Team:SJTU-BioX-Shanghai/Design

Design

This year, we are trying to create a visualized monitor which can detect two or even more factors.

To achieve our goal of visualization, we use chromoproteins as our reporter. Chromoproteins, containing different pigmented nonprotein groups, are well known that their colors can be easily identified by naked eyes, so that we don't need to take certain equipment with us. Impressed by the excellent result that iGEM16_Imperial received, we decided to use small transcription activating RNAs (STAR) to control our circuits. We add specific terminators (STAR target or Target) before RBS to stop transcription, while a small RNA (STAR antisense or Antisense) can control the target to restart transcription.

Born of STAR3 system

Since we called the STAR system in IGEM16_Imperial STAR 1and our STAR system STAR 3, supposed that you must be curious about why we skipped the number ‘’’2’’’ and named the new system directly ‘’’3’’’. Now, we will tell you the story about how the new STAR system came into being.

Both Imperial’s and our project are based on the article which first described STARs. (Chappell J, Takahashi MK, Lucks JB. 2015. Creating small transcription activating RNAs. Nat Chem Biol 11:214–220.)

Figure 1. The STAR construct. In the absence of antisense, sense target RNA will form a stem-loop structure, functioning as a terminator and stop the following transcription. When antisense RNA, a mRNA fragment which is complementary to part of the target RNA, is transcribed, subsequently the terminator structure will be disrupted and then switch on the inhibited transcription.

In the article, there are four attenuators of four different strains have been examined and their performances are measured by fold activation (simply dividing fluorescent extension by OD600 of the bacteria) in the absence or in the presence of STAR Antisense. Imperial College chose the best STAR system originated from AD1 strain, which was shown that had the greatest fold activation.

Figure 2. Fold activation of modified attenuators. Fold activation of four modified attenuators of strain Anti-anti, T181, pbuE and AD1 respectively are labelled *. Fluorescence characterization was performed (measured in units of fluorescence/OD at 600 nm) in the absence of Antisense and the presence of Antisense in E.coli strain K12 MG1655. (Chappell J, Takahashi MK, Lucks JB. 2015. Creating small transcription activating RNAs. Nat Chem Biol 11:214–220.)

With the purpose of building a multifactorial detection system, we must have more than one RNA switch to control our signaling circuits. Thus, we need to apply more STARs in our project. Noticed that only strain T181 and AD1 have a fold activation above 10, we decided to use attenuator of strain T181 in our project and named STAR 2 system. The completed sequences of both Target and Antisense of four strains were presented in the article.

However, what we were worried about was that STAR 2 system didn’t work well as the STAR 1 because STAR 2 only has 17 fold activation compared to that of STAR 1, which has 153. What’s worse, according to record of Imperial, the actual fold activation of STAR 1 wasn’t so great compared to that in the article. As a matter of fact, we found that the STAR 2 system didn’t function well as we expected, with basic expression at a non-neglectable level and less fold activation.

During the summer, we have found some strategies to improve the STAR 2 system (according to another article Meyer, S., Chappell, J., Sankar, S., Chew, R., and Lucks, J. B. (2016) Improving fold activation of small transcription activating RNAs (STARs) with rational RNA engineering strategies Biotechnol. Bioeng. 113, 216) by changing constitutive promoter J23119 into a less efficient one J23150 and also add a hairpin to 5' end and a scaffold to 3' end of the Antisense in order to make Antisense 2 more stable, existing longer before being degraded. But our attempt failed due to unsuccess of inserting Antisense 2 into the vector. Nevertheless, we figured out another strategies to improve our STAR system that differed from those mentioned in the article.

We noticed that although target of STAR 1 (denotes Target 1) is shorter than that of STAR 2 (denote Target 2), it still worked better. So before we stepped further, we used RNAstructure to predict the structure of both Target 1 and Target 2. The result is shown below. (RNAstructure)

a

b

Figure 3. Structure prediction of Targets. To predict such RNA secondary structures according to DNA sequence, a online tool RNAstructure is used here. The color of each base pair indicates the probability of the structure (red indicates the greatest probability). Set Temperature to 310.15 K (37℃), Maximal Loop Size to 8, Maximum Percentage Energy Difference to 10, Window Size to 3 and Minimum Helix Length to 6. (a) Structure predicted for Target 1. (b) Structure predicted for Target 2.

Given the fact that Target 2 couldn’t stop transcription effectively as Target 1, we assumed that the two stem-loops together contributed to the high fold activation. Therefore we retained the first stem-loop of Target 2 and use the sequence upstream to make a second stem-loop, with the downstream sequence of poly T discarded. A little modification was done on order to make sure the length of the stem and length of the loop consistent with Target 1. The new STAR system we design from STAR 2 was denoted as STAR 3 system. DNA Sequence of Target 3 can be found at BBa_K2285020 and Antisense 3 at BBa_K2285010.

Figure 4. Structure prediction of Target 1 and Target 3. Condition Setting was the same as mentioned above. The red color of base pair indicates the greatest probability. We can see that the lowest energy of T3 (Target 3) is -41.2, lower than that of T1 (Target1) which is -32.6. More red regions appear at the first loop of T3 compared to that of T1.

