Team:Hong Kong HKUST/riboregulator/characterization

Riboregulator is a type regulatory RNA that can regulate translation. One component of the riboregulator system, cis-repressing RNA (crRNA). CrRNA contains a cis-repressing sequence which is located 5’ of the RBS and the gene of interest. When the transcript is formed, the cis-repressing sequence can form a loop to form a complementary base pairs with the RBS and blocking the ribosome entry to RBS. CrRNA is commonly called “lock” because it “locks” the translation of proteins. When there is a lock, we need a “key”. Component of the system that act as a key is the taRNA. It can interact (in trans) with the cis-repressing sequence to unlock the RBS and therefore activate translation. The HKUST iGEM 2014 team characterized 4 riboregulator already available in the Part Registry and 1 riboregulator introduced by our team.
Table 1 List of riboregulator pairs characterized by HKUST iGEM 2014 team:

Name and registry code	Group	Cognate pair
Lock 1 BBa_J01010 Key 1BBa_J01008	iGEM 2005_Berkeley (Golden Bear)	Yes
Lock 3 BBa_J01080 Key3 BBa_J01086	iGEM 2005_Berkeley (Golden Bear)	Yes
Medium lock BBa_K175031 Key for medium lock BBa_K175032	iGEM09_TUDelft	Yes
Lock 1 BBa_J01010 Key 1BBa_J01008	iGEM 2005_Berkeley (Golden Bear)	Yes
Lock 1 BBa_J01010 Key 1BBa_J01008	iGEM 2005_Berkeley (Golden Bear)	Yes
Lock 1 BBa_J01010 Key 1 BBa_J01008	iGEM 2005_Berkeley (Golden Bear)	Yes

Riboregulators have cognate pairs. For certain crRNA, there is a corresponding taRNA that can activate “unlock” the repression by crRNA. We originally thought that Lock 3c and Key 3c (Table 1.) were cognate pairs, but they turned out to be that iGEM 2006_Berekley simply made different variants of Lock 3 and Key 3. They gave put an alphabet at the end of the name every time they produced different variant of lock 3 and key 3. The lock 3 and key 3 variants were created independently from each other so the letters at the end of name does not mean correspondence. Other teams should take note of this when they consider using riboregulators variants from iGEM 2006_Berkeley.

P_BAD promoter is an arabinose inducible promoter. In nature, the promoter exist in the arabinose operon to regulate the transcription of araB, araA, and araD. The arabinose operon or the ara operon encode enzymes needed or the catabolism of arabinose to xylulose 5- phosphate which is an intermediate of the pentose phosphate pathway. The Pc promoter which is adjacent to the P_BAD promoter transcribes the araC gene in the opposite direction. AraC protein is responsible to repress the activity of the P_BAD promoter when arabinose is absent. Once arabinose is present,the AraC protein binds to the arabinose and dimerize. The dimerize form of AraC-arabinose can activate the P_BAD promoter (Schleif, 2010).

There are several P_BAD promoters in the Part Registry. The promoter that we were interested was BBa_I0500 because of two reasons. First, BBa_I0500, along with P_BAD, has araC gene regulated by the Pc promoter. Without the AraC, the repression and induction of P_BAD can only work on strain that are AraC+. By coupling the araC gene with the P_BAD promoter, we can be free from such restraints. Second, BBa_I0500 needed debugging. BBa_I0500, although it is useful, it is not requestable because of inconsistency in sequencing. Also, in the experience page, two teams, Groningen 2011 and Cambridge 2011 had some discrepancy between how the promoter responded to the arabinose induction. In brief, Groningen results show that the induction of the promoter by arabinose was gradual while Cambridge results show that it was an “on-or-off” response. We wanted to analyze these problems so that the part could be more reliable for other users. Cambridge cited a paper that mentioned that variation in response could have resulted from cell strain variation Khelbnikov, 2001)

All-or-none response observed for individual cells.
Flow cytometry can measure the fluorescence of individual cells. The individual measurement of cell fluorescence can be plotted in a forward scatter (FSC) intensity value versus fluorescence graph. An arbitrary vertical line divided the region of the graph with low fluorescence (Q3) and high fluorescence (Q4).

