Team:ShanghaiTech/Project NFBL
Overview
We introduced Negative Feedback Loop(NFBL) to our control system, which contains three parts - named A, B and C - to ensure the precise control. As shown in Fig.1, part A activates part B, which then upregulates part C. However, part C will repress the expression of the part A.
Although the two-node loop, in which part A activates part C while part C represses part A in return, is enough for a negative feedback, it is more unstable compared to the three-node loop, indicating that it is harder for the output of two-node loop to follow the input owing to the huge fluctuations. The improved performance is resulted from the existence of part B in the three-node feedback loop. By working as a buffer in the system, the output downstream part B is allowed higher sensitivity and fidelity.
Fig.1 Negative feedback loop compared with two-node loop.
Lux is put into our negative feedback loop as the part A. It is reasonable to choose it rather than other induction systems as Lux shows better input-output linearization, which is a good beginning for the output to follow the input.
As the signal has been input to the system by part A, we want to ensure the fast response of the other parts. There is no better choice than the RNA switches, since they do not need to be translated, which save much time. Therefore, STAR and pT181 are introduced to our system as part B and part C.
Besides the fast response of the two parts, they also show advantages on the efficiency. The activation caused by STAR could reach almost 100-fold, which means we could reduce the signal attenuation as much as possible in our system. Meanwhile, the repression of pT181 is 98%. The high repression effect can reduce the inevitable leak as just tiny quantity of it will prevent the expression of part A.
In conclusion, the feedback loop and the parts we choose can ensure not only the fast response of the system, but also the high-fidelity of the output.
Lux
The Lux Module is the input signal controller of our Negative Feedback Loop(NFBL) as the part A. In our project, we utilized quorum sensing system's sensing part as the signal sensor of the Module, this allows downstream gene expression to be operated by external signal. We have examined four quorum sensing systems(Lux, Tra, Rhl, Rpa) and finally choose the Lux as the input part of our module.
Key Achievements
Construct Lux, Tra, Rhl, Rpa quorum sensing system
Evaluation for each quorum sensing system
Overview
Quorum sensing is a natural occurring mechanism that certain strains of bacteria use to regulate gene expression in response to their population density. As quorum response proteins be translated, signaling molecules such as N-acyl homoserine lactones or AHLs could bind transcription factors and activate downstream gene expression as well the cell’s growth.
In our project, quorum sensing response genes were connected to receptors from other modules, and associated AHLs were added to control downstream gene expression. We have examined four quorum sensing systems (Lux, Tra, Rhl, Rpa) to evaluate weather each quorum sensing system is suitable for our Negative Feedback Loop (NFBL) or not.
Fig.2 Mechanism of the Lux Module
Through experiments, Lux show advantages in its robustness and sensitivity, including:
Lower background leakage: Downstream expression perform very low when associate AHL was note added.
High sensitivity: Only very little amount of associate AHL were needed to activate downstream expression.
Stable expression: Downstream gene expression performed stable even after 8 hours after adding associate AHL.
According to Lux’s outstanding performance its sensitivity and robustness, we chose the Lux quorum system to control module A.
Our Approach
The first node of our three-node negative feedback loop, was controlled by the Lux quorum sensing system and the repressive RNA switch pT181 from module C, which was designed as “pCon(J23100) + pT181 Target + RBS(BBa_B0034) + LuxR + TER(BBa_B0015)”.
We believe that, as pT181 RNA switch is open, the Lux transcriptional activator, LuxR, is translated. With the induction of exogenous AHL, LuxR will combine with AHL molecules and activate the Lux transcriptional promoter, pLux, and lead downstream gene expression.
In the absence of AHL, the pLux promoter will not be activated, which means the system keeps off without input signal. When the concentration of AHLs changes, the Lux module start to work, activate the part B. Due to the nice input-output linearization of Lux, the input signal waveform will be well introduced to our system, without being changed.
Experimental design
Experiments were carriedout to determine which quorum sensing system should be chosen. We designed different plasmids whom have different Quorum resporter devices and add Aequoria victoria green fluorescent protein at the downstream of Quorum transcriptional promoters to detect the performance of different quorum sensing concentrations.