To achieve multifactorial detection, we must ensure that the two STAR systems can control their own genetic circuit respectively and wouldn’t affect each other. Noticed that STAR 3 system in fact was designed from part of STAR 2 system, although we named it a new STAR system, it still keeps a great orthogonality with STAR 1 system because STAR 1 and STAR 2 system share no significance similarity, which had been proved by using Blast, and so the same as our STAR 3 system.

To characterize our STAR system, we use sfGFP (superfolder green fluorescent protein) as the reporter gene. Here are the plasmids we constructed. Plasmids pETDuet1 and pCDF-Duet1 are used for both of them have two multiple cloning sites (MCS), which can perfectly suit insertion of Target and Antisense fragments.

Figure 5. Characterization of STAR 3 construct. In the absence of Antisense 3, little sfGFP will be expressed with undetectable fluorescence intensity. When Antisense 3 is transcribed, fluorescence will be emitted under ultraviolet light. Both Target 3 and Antisense 3 are under the same kind of Anderson’s promoter J23119.

Reference:

1. Chappell J, Takahashi MK, Lucks JB. 2015. Creating small transcription activating RNAs. Nat Chem Biol 11:214–220.

2. Meyer, S., Chappell, J., Sankar, S., Chew, R., and Lucks, J. B. (2016) Improving fold activation of small transcription activating RNAs (STARs) with rational RNA engineering strategies Biotechnol. Bioeng. 113, 216.

Chromoprotein

Next step, we used chromoproteins to replace sfGFP in order to achieve better visualization.

Figure 1. Optimized chromoprotein expression construct. In order to make our monitor system more visible and more convenient for people to use, we use chromoproteins as our reporter gene instead, which exhibit colors without ultraviolet light.

And by mixing two types of chromoproteins, we can create a third color with a series of different chromaticity, which can be taken as the characterization of the relative abundance of two signal molecules.

We believe our monitor can be applied in various situations. To demonstrate its feasibility, we choose heavy metal ion detection as an application. We plan to use cjBlue and eforRed to indicate Co and As. Not only can we determine whether Co or Hg exists, but also can get their rough concentrations by analyzing mixed color. What makes it special is that we can get information about two factors at the same time only by working on one indication.

Figure 2. Vector design for Co2+ detection. RcnR is a repressor for Co promoter which control the transcription of Antisense 1, then control the expression of chromoprotein cjBlue. In the absence of Co2+, RcnR binds to the Co operator region and blocks the transcription of Antisense 1. When Co2+ exists, Co2+ binds to the RcnR and inactivates it so it no longer binds to the operator. Then Antisense 1 disrupts the stem-loop of Target 1 and cjBlue is expressed.

Figure 3. Vector design for As3+ detection. AsrR is a repressor for As promoter which control the transcription of Antisense 3, then control the expression of chromoprotein eforRed. In the absence of As3+, AsrR binds to the As operator region and blocks the transcription of Antisense 3. When As3+ exists, As3+ binds to the AsrR and inactivates it so it no longer binds to the operator. Then Antisense 3 disrupts the stem-loop of Target 3 and eforRed is expressed.

small CAT

Loader

In the way of searching for a tool to load our bacteria, we have tried many possibilities. Microfluid chip is too small to vision the color, dry bacteria might be hard to control. Finally, we found glass fiber filter to act as a loader, which means to conserve and shelter the bacteria.

Glass fiber filter has the exact pore size as the E.coli. So, if we filtrate the culture solution through this filter, it will conserve the thallus and leave the nutrient liquid.

In this operation, we complete two additional purposes, condense and restrict the bacteria. When looking through the culture solution, it is hard to reckon some tiny color change because of the shallow concentration and the color of nutrient liquid. After condense onto this filter, many small color changes can be much easier be captured by camera or naked eyes. Also, we need a container to restrict and standardize the area to make it possible for analyzers to compare between different experiments.

APP

To visualize the result of the protein, we plan to take picture and transfer the result by the image and RGB value and finally get the concentration. So, the APP ColorAnalyze we designed has two main functions, taking photo samples and analyzing the color.

When sampling photos, we got two methods. First, the APP can import photos from album, so we choose the pre-cut photos to calculate the colorimetric values. The cutting operation is controlled by users, so users cut off the incorrect edge and remain the sample by subjective opinion. Another way to select samples from the photo is setting the limits. Users can set a maximum and minimum value to restrict the RGB value. We use this method to get rid of the white background and dark dots.

When analyzing the color, there are many extraneous factors need to be controlled, the most important one is the light change. We designed a box to insulate the environment, so that we can control the light consistency inside the box. Except for the box, we also develop a method to calculate the standard curve to avoid the environment interference. Users let some bacteria reacting with the solution with standard concentration, and then we can reach the concentration of the sample by searching in the standard curve.

MATLAB

This part intends to realize the function: get a standard color chart from a n*n matrix which stores color information of combinations of two kinds of inducers. In different combinations, difference of concentration would result in difference of color. We use image processing method to generate a color chart. There are a few procedures in the program.

First,Read an image.

Second,Find standard area center.

Third,Calculate standard area RGB separately.

Finally,Make a gradient color figure by linear interpolation.