For the negative control, DH10B transformed with pSB3K3-BBa_E0240, we see cells distributed in Q3 while absent in Q4. Converse was true for the positive control, DH10B transformed with pSB3K3-BBa_I20260 (GFP generator regulated by BBa_J23101). We see most cells distributed in Q4. For our experimental sample, (pSB3K3-BBa_I0500-BBa_E0240) similar trend could observed for three strains. At no arabinose induction, we see most of the cells in Q3. The distribution of cells shifted to Q4 when arabinose was added to the medium. We also see that the cells remained in Q3 even after arabinose induction. The bimodal distribution indicate that some cells got induced while other cells were in repressed state.

Figure 1. Forward scatter intensity (FSC) versus GFP graphs for samples with P_BAD promoter regulating GFP generator.

All samples were inoculated in M9 minimal salt medium overnight in various arabinose concentrations (%w/v). The samples were diluted around 10 fold the next day. Sample were fixed and the fluorescence was measured using flow cytometer. The graphs were plotted for the control constructs, pSB3K3-BBa_E0240 (-) and pSB3K3-BBa_I20260 (BBa_J23101) in the absence of arabinose. FSC versus GFP graphs for pSB3K3-BBa_I0500-BBa_E0240(BBa_I0500) in 0, 0.2 and 1.0% arabinose concentration were plotted. Each set of graphs were obtained for three different cell strains, DH10B, DH5α and BW25113.

The distribution can also give us some idea of P_BAD promoter leakage in different cells strain. If the data points are on or near the boundary between Q3 and Q4, we know that the promoter is leaky. The results indicate that the promoter is relatively more leaky in DH10B compared to other cell strains.

The percentages of cells in each Q3 and Q4 are highlighted in Figure 2. After addition of arabinose, we clearly see more percentage in Q4. Around 80% of the cells for DH10B and BW25113 are in Q4. Relatively lower percentage of cells in Q4 is observed for DH5α. Only around 60 to 70% of cells were in Q4 after arabinose induction. The shift in the distribution from Q3 to Q4 also corresponded to the increase of RPU across different arabinose concentration. At 0% arabinose concentration, for all three strain, higher percentage of cells are in Q3 and therefore the RPU is low. After arabinose induction, more cells are in Q4 and RPU is high. The percentage of cells in Q4 levels off for arabinose concentration that we have tested. The trend was observed for RPU measurement.

We can also conclude that even after maximum period of induction (overnight induction) of arabinose, we still see some cells uninduced (in Q3).

Figure 2. The percentage of cells in induced and uninduced state, and RPU across different arabinose concentration.

Q3 and Q4 represent the 3^rd and 4^th quadrants of the forward scatter versus GFP curve mentioned in Figure 2. The experimental condition was same as the procedure mentioned in the caption of the Figure 1. The left y-axis is for the percent of cells in Q3 and Q4 while the right y-axis is for RPU. Graphs depict the triplicate mean ± standard deviation. (A) Graph for pSB3K3-BBa_I0500-BBa_E0240 in DH10B. (B) Graph for pSBK3K3-BBa_I0500-BBa_E0240 in DH5α. (C) Graph for pSB3K3-BBa_I0500-BBa_E0240 in BW25113.

RPU of PBAD in different cell strains
In order to solve the discrepancy between Groningen 2011 and Cambridge 2011, we have calculated the Relative Promoter Unit of the P_BAD promoter in three different cell strains across increasing arabinose concentration. The three strains that have chosen was: DH10B, BW25113, and DH5alpha. Groningen 2011 used cell strain DH5alpha to obtain the 3-D graph while Cambridge 2011 used BW27783. Unfortunately, we did not had access to the strain, so we had to use another strain, BW25113, which is commonly used by synthetic biologists. Only DH5alpha has araC in its genome. All other strains have deletion of araC (and other genes in the L-arabinose operon).