Fig.3 "pCon(J23100) + RBS(BBa_B0034) + QuorumR + TER(BBa_B0015) + pQuorum + RBS(BBa_B0034) + GFP + TER(BBa_B0015)"
DH5-α cells whom contains plasmid of constructed resporter devices were cultured to Abs600 0.60 and then transferred to 96-well microplates where they were induced with appropriate AHL concentrations. By measuring the fluorescence output from each of the constructed devices by inducing cell cultures with various concentrations of AHL molecules. The induced cell cultures were grown in the microplates at 37°C and the fluorescence signal were monitored over time by using a microplate reader.
We needed to characterize the response of the construct to different concentrations of AHL so that we could use the data in our model to predict how the system could function. The activation ranges were compared between the quorum sensing systems in order to determine their robustness, sensitivity, stability and if it is suitable to be used in our Negative Feedback Loop.
Results
Through experiments, concentration ranges of AHLs required for activation in each quorum sensing system were calculated to be 100nM-10uM for Rhl and 100pM-10nM for Lux, Tra, Rpa. Rhl were found to differ by a 1,000-fold sensitive difference than other quorum sensing systems, which meas Lux, Tra, Rpa has higher sensitivity.
Fig.4 LuxR-AHL Fluorescence 485-535 absorbance related to time.
Fig.5 RpaR-AHL Fluorescence 485-535 absorbance related to time.
Compared between different quorum sensing systems, the Lux quorum sensing systems has lower background leak, higher orthogonality and robustness and better expression stability. Therefore, we choose Lux quorum sensing systems as the input signal controller for module A.
STAR
Small transcription activating RNA (STAR) is the part B of our feedback loop (NFBL), which works as a buffer in our three-node feedback loop. STAR is a significant switch which determines whether the gene of interest downstream could be expressed or not. With the presence of STAR, the target gene will be activated due to the combination of it with the terminator RNA upstream the gene of interest. While the target gene is inhibited without STAR. STAR is a good part as small RNA to construct synthetic gene networks that precisely control gene expression.
Key Achievements
Characterization of STAR-mediated target gene activation over a period of time.
Overview
As the Three-node Negative Feedback Loop is needed for the precise control of the transcriptional level, a buffer is necessary to be the upstream of the target gene. STAR is chosen to play that role.
The STAR contains a terminator and a STAR antisense.
The terminator resides in the 5΄ untranslated region of the target gene. In the absence of the STAR antisense, the terminator will be allowed to form, which will prevent the transcription of the target gene.
With STAR antisense present, it will anneal with the 5’ region of the terminator stem, which will prevent the formation of the terminator. In this way, the RNA polymerase is able to start transcribe.
There are obvious advantages for introducing STAR into our system, including:
Low metabolic burden: Compared to the universal protein-based regulatory systems, STAR dispense with translation step as a RNA regulator, which saves a great amount of resources.
High efficiency: The presence of STAR produces an expression of the downstream gene whose activation is 100-fold compared with the condition when STAR is absent, which provides the efficiency and validity.
Fast response: STAR acts quickly when receiving signal molecules and causes little delay in the circuit because of rapid RNA degradation.
Our approach
STAR is the part B in our negative feedback loop. With the existence of the part B, which works as a buffer in our system, our three-node feedback loop will show higher fidelity than the simple two-node feedback loop.
STAR will be activated by part A, since it is regulated by the Lac I promoter. Meanwhile, STAR will upregulate the expression of part C since the terminator part of the STAR, which will be regulated by the STAR antisense is upstream the gene for part C. Consequently, the upregulation of part A caused by the input signal will indirectly activate part C with the help of the buffer - STAR. In this case, the whole system is ready to response change of the input signal.
After the input is introduced into the system, we want the following parts response fast to the signal. Since STAR shows fast response due to the lack of translation step and rapid RNA degradation, it is the appropriate choice for the part B.
In the absence of input signal, the system is expected to be shut down. However, the leak is inevitable in all the control systems. But it seems to be harmless as the STAR shows activation about 100-fold. It will also be reduced by the pT181 attenuator, which will be mentioned in the part C.
