Matching and Testing
The dynamics pattern of each pair of pInv-gen and pInv-rep
Fig-T-5: The elements of our concern in invertase module. Through changing such components we wish to understand their property and illustrate an optimized combination of them.
We co-transform pairs of pInv-gen and pInv-rep into E. coli BL21(DE3) or Top10, and have measured till now 13 different combinations them (see Table-T-1). A detailed pattern of relationships between time and RFU/OD can be hence revealed. Most of our data indicate the RFU/OD – time graphs shares a similar pattern (See Fig-T-6). The signal of Cre fused with EGFP stably increase after induction, perhaps a linear relationship; the mcherry signal, however, resembles an S-type curve that show a tiny or no growth and suddenly erupt a period after induction. Later, the increasing rate damped and the curve moves to a plateau phase. But technically, each pattern is slightly different because the molecular element of in this pathway varies from each other.
There are 3 major variants in this study, invertase (itself), promoter, and the ssra-mediated degradation, and additionally the fusion site of EGFP onto invertase does significantly count (See Fig-T-5). Using a series of combination of these variants we can understand their effect on invertase module.
Table-T-1: A detailed list of information of each pInv-rep and pInv-gen in this study.
Fig-T-6: The measured dynamics pattern of all 13 combinations of pInv-gen and pInv-rep. For relationship of group name and there corresponding device, please check Table-T-1.
The properties of elements in invertase module
1.Promoter intensity is positively-correlated to expression.
It is easy to understand that the higher intensity of a promoter, the better it performs to initiate transcription. We tested 5 constructive promoters, that is, J23100, J23101, J23106, J23110, and J23116, on pInv-rep (see Table-T-1 and Fig-T-7). On the one hand, the plateau phase RFU for different promoter is positively-correlated with its promoter intensity; on the other hand, another important value, the length of time from adding IPTG to burst of expression rate, is negatively-correlated to promoter intensity. The later value is perhaps more meaningful, because it can be used as a standard to help us define “inversion time” of an invertase module.
Through this series of data we can understand that promoter element has a significant potential to control the inversion time. Specifically, a stronger promoter leads to a shorter timing length. This helps us to modify the module into intended timing length.
Fig-T-7: The comparison of 5 constructive promoters on pInv-rep. The curves represents mcherry signals by pInv-rep. All condition are controlled expect for promoter of pInv-rep.
2.The N-term fusion of FP onto invertase deteriorate its activity but is more stable
All the invertase in this study is a fusion protein to EGFP. However, we must be cautious when using such material, in that a fusion (especially of an entire protein) might interfere its 3D-structure and folding due to steric hindrance.
Hence, we prepared EGFP fusions onto either N-term or C-term of Cre (namely EGFP-Cre or Cre-EGFP), with an 8 AA flexible chain as linker to provide larger space for proper folding. The dynamics of the two proteins show dramatically different pattern (see Fig-T-8). When EGFP is linked to the C-term of Cre, it can be expressed at a slow but stable rate. Cre-EGFP works pretty good since we obtained obvious burst and increasing of mcherry signal. However, if EGFP is fused onto Cre’s N-term, it renders dramatically high concentration of EGFP-Cre expression, but the invertase activity is almost lost. The phenomenon indicates a possibility that if the N-term of Cre is linked to EGFP, its 3D structure might be interfered.
Fig-T-8: The comparison of effect of different fusion sites on Cre. The curves represents Cre-EGFP and mcherry signals. All conditions are controlled except for fusion sites.
Fig-T-9: The 3D structure of tetramer of Cre, the active form of Cre. Image is obtained from RCSB Protein Data Bank ( Http://www.rcsb.org/pdb/explore/explore.do?structureId=3MGV). A, the Z-axis view of Cre tetramer, 4 colors indicates 4 monomers. B, the N-term of Cre, blue arrows marks the M28, the starting amino acid of simplified Cre of BBa_K1179058; linking to a large protein like EGFP at this point take risks to generate steric hindrance. C, the C-term of Cre. Yellow arrow marks the final amino acid; a fusion to this site can be much better to maintain the structure.
We find some evidence to support this hypothesis. The activity of invertase requires the formation of tetramer of Cre in combination with target DNA (see Fig-T-9 A), forming the structure so-called Holiday Junction. At this stage, Both N and C terminus seems to be loose with relatively open space, which is a good structure compatible of fusion (see Fig-T-9 B, C). However, the Cre we used, BBa_K1179058, is a truncated type of Cre that only maintains perhaps necessary part of the protein, so actually the translation starts from the No.28 Met, as marked in Fig-T-8 B, which is at the middle of an alpha helix locating pretty close to other part of the protein. Hence, we believe this is one of the reason why N-term linking to EGFP deteriorate its activity. Yet although this protein shows less activity, the expression rate is far higher than Cre-EGFP, probably because such structure can be folded quicker.
Therefore, since other invertases belongs to Cre-family, we construct only invertase-EGFP fusion with no EGFP-invertase issues.