Figure 3. RPU of P_BAD promoter in three different cell strains across different arabinose concentration.

Relative Promoter Unit of P_BAD promoter was calculated in three strains: DH10B, BW25113 and DH5alpha. Gradient arabinose concentration (% w/v) from 0% to 1.0% with 0.2% increments was used to test the variation of promoter strength (RPU) in different concentration of arabinose. Each strain of cells inoculated overnight in various arabinose concentration above. The cells were diluted around 10 fold and grown until they reached mid-log phase (OD600 0.3-0.5). Cells were fixed and fluorescence was measured using flow cytometer. The graph represent triplicate mean ±SD.

We thought that Relative Promoter Unit (RPU) defined by Endy et al. would be a better a measure of promoter strength at different arabinose concentration than simply comparing the fluorescence measurement of different strains. This is because simply measuring and comparing fluorescence as an output can also be affected by other experimental factors such as genetic context. Because RPU is a ratio to fluorescence measurement, the effects caused by these factors can be minimized (See Methods for RPU calculation).

We observed clearly a different promoter response among the three cell strains. The all-or-none response was observed for the strains, but the levels of RPU were different. For DH10B, RPU leveled off around 0.45. For BW25113, RPU leveled off around 0.3. For DH10B and BW25113 we see a very clear “all-or-none” response. The RPU reaches a plateau at 0.2% of arabinose. For DH5α, it is less obvious, because the initial increase of RPU at 0.2% arabinose is around 0.1, lower than the other two strains. With the given graph, statically, we cannot say that the RPU level gradually increases as % arabinose increases.

The lower RPU of DH5α was expected because DH5α has araC gene, gene for the repressor, in its genome. This could have caused the lower promoter strength of P_BAD in DH5α. All in all, all three strains produced all-or-none response.

For DH10B and BW25113 we see a very clear “all-or-none” response. The RPU reaches a plateau at 0.2% of arabinose. For DH5α, it is less obvious, because the initial increase of RPU at 0.2% arabinose is around 0.1, lower than the other two strains. With the given graph, statically, we cannot say that the RPU level gradually increases as % arabinose increases. All in all, all three strains produced all-or-none response.

P_BAD Leakage
Leakage was observed for P_BAD promoter in DH10B. Although the RPU was relatively higher compared to that of other two strains, DH10B showed leakage at no arabinose induction. The RPU was around 0.16 for DH10B which is quite significant if we want to have a non-leaky system. For BW25113 and DH5α the leakages were below RPU of 0.1. The lowest leakage was observed for DH5α. The low leakage in DH5α, once again, could be explained by the additional copy of araC in the genome. The low leakage BW25113, however, is harder to explain. Nonetheless, if we want a low-leakage system that can reach relatively high RPU upon arabinose induction, out of the three strain we have used, BW25113 would be the best option.

3-D graphs for DH10B and DH5α
We also tried to produce a 3-D graph. We, however, had some changes from the genetic context that Groningen 2011 used. At 3’ of the P_BAD promoter, instead of using BBa_E0840 (GFP generator), we used BBa_E0240 (GFP generator). We also used a low copy pSB3K3 plasmid as the backbone instead of the high copy pSB1C3. We also could not take measurements every 15 minutes interval because we had to measure the fluorescence at each time point manually. Fluorescence was instead measured every two hours. Lastly, OD600 was also monitored and graphed (Figure 4. A and B)

Figure 4. Fluorescence and OD600 measurements of DH10B and DH5α induced in different arabinose concentrations.