When the input signal is introduced into the system by the part A, STAR will send it to both the regulator- the part C - and the output downstream it due to the fast response mentioned before. In this case, the high efficiency of the activation of STAR, the signal attenuation will be reduced at utmost.
In brief, the STAR can help our system respond to the input with high-fidelity in a fast way.
Experimental design
The experiments on STAR can be generalized into two parts. Almost all the experience about STAR are operated under 20-25 degrees Celsius. Streptomycin medium has been used to cultivate bacteria in need. We mainly use plate reader and fluorescence microscope to get results of how STAR has been working. Massive data about corresponding bacterial density and fluorescence intensity have been detected every experience and have been compared(FL/ABS) in order to analyze the variation of fluorescence intensity of a single bacterium.
We really appreciate the precious gift from SJTU, which is the STAR plasmid conduct by Imperial College in 2016 from SJTU. The plasmid consists of a STAR target upstream of a superfolderGFP with a RBS(BBa_B0034) and a terminator(t500) under control of a constitutive promoter (J23119). We transformed it into DH5-α E.coli cells.
Fig.6 STAR plasmid form Imperial College pCon(J23119) + STAR.Target1 + RBS(BBa_B0034) + superfolderGFP + Terminator(t500)
Firstly, two different types of cells that contain different GFP (SFGFP as enhanced group and generic GFP as control group) were cultivated and examined using plate reader and fluorescence microscope, from this part of the experiment we wanted to verify the effectiveness of GFP in part B in order to make sure that it could express well in our system later. Additionally, by comparing the light intensity of different types of GFP, part B would be proved to be sensitive to embody its working status, which is helpful for our system to handle the upstream information, also convenient when manipulators want to analyze the situation of system.
The second part was designed to measure the growth and green fluorescence of cells under time span, in order to coordinate with data and curve from other parts in our experiment, the graph we drew based on collected data would help us find out the characteristic and change of expression of STAR during a long period. Then FL/ABS is calculated likewise to give a clear expression of how the expression of a single bacterium varies from 0 to 15h.
Result
By analyzing these graphs below, the result shows that the GFP used in part B is an enhanced GFP whose intensity is nearly 10 times as the control group, which proves to be greatly sensitive. Meanwhile, this experiment proved part B works and creates a brighter image when observing under fluorescence microscope.
Fig.7 The growth of cells harboring plasmids containing two different types of GFP (enhanced, control) at 15 hours, which is shown to approach the same level.
Fig.8 The expression of two different types of GFP (enhanced, control) at 15 hours, which is shown to approach the same level.
Fig.9 The logarithmic form of fluorescence intensity divided by cell density respectively for STAR1, STAR2, Blank group at 15 hours.
Fig.10 100x image of STAR1 under fluorescence microscope.
Fig.11 100x image of STAR1 under fluorescence microscope.
Fig.12 100x image of STAR1 under fluorescence microscope.
Plus, how STAR works according to time can be seen from the graphs below.
Fig.13 Bacterial density of STAR used in part B. The growth shows a steady increase from 0 to about 11h, then the numerical result fluctuates at approximately 1.04.
Fig.14 The expression quantity of STAR varies with time and stabilizes at about 15 hours.
Fig.15 The ration of Fluorescence intensity divided by cell density. The curve first decreases sharply and then increases slowly, eventually stabilizes at about 15h.
Reference
A network of orthogonal ribosome·mRNA pairs, Oliver Rackham ; Jason W Chin, Nature Chemical Biology, 2005, Vol.1(3), p.159
pT181
The Repressive RNA Switch pT181 attenuator is the extra regulator of our Negative Feedback Loop (NFBL). We are utilizing pT181 attenuator – a dual control repressors – to regulate both gene transcription and translation in a fast and robust way. We have submitted it as our improved part since it increases repression from 84% to 98% compared with that of Kyoto 2013 submitted. As RNA transcriptional regulators are emerging as versatile components for genetic network construction, we believe that improving the part in this library is essential for advancing synthetic biology. We hope our improvement of pT181 attenuator to iGEM parts will encourage future teams to implement this versatile, highly orthogonal, and effective regulator in their circuits.
