Our nitrate sensing device was originally designed to have three parts:
- A part with a constitutive promoter that expresses NsrR, a nitrate-dependent transcriptional repressor.
- A part that expresses mScarlet, a lambda repressor c1, and Lon which facilitates the degradation of proteins tagged with ssrA. The promoter used for this part is PyeaR, which can be repressed by the binding of NsrR to the promoter.
- A part that expresses mVenus-ssrA. The promoter used for this part is promoter PL, which can be repressed by the lambda repressor c1.
Before adding all three elements for the experiment, we first mainly focused on the first two parts, NsrR and mScarlet. When NsrR is present at high concentrations, a large amount of nitrate is required to activate the reporter system. This means that the concentration of NsrR in the cell determines the threshold of nitrate at which the reporter system can be triggered. In order to maintain different intracellular concentrations of NsrR, we used the following promoters that have been used in previous iGEMs as constitutive promoters for the expression of NsrR: P43, the Bacillus promoter, PampR, the promoter of the beta-lactamase gene, and PlacI, the promoter of the lac repressor (lacI) and its variants.
Table 1
A total of six constitutive expression units were tested with the combination of each promoter and ribosome binding site (RBS). Briefly, the plasmid containing mScarlet-I3 transcriptionally fused to each promoter was transformed into DH5a-pir and fluorescence images were acquired using a fluorescence microscope equipped with a CCD camera (Figure 1A). The RGB channel was separated using imageJ, in which pixel intensity was calculated from the red channel. Using these methods, we compared the strength of the expression units for PlacIq1, P43-creRBS, PampR-RBS, PampR, Placiq1s, and PlacI, as shown in Table 1 and Figure 1.
Figure 1
Next, we constructed plasmids to express NsrR under these promoters and observed how the transcriptional activity of PyeaR depends on the degree of intracellular expression of NsrR by nitrate treatment. We used sodium nitroprusside (SNP) as a nitrate donor. Single colonies obtained after transformation of each plasmid were inoculated, and the well-grown transformed E. coli were transferred to fresh media and incubated for 2 hours at 600 rpm at 35 degrees Celsius using a plate shaking incubator. Next, the SNPs were treated with the concentrations mentioned in Fiqure 1 for 6 hours with the same incubation conditions as above. Finally, the OD600 and fluorescence intensity of mScarlet-I3 were measured using a micro plate reader (Table 2).
Table 2
For PlacI and PlacIq1s with weak promoter strength, high red fluorescence values were measured even without SNP treatment. This suggests that the activity of PyeaR was not adequately controlled due to the low concentration of NsrR in the cell. On the other hand, in the case of P43-creRBS, which used a stronger promoter, the basal level of red fluorescence remained low. Interestingly, however, P43-creRBS system showed reduced sensitivity to nitrate, indicating that a high threshold concentration of nitrate is required to activate the nitrate sensor. For the remaining three expression units (PampR, PampR-RBS, and Placq1), a pattern of low basal levels and a linear increase in fluorescence values with increasing concentration