Triplicate of DH10B and DH5α samples were inoculated in deep well 96 well plate overnight in M9 minimal salt medium. Arabinose was added to match the final working concentration from 0 to 1.0 % (w/v) with 0.2% increments. Fluorescence and OD600 was measured every two hours for ten hours. (A) Increase of OD600 measurement for DH10B strain in different arabinose concentration. The graph represents triplicate mean ± SD (B) Increase of OD600 measurement for DH5α strain in different arabinose concentration. (C) Fluorescence VS arabinose concentration VS Time 3-D graph for DH10B. Each point represent triplicate mean. (D) Fluorescence VS arabinose concentration VS Time 3-D graph for DH5α. Each point represent triplicate mean.

The OD600 value reflects cell concentration. OD600 for both strain in different arabinose concentration increased exponentially. Growth rate for cells without arabinose induction (0%) was greatest for both DH10B and DH5α. For other concentration of arabinose induction the growth rates were similar for both of the cell strains (Figure 4. A and B). The similar growth rate across the 10 hour period indicates that the cell concentration for samples in different arabinose concentration increased similarly and therefore the cell concentration at each point of measurement was similar across different arabinose concentration. We can therefore assume that differences in the cell concentrations have minimal effect on the different fluorescence levels at different arabinose concentration.

Unlike what Groningen 2011 observed in their experiment, (Experience page.) we observed all-or-none response for both of the cell strains. For DH10B and DH5α the fluorescence levelled off at 0.2% arabinose concentration. In the process of subtracting autofluorescence, we got some negative corrected fluorescence. These region correspond to the blue areas of the 3-D graphs.

Groningen 2011 results may not truly represent the gradual induction of P_BAD promoter

We believe that it is actually difficult to analyze the promoter’s gradual or all-or-none response looking the 3-dimensional graph that Groningen 2011 team presented. The 3 dimensional graph has three parameters: time, various arabinose concentrations and fluorescence. The graph does not consider the OD600 value which can represent the concentration of cells. If the growth rate of the cells are different in different arabinose concentration, the final concentration of cells at given point of time can vary. Groningen 2011 team did not provide with OD600 versus time graphs. Therefore it is hard to tell well there was variation in the cell growth under the context of Groningen 2011’s experiment. Because the fluorimetry measurements, the method that Groningen 2011 used, measures the fluorescence of the entire population of cells, the fluorescence can be affected by the concentration of cells and hence show a response that is not all-or-none. Looking at Groningen’s results, it would be more appropriate to say that the population of cells that is transformed with plasmid containing BBa_K607036(BBa_I0500 -BBa_E0840 ) showed gradual response to increasing arabinose concentration.

Different genetic context could be responsible for the different response of the P_BAD promoter.

Due to lack of human-power, unfortunately, we could not produce 3-D graphs for cells with pSB1C3 backbone as controls. Assuming that Groningen 2011 result was valid, we can speculate that difference of the P_BAD promoter response could have resulted from the different backbone that we have used to characterize the promoter. In fact, Cambridge 2011, who also observed all-or-none response (Experience page.) used pSB3K3 as backbone for their constructs. Without proper controls, it is hard to conclude

Construction

1. Construct pSB3K3-BBa_I0500-BBa_E0240

2. Transforming pSB3K3-BBa_I0500-BBa_E0240 to DH10B, DH5alpha, and BW25113.

3. Transforming pSB3K3-BBa_I20260 (Standard Constitutive Promoter/Reference Promoter) from the 2014 Distribution Kit to DH10B, DH5alpha, and BW25113.

4. Transforming pSB3K3BBa_E0240 (GFP generator) from the 2014 Distribution Kit to DH10B, DH5alpha, and BW25113.

Measurement

1. Preparing supplemented M9 medium
(M9 Minimal salt medium protocols could be seen on the Protocols page, or download the PDF file)

2. Culturing E. coli DH10B strain carrying the whole construct listed on the construction part. Grow cell culture overnight (Incubate 37°C and shake for 15 hours) with M9 minimal medium (we used Corning® 96 well storage system storage block, 2 mL, V-bottom, sterile to culture the cells, and Corning® microplate sealing tape white Rayon (with acrylic), sterile, suitable for cell/tissue culture applications, breathable sterile membrane.)