Key achievements
Characterization of pT181 attenuator -mediated target gene activation in various conditions.
Improvement of a new sense target sequence of pT181 attenuator.
Update of the BioBrick Registry library by improving a RNA-logic toolset.
Overview
Our engineered cells need a Three-Node Negative Feedback Loop to construct a more sensitive and high-fidelity control system. And pT181 attenuator is the part that plays the role of repressor in this loop.
pT181 attenuator is a part composed of a sense target sequence and an antisense RNA that can regulate gene transcription and translation. Residing in the 5΄ untranslated region of the target gene, it can regulate the expression of a downstream gene at both transcriptional and translational levels.
Fig.16 A schematic representation of the pT181 attenuator in action
At transcriptional level, the anti-terminator, which is a part of the sense target sequence will anneal with the 5’ region of the terminator stem without antisense RNA. In this case, the terminator could not format, which means the RNA polymerase can start transcribe. At translational level, in the absence of antisense RNA, a ribosome binding site (RBS) for the gene of interesting is exposed, so that the ribosome could bind to it to begin translation.
While the antisense RNA is present, the formation of terminator will be allowed as the anti-terminator is sequestered due to the kissing hairpin interaction between the antisense RNA and the sense target sequence. In this way, the downstream transcription will be prevented. As for the translational level, the occlusion of the RBS by the terminator hairpin will prevent the translation of the target gene.
In conclusion, while antisense RNA is not present, the gene will express as normal, while antisense RNA is produced, the downstream target gene will be repressed effectively.
The RNA regulators show sufficient advantages over traditional protein-based regulatory systems, including:
Programmability: As Watson-Crick base pairing is predictable, the RNA-RNA interaction can be predicted by sophisticated software tools. In this way, a RNA switch can be designed artificially, which are difficult for proteins.
Lower metabolic cost: Compared with proteins, the RNA switches dispense with translation step, which saves a great amount of resources.
Fast response: RNA switches could propagate signals faster than proteins considering the fast degradation rates of RNAs.
Despite these advantages, RNA regulators still suffer from incomplete repression in their OFF state, making the dynamic range less than that of the proteins. This leak can cause the network to function incorrectly. Therefore, we submit the dual-control pT181 attenuator, which can solve this problem.
The dual-control pT181 attenuator we submitted offers a significant advantage over previous iGEM parts that submitted in 2013:
Reduce leak: As our pT181 attenuator could regulate both transcription and translation in a single compact RNA mechanism, which means it could provide stronger functions without increasing burden. This dual control repressor is able to increases repression from 85% to 98%.
Our approach
pT181 attenuator is the part C in our negative feedback loop. As the regulator in our system, pT181 attenuator will be activated by part B, since we place part B’s target sequence upstream the gene for pT181 attenuator antisense. Meanwhile, pT181 attenuator will repress the part A due to the pT181 attenuator sense target upstream the gene for part A. Consequently, the upregulation of part B caused by the expression of part A will indirectly repress part A. In this case, the whole system is ready to response change of the input signal.
In the absence of input signal, we hope that there is no expression of the output. However, due to the inevitable leak from the part A and part B, it seems to be impossible to avoid expression of the output. With pT181 attenuator in our system, we could avert it at utmost because of the efficient repression of it – just tiny quantity of pT181 attenuator will prevent most of the leak. This ensures the minimum expression of the output without input signal.
When the input signal is present, it will induce the expression of part A, which will upregulate the part B a lot. In this way, the output will be strongly expressed. But the upregulation of part B would cause the expression of pT181 attenuator, which will repress the part A. And this will cause the low expression of part A and indirect low expression of part B. Although the remains in the environment would keep the amount of output, the low quantity of each part of the system will prepare it for any change from the input signal. In this case, the output of the system would response to the change of the input in a very short time. This can also eliminate the possible superposition between outputs from different input signals.
We used our model to predict whether the high repression effect of pT181 attenuator will cause the silence of the output, as it may cause the silence of part A. However, our model shows that the output will respond to the input perfectly, which support our experiment a lot.
