From 2006.igem.org

Home	Background and Signalling Pathway	Project & Model	Methods	Terms & References	Team members

Results

Construction of Biobricks

We prepared individual BioBricks that could be used in mammalian cells. The following BioBricks were used:

• promoter (constitutive CMV and NF-κB inducible promoter),
• terminator (which we included into plasmids to avoid repeated additional steps in construction of each functional Part)
• protein coding sequences, which include:
  • two inhibitory (dominant negative) proteins of the signaling cascade (dnMyD88 and dnTRAF6),
  • two reporters: Renilla luciferase (in order to allow simultaneous dual luciferase assay to normalize the luminescence for the efficiency of transfection and cell number) and GFP
  • PEST sequence to decrease the lifetime of the inhibitor.

Linkers between the protein coding sequences (e.g. MyD88+luciferase+PEST) were prepared by PCR ligation to avoid addition of unwanted aminoacids in the linker region between the protein domains.

Detection of phosphorylation of ERK kinases fy flow cytometry reflects the activation of the TLR signaling

Figure 19: Flow cytometric profile of phosphorylated ERK kinases in HEK293 cells stimulated by LPS for 0, 20 and 80 minutes.

Flow Cytometry experiments showed that after 80 minute stimulation the increase in the level of phosphorylated ERK kinases was evident(Figure 19). The method was not as sensitive to the rapid response as we hoped. Further optimization of the method would be necessary in order to obtain the quantitative data to accurately model the cell signaling.

Detection of transcriptionally competentNF-κB in whole cell extracts dependent on the time after stimulation with LPS determined with promoter ELISA method

Figure 20: Activity of peroxidase (conjugated on secondary antibodies) in luminescence units as a function of time after stimulation with LPS. The amount of free NF-κB (and consequently luminescence) increases with time of stimulation (7 hours) and decreases after extended (23 hours) stimulation.

Cells were transfected with TLR4 plasmid and MD-2 plasmid in ordeer to confer the LPS responsiveness. We have shown that cells with TRL4 receptor respond to LPS stimulation and that ELISA is apropriate and sensitive method for the detection of stimulation. However we did not get satisfying results when stimulating (with LPS or flagellin) cells transfected neither with dnMyD88 nor TRAF6. The response of the ELISA method, as the most direct measurement of the free NF-κB was not sufficiently reproducible and linear in comparison to the method of luciferase reporter plasmids (see below).

Inducible transcription by the stimulation of Toll-like receptor

Stimulation of HEK293 cells transfected by TLR with their agonist (flagellin in case of TLR5) stimulates the translocation of NF-κB into the nucleus and activation of transcription of NF-κB-responsive genes. We have used the firefly luciferase under the NF-κB dependent promoter to monitor the promotor activation. We tested the functionality of our part containing NF-κB dependent promotor in a construct with dnMyD88-rluc. We detected the R-luciferase activity only in stimulated cells with significant activity appearing a few hours after the cell stimulation, which proves the functionality of this part (inducible promoter).

Figure 21:Transcription of dnMyD88 with Rluc reporter is induced by bacterial flagellin Activity of Rluciferase stimulated by the addition of a pulse of flagellin. HEK293 cells were transfected with device comprising NF-κB-inducible fusion of dnMyD88-Rluc.

Figure 22:Transcription of dnMyD88 with Rluc reporter in our device is inducible by bacterial flagellin. Activity of the Fluciferase stimulated by the addition of two pulses of flagellin.

We also measured NF-kB-responsive F-luciferase activity of this device. We expected a decrease in the activity after the amount of time sufficient for the sysnthesis of inhibitory fusion protein in cells. To our dissapointment no decrease of the cell activation was observed. One of the possible causes for the lack of inhibition of the dnMyD88-Rluc protein fusion could be the sterical hindrance of the activity of dnMyD88. This domain interacts with TIR domain of TLRs in the direct proximity of the cellular membrane, which could prevent interaction with larger protein domains. Based on those results we also conducted our experiments with a device containing dnMyD88 without of any protein appendix.