3. Take out 20-30μl of overnight cell culture (we used Multichannel Pipetman) and mix it with M9 medium and arabinose with specific concentration (0%, 0.2%, 0.4%, 0.6%, 0.8%, 1%) in the 96 Deep Well plate.

4. Incubate in 37°C and shake for 3 - 4 hours.

5. Take out 200ul of cells from the 96 deep well plates, and put it on a micro test plate 96 well flat bottom. (we used Micro test plate 96 well flat bottom, made by SARSTEDT.)

6. Measuring the GFP intensity and OD595 values (we used Envision Multilabel Reader) every 30 minutes after the above mentioned E. coli strains are cultured to mid-log phase (OD600 = 0.3 - 0.5)

Filter used on Envision Multilabel Reader:
- Absorbance :Photometric 595nm,
- Excitation :485nm FITC,
- Emission :535nm FITC,
- Mirror module : FITC (403) at bottom.

- In between measurements, keep incubating the cells in 37°C while shaking.

OR

- When cells were in the mid-log phase, cells were fixed and fluorescence was measured using FACS.

7. Calculating the Relative Promoter Units (RPU) using the obtained data;

Data Processing for data from Envision Multilabel Reader

1. After E. coli carrying the right construct was grown to mid-log phase, GFP intensity and OD595 were measured every 30 minutes (up to 120min);

2. GFP intensity are subtracted with the background fluorescence which is the fluorescence of pSB3K3-BBa_E0240. Curve reflecting GFP expression change was plotted (from 4 measurements from time=0 to time=120); OD595 was converted to OD600, and average values were taken;

3. GFP synthesis rate was then obtained by calculating the slope of the above mentioned curve;

4. Absolute promoter activity of P_BAD and BBa_I20260 were calculated by dividing the GFP synthesis rate with the average OD600 value;

5. Averaged absolute promoter activity was then obtained by averaging the respective 3 sets of absolute promoter activity values;

6. Finally, R.P.U was calculated by dividing the averaged P_BAD and absolute promoter activity over the averaged BBa_J23101 absolute promoter activity. R.P.U value of P_BAD in different concentration of arabinose is shown. Leakage could be analyzed according to the R.P.U value that shows the GFP expression of P_BAD promoter in the absence of arabinose (0%).

Equation of the RPU calculation is shown below:

Based on this equation, the GFP synthesis rate (dF/dt) for the experiment (BBa_I0500-BBa_E0240) divided by the ABS (which means the absorbance or OD600) of the cells containing the experiment construct. GFP synthesis rate (dF/dt) for the reference promoter (BBa_J23101-BBa_E0240) divided by the ABS (which means the absorbance or OD600) of the cells containing the reference promoter construct.

Then, RPU is obtained by dividing the GFP synthesis rate/ABS of experiment with the GFP synthesis rate/ABS of the reference promoter.

Data Processing for data from FACS

1. RPU was calculated by first subtracting the autofluorescence (fluorescence of cells with pSB3K3-BBa_E0240 )

2. Dividing the fluorescence of cells containing pSB3K3-BBa_I0500 -BBa_E0240 with cells containing pSB3K3-BBa_I20260 .

Equation is shown below:

Based on this equation, the amount of fluorescence [F] for the experiment (BBa_I0500-BBa_E0240) multiplied by the growth rate (μ) of the cells containing the experiment construct. Amount of fluorescence [F] for the reference promoter (BBa_J23101-BBa_E0240) multiplied by the growth rate (μ) of the cells containing the reference promoter construct.

Then, RPU is obtained by dividing the fluorescence*growth rate of experiment with the fluorescence*growth rate of the reference promoter.

3. For simplicity, we assumed the growth rate of the cells transformed with pSB3K3-BBa_I0500 -BBa_E0240 and cells transformed with pSB3K3-BBa_I20260 had same similar growth rate.