In conclusion, the presence of pT181 attenuator will reduce leak of our system at utmost, as well as allow it to rapidly respond to the changing signal.
Experimental design
Fig.17 A schematic representation of the experimental group plasmid. This has the basic pT181 attenuator Antisense under control of a constitutive promoter, as well as a GFP gene downstream of the pT181 attenuator sense target under the control of a constitutive promoter.
Fig.18 A schematic representation of the positive control plasmid with the GFP gene downstream of the pT181 attenuator sense target under the control of a constitutive promoter, without a pT181 attenuator Antisense on it.
We generated two plasmids, one is for experimental group, the other is for a positive control. The experimental plasmid, which is the pT181 attenuator in this experience, contains the antisense sequence downstream of a constitutive promoter and followed by a double terminator on a high-copy plasmid. Meanwhile, there are also a GFP gene with a ribosome binding site downstream of the pT181 attenuator sense target sequence. The GFP coding sequence is also downstream of a constitutive promoter and followed by a double terminator. The positive control plasmid, which is the blank in this experience, contains the same as the experimental plasmid except for the antisense sequence.
We did a group of pT181 attenuator expression experiments. First, as a part that needs to show strong inhibition, we should ensure that its inhibitory effect is obvious enough. Therefore, we compared experimental group and positive control group, which is transformed into a normal GFP plasmid. Depending on the GFP expression, we can prove that our work has a high credibility. Additionally, in the group above, three types of flora from Interlab are cultivated for contrast. The aim is to compare the statistics of pT181 attenuator and verified Interlab to find out which repressor level pT181 attenuator is in when put into practical application.
For this group, we transform different plasmids into the E. coli in the tubes and cultivate for hours (37℃, 220RPM). Then we used ELISA plate to detect the change of fluorescence and OD600 over time. What should be noticed is that we set the original flora at OD600=0.05 to guarantee flora proliferating at the same concentration. As the repression of pT181 attenuator attenuator is so powerful that the fluorescence of the experimental group is hard to detect. As a result, to remove LB medium’s fluorescence background, we centrifuge fluid, take out supernatant, add PBS buffer and resuspend before detect.
Results
The form showed in figure 21 tells that the expression of GFP is greatly decreased in the presence of pT181 attenuator of 18 hours of culturing. The data from figure 19 shows that the pT181 attenuator is not harmful to the cells. The data suggests that pT181 attenuator is a promising tool for the regulator part in our Three-Node Feedback Loop.
Fig.19 400x image of positive control under fluorescence microscope.
Fig.20 400x image of experimental group under fluorescence microscope.
Fig.21 Characterization of pT181 attenuator in DH5-α E.coli cells. OD600 monitored over time for cell lines incorporating the pT181 attenuator in the absence or presence of the pT181 antisense. The result shows that the pT181 antisense is not harmful to the E.coli, which provides convenience for test for fluorescence as we do not need to normalize the OD600.
Fig.22 Characterization of pT181 attenuator in DH5-α E.coli cells. The figures show the fluorescence for cells with or without pT181 antisense. (a) Fluorescence monitored over time for cell lines incorporating the pT181 system with pT181 antisense. It shows that the GFP can be expressed in the pT181-attenuator, and the expression level increases gradually. (b) Fluorescence monitored over time for cell lines incorporating the pT181 system without pT181 antisense. It matches the curve of how GFP’s expression increases without being repressed, which establishes foundation for measure the repression effect of pT181-attenuator. (c) The combination of (1) and (2). We could see the sharp difference in the fluorescence between the two curves. This proves our pT181 could repress the expression of GFP as expected, which means our part C is able to produces repression effect as anticipated. This shows that the controller in our Three-Node Feedback Loop is constructed successfully.
Fig.23 Characterization of pT181 attenuator in DH5-α E.coli cells. endpoint fluorescence (18 hours) for cell lines in the absence or presence of Pt181. The data shows that our Pt181 attenuator could repress the target gene for 98%.
Reference
Achieving large dynamic range control of gene expression with a compact RNA transcription-translation regulator, Westbrook, Alexandra M ; Lucks, Julius B, Nucleic acids research, 19 May 2017, Vol.45(9), pp.5614-5624