Inhibition of cell signaling by a dnMyD88 feedback device

An adapter protein MyD88, consisting of a TIR and death domain, is at the crossroads of Toll-like receptors. Therefore activation of each of the surface expressed TLR recruits MyD88 to the cell membrane. We selected to use the receptor TLR5 and flagellin instead of the often used TLR4 and its agonist LPS, because the TLR4 system is extremely sensitive to the contamination with bacterial LPS, which often copurifies with the plasmid. In this way we have avoided the contamination issue, but our system is designed to work on any signalling receptor that has MyD88 in its signalling pathway (i.e. TLR1,2,4,5,6,7,8,9,11,IL-1).

We expect that after stimulation with flagellin cytoplasmic NF-κB is released from its inhibitor. It then migrates into the nucleus and induces the transcription of our construct. Accumulation of the dominant negative protein should compete with wild-type MyD88 and attenuate the downstream signalling pathway. We can detect the NF-κB translocation, transcriptional activation and phosphorylation by our detection systems - luminescence, free NF-κB, phosphorylation detected by flow cytometry and confocal microscopy. In case that our construct is followed by PEST (degradation) sequence, it is expected that product should degrade rapidly and shorten the inhibition period. If the stimulus is still present, cells should again respond to it and the cycle with inhibition will repeat as well.

We have designed the experiment by two consecutive pulses of TLR stimulation separated by 4-6 hours, which is sufficient time to accumulate the inducible inhibitor. Normal cells should respond to both stimuli, while the cells with our device should respond to the same extent only to the first pulse -- this induces the transcription of the inhibitor, therefore the response to the second stimulus should be decreased. However if the stimuli are separated by a longer timespan, the system should respond normally, as the inhibitor has already degraded and the systemn has reset to the normal state.

Figure 23:dnMyD88 feedback loop decreases the cellular activation by repeated stimulation with bacterial flagellin. Normalized luciferase activity (Fluc/Rluc) of cells stimulated by the addition of two pulses of flagellin separated by 6 hours, indicated time after the second pulse is shown. HEK293 cells were transfected with device comprising NFκB-inducible dnMyD88 in comparison to wild type.

Our results show that the system with inducible dnMyD88 indeed responds weaker to the second stimulus. The minimal delay between two stimuli should correspond to the time required for the synthesis of the inhibitor, which is exactly as planned, since it does not completely abolish the cellular response but decreases it after extended stimulation.

Decrease of the protein lifetime by the PEST sequence

Rapid degradation signature motif - "PEST" was considered to tune the lifetime of the inhibitor in order to reset the cells to the normal responsiveness more rapidly. Fusion of the luciferase and PEST sequence under the control of CMV (Device [http://partsregistry.org/Part:BBa_J52039 BBa_J52039]) was used to test the effectiveness of the PEST sequence. We have used the cycloheximide to stop the protein synthesis and analyse the lifetime of the fusion protein with PEST sequence. The steady state of the luciferase activity was 5-10 fold lower in the Rluc-PEST fusin than without of the PEST signature, as evident from the chart below.

Figure 24:Addition of the PEST degradation tag decreases the steady state of the protein in cells due to its rapid degradation . Cycloheximide was added to cells in order to stop the protein synthesis and monitor the protein degradation kinetics.

Constructs of the dnMyD88 domain and PEST sequence were not effective as inhibitors of signaling because of the interference of the C-terminal extension of the MyD88 inhibitory domain, as described above for the fusion constructs with the reporter luciferase or GFP.

Suggestions

• for mammalian cell systems an additional set of vectors may be constructed based on the backbone, which includes the selection marker for stable transfection or retroviral integration

Home	Background and Signalling Pathway	Project & Model	Methods	Terms & References	Team members