3. Ssra can significantly avoid leakage but slightly reduce the inversion efficiency
Ssra-tag (e.g. BBa_M0051), if linked to the C-term of a protein in E. coli, will lead this protein ClpX or ClpA protease, rendering effective degradation. In this study we wish to utilize this tool to clean up certain invertases when going to next round of timing, and reduce the leakage of expression when not induced. Hence, ssra-tag, theoretically, will decrease both protein expression and leakage, prolonging the time for total inversion of an invertase module.
Fig-T-10: the effect of ssra-tag on invertase dynamics. A, a comparison of EFGP expression of all pInv-gen. Adding an ssra tag can always reduce net expression rate, comparing to corresponding group. B, a comparison between group “pET-CG & 101Lox-M” and “pET-CGS & 101Lox-M”. The relatively slow net expression rate of Cre-EGFP-ssra renders a reduced intensity of mcherry expression.
Fig-T-11: Adding an ssra-tag can restore the activity of EGFP-Cre. Although strains with EGFP-Cre grows at a significantly slow speed and shows tremendous accumulation of EGFP-Cre while hardly express mcherry, adding an ssra tag dramatically restarts the inversion.
This is confirmed by our experiment. When adding an ssra-tag onto either mcherry or Cre-EGFP, its expression can be repressed at a certain level, comparing to those without ssra (see Fig-T-10). However, we do not have to worry about the risk that ssra is so strong that no enough invertase can be generated: even if we uses ssra-tag to modify Cre-EGFP, it still performs pretty good (See Fig-T-10 B).
Additionally, ssra tag can even restore the enzymatic activity of EGFP-Cre. When EGFP-Cre-ssra is expressed, it successfully overturns the reporter, although at a lower speed than Cre-EGFP-ssra (see Fig-T-11).
We decide to use ssra on other invertases.
The comparison of efficiency of different invertases (In fact, a good lesson!)
There are 5 other Tyr-family recombinases, which have invertase activity. 4 among them, Dre, Vcre, Scre, and Vika are Cre-like invertases. Our team this year synthesized these 4 Cre-like invertases which is not previously in the biobricks and wish to test their function and utilize them to construct more timer.
Yes, perhaps we can catch up with the DDL of sending parts of them and will show the result in our poster and presentation. But we should have been able to finish this job far earlier and should have already measured their real-time dynamics – if it is not due to a horrible oligo synthesis company (Guangzhou IGE Biotechnology Ltd.) that totally ruined our clones by providing impure primers (even so-called “PAGE purified”) with deleted bases and making tremendous numbers of frameshift CDS.
This is a good lesson; we strongly recommend all iGEMers be careful to choose primer producer.
The strategy to define inversion time of each invertase module
To understand the length of timing for each combination of pInv-gen and pInv-rep is our primary concern. Roughly, this time can be considered as the time interval between the initiation of burst of green and red signal. This definition through visual observation is pretty subjective and not yet precise. We hence wish to design a precise method to understand inversion time for each module.
Luckily, due to our modeling work aims to solve this problem, we can find a far better solution. In real-time invertase dynamics measurement system, we can describe this process totally through mathematical function by data fitting. One of the greatest outcome is that we can define “time interval” be quantificational method. We will use the first derivative of both green and red signals, and the timing length of the invertase module is defined as the length of time between the two points when growth rate of two signal reaches 1/2 maximum level.
To see how data is further processed and invertase module timing length is calculated, please turn to MODELING for detail.
Bacteria Timer
Introduction
A report from Science[1]（在本版面最后添加参考文献）, by which we were inspired, tries to explain that synthetic gene networks can be constructed to emulate a cellular counter that would enable complex synthetic programming and a variety of biotechnology applications.
One of the figures (fig. 1) from this article indicates how genes can work in a counting system by reversal of recombinases. Two recombinases in the circuit, Flpe and Cre, in conjunction with their specific recognition site, FRT and LoxP, accomplishes the whole flipping process. Convenient to distinguish, we'll call it circuit 1.
Based on Circuit 1, we designed what we call circuit 2 (fig. 2) , which is, to our perspective, more functional and less induced, by principally altering the location of genes.
The comparison between them indicates 2 advantages of circuit 2 over circuit 1 (fig. 3) .
First, it continues transcripting and flipping in a circulation once induced. It happens theoratically because it may be bothered by objective resistence, but it provides us with a possibility to time gene reaction and control certain protein expression in a time scale.
Second, circuit 2 can transfer certain DNA sequences unit after unit. Imagine if there is a target gene between the first frt and loxP in the initial phase, it would pass on and on in the continuous units after flipping.
Construction
Due to time limits, we focused on constructing circuit 2, our original design.
We added florescent protein eCFP（标上砖号） and mCherry（标上砖号） right in the downstream of the ssrA-tag of recombinase flpe and cre, respectively. And we add a final GFP（标上砖号） as a reporter. pSB1C3 was used as vector for cloning and we tried to transfer the entire circuit to pSB3K3 in order to test its viability (fig. 4) .
