of the readout (5 h after start of induction).
The experimental setup was changed from induction in thermocycler to induction in shaking incubators. Shake flasks were used and the same experiment repeated.
Figure 4. Fluorescence readouts of the TlpA Heat sensor after 3h induction at 45 °C in shake flasks in a shaking incubator. Biological triplicates were measured and the standard deviation is depicted. The fluorescence of the supernatant was also measured to find out if we can use supernatant fluorescence in future cell lysis measurements.
These experiments showed that induction is possible, but the fold change of ~6 is very low. This is either due to a low maximum induction of the promotor compared to a constitutive promotor of gfp, or to a high expression when not induced. We hypothesized that the RBS of TlpA was not inducing translation enough, and low amounts of the TlpA reporessor protein don't inactivate base level expression enough. Double transformations with protein E under the TlpA operator suggested that we need a tighter regulation (they did not yield any transformant colonies due to insufficient inhibition of protein E expression leading to cell lysis). An RBS library was then created to find variants translating more TlpA RNA.
Figure 5. RBS Library

To read more about each of these experiments, click on the buttons below. For a detailed protocol describing each experiment, visit Protocols.

Quorum Sensing End-Point Characterization

OBJECTIVE
Determine the population density at which the quorum system gets activated and provide the modellers with data to infer aLuxI, the production rate of LuxI.

PROCEDURE
We transformed E. Coli with a regulator and an actuator plasmid, containing constitutive LuxR and Plux, sfGFP, mCherry and LuxI respectively (Figure 2).

Figure 2. Depictions of the two transformed plasmids. One contains the regulator, LuxR. The other one Plux which responds to dimerized LuxR. LuxR dimerizes upon binding to AHL which synthesis is catalyzed by LuxI.

Subsequently, we let these colonies grow to different final population densities. This was achieved by varying glucose concentrations in a defined medium. [1] Population density was assessed by measuring absorbance at 600 nm wavelength. Fluorescence emitted by sfGFP and mCherry served as a read-out of the level of activation. A detailed protocol is available in Protocols.

RESULTS

Figure 3. A) Fluorescense per A600 in response to population density. Colonies were grown over night in media with varying glucose concentrations that lead to different final population denisities. With increasing absorbances at 600 nm, increasing fluorescence levels are observed. B) Proof of concept that final population densities can be modulated with the amount of glucose in a defined medium.

CONCLUSION

  • We can modulate the density a bacterial population reaches in defined medium by varying the amount of glucose.
  • The quorum sensing system shows a response to increasing population densities.

AND-gate without Quorum Sensing

OBJECTIVE
Determine expression levels of GFP production under the control of the AND-gate with different inducer concentrations. In this experiment we wanted to assess whether our designs would be capable to distinguish healthy and tumor tissue based on lactate and expected AHL concentrations.

PROCEDURE
Two plasmids required for the AND-gate were transformed into E. coli Top 10 (Figure 4). Cultures were grown in microtiter plates under combinations of 8 different AHL and 8 different lactate concentrations and measured after 5.5 hours. A detailed protocol is available in Protocols.

Figure 4. Schematic depiction of the two plasmids that were transformed for this experiment. Both lactate and AHL were manually provided in this experiment.

RESULTS
The different conditions cleary have an impact on expression levels of sfGFP under control of the AND-gate promoter. All three designs show increasing activation with increasing inducer concentration, even if the second inducer is not present. The highest fold-change for all designs however, is observed if both inducers are present in high amounts.

Figure 5. AHL Dose-Response Curve obtained by measuring fluorescence.

CONCLUSION

  • Leakiness of the synthetic promoter increases with increasing amounts of either inducer in the absence of the other.
  • Increasing AHL amounts have a greater influence on the leakiness in absence of lactate.
  • All three AND-gates exhibit highest inductions in presence of both inducers.
  • At lactate levels found in healthy tissue and low AHL concentrations, all designs are only weakly activated.
  • Design B performed best at distinguishing “healthy tissue lactate”, low AHL vs. “tumor tissue lactate”, high AHL. Design C, on the other hand, performed worst.

AND-gate with Quorum Sensing

OBJECTIVE
Verify the findings of the AND-gate characterization without quorum sensing with strains of E. Coli that contain additionally to the AND-gate also LuxI, the enzyme that catalzyes AHL production.

PROCEDURE
Two plasmids required for the AND-gate were transformed into E. coli Top 10 (Figure 6). Cultures were grown in microtiter plates in media with varying lactate concentrations. Density and fluorescence measurements were taken every 15 minutes to ensure a high enough time-resolution. A detailed protocol is available in Protocols.

Figure 6. Schematic depiction of the two plasmids that were transformed for this experiment. Lactate is provided to the system in this experiment, AHL is synthesized by the cells themselves.

RESULTS
The data is very noisy and it’s hard to make general statements about this systems behaviour. Despite this, a clear trend is visible for GFP to be higher expressed under lactate concentrations similar to tumor tissue than under those resembling healthy tissue or no lactate at all. With increasing population densities this effect becomes less pronounced (Figure 7).

Figure 7. Fluorescence normalized to population density vs. population density. Blue circles correspond to media lacking lactate, green to media containing 1 mM lactate, and red to 5 mM lactate. Circle styles correspond to three different biological replicates. It becomes apparent that with higher densities comes higher activation and that for lower population densities, lactate has a positive influence on GFP expression levels.

CONCLUSION

  • Due to a lot of noise in the data, conclusions have to be drawn with caution
  • Under lactate concentration mimicking tumor tissue, GFP gets stronger expressed than under lactate levels associated with healthy tissue.
  • Fold-changes are around 4 for design B and 2 for design A which is considerably less than observed in Figure 5. This might be due to a somewhat different experimental setup.

References

  1. Contois, D. E. "Kinetics of bacterial growth: relationship between population density and specific growth rate of continuous cultures." Microbiology 21.1 (1959): 40-50. doi: 10.1099/00221287-21-1-40