In terms of ligation efficiency, we resembles small fragments (promoters, FRTs and LoxPs) and deliver them to IDT for complete synthesis. Then we ligate long fragments in between according to our design (fig. 5) . Worth of attention, we created a new "2A" assembly by using DNA clean up and arranging gene segments with different resistances (fig. 6).
Testing
Experimental proof must be accomplished after construction.
2 objectives for circuit 2 must be achieved during testing. One is that we must prove it actually has capability to reverse. The other one is that we must test its efficiency and explore how close it is to an ideal biological device that can really calculate time.
For the first goal, we thought that it could be solved by digestion (fig. 7) . When sequences flip, certain restricted enzyme cutting sites won’t change. What have changed are their locations. Gene length can be altered when sequences reverse, which can be visualized in an elecrophoretic way, as shown in fig. 7.
For the second goal, we used qPCR to verify feasibility of the circuit. We design primers shown in the picture (fig. 8) . We can tell from amplication curves (fig. 9) whether it reverses and calculates time when phases have altered.
Yeast Timer
(Our project, micro-timer, is to construct a counter on DNA that can imprint time on microbes.)
Micro timer 2.0 (Eu-timer) is constructed by DNA-based counting motifs that are inserted into different sites of chromosomes, creating a relatively large-scale system with more motifs than that in Micro timer 1.0.
The Eu-timer uses recombinases from Ser family such as Bxb1, which typically catalyzes site-specific recombination between an attachment site on the infecting phage chromosome (attP) and an attachment site in the host chromosome (attB) in natural system. The resulting integration reaction inserts the phage genome into the host chromosome bracketed by newly formed attL and attR (LR) sites. When attB and attP are engineered to be opposite BP sites, the integrase alone catalyzes the inversion of sequences flanked by BP sites, changing BP sites into LR sites, and will not revert the DNA flanked by LR sites.
In the design of Eu-timer (Fig 1) , each inverted promoter flanked by BP sites is downstream of an inverted reporter gene and followed by a ser integrase gene. The reporter i(inverted reporter gene)-attP-promoter i-attB-integrase unit is defined as a counting motif, named eu-timer integrase motif (EIM).
The circuit can be programmed to record time by counting a specific type of events like the expression of cyclins. Once the motifs are activated, the downstream expression can work automatically and will not be terminated or reset by the hosts themselves, which is the reason why we believe that such a system can imprint “the same time” on microbes derived from a single clone.
We then designed an telomere-like device by making little changes upon Micro timer 1.0, named Micro timer 1.1 (Fig 2) , in which the flanking site are of same direction. With every cell division, this device will sequentially truncate a part of the sequence, and finally lead to cell death, working like telomere.
Modelling
First of all, we get two tables of one certain combination, including different kinds of plasmid with certain type of promoter, invert-ase together with its recognition site on the reporter and a ssra tag of a specific intensity which we will calculate later. From the second table in each group, we show the RFU changing with respect to time and obviously it reflects the quantity of the protein. The second derivative of the fitting curve is the enzymatic activity, which is the product of the enzymatic activity of one single enzyme at a given moment and the total quantity of the enzyme. Now let consider it separately. As for the enzymatic activity, we use the Michaelis-Menten equation to describe it.
S in this equation is the quantity of the plasmid to be inverted in the bacterial population per volume . S= S initial value－the first order integral of V0. So the expression of V0 is an ODE model. Another important fact that we need to take into consideration is that since the quantity of the plasmid to be inverted in the bacterial population per volume is limited, the total feedback is on the enzymatic activity. So the ODE model can be only used from the start moment till the first derivative of the curve reaching its maximum. As for the quantity of the enzyme, we can use the cftool in matlab to fit a function to show the quantity with respect to the time using the data in the first table of each group. However, in order to get the total quantity of the enzyme produced in the process we must add the expression leakage of RFU (expression: OD0/OD * G0 . OD0 is the first line of the first table in moment 0. OD is shown in table 3. G0 is also constant. ) to the actual value of the RFU in the first table of each group which supposed to be linear. Up to now, we can get the coefficient Vmax and Km in this function, that is, we get the exact function of the fitting curve of the second table in each group. In order to get the rest part of the function, let’s move forward to the meaning of the first order derivative of the fitting curve = quantity of inverted promoters equal (=S) * promoter intensity – the degeneration rate of the protein with the ssra tag – the nutrition correction equation which called Logistic equation. The degeneration rate of the protein with the ssra tag can be obtain in the following method. We use the principal component analysis on each four difference between row OCG and OCGS, OGC and OGCS, MCG and MCGS, MGC and MGCS of the same line which means at the same moment, to integrate one data for each line. Then we use the data at moment 0,1,2 and so on to construct the function of the degeneration rate of the protein caused by the ssra tag using the cftool in matlab.
Now we get the complete function of the second table and the constant Vmax and Km, which is the property of the certain kind of enzyme we use in each group.
The abbreviation of the term in each row.
Now we use the data of MCG group as an example, the fitting curve of the second table is the red one, the blue one is the scatter diagram linked together with line showing quantity of the enzyme with respect to the time and since we know the form of the function is linear, we can easily derivate the exact function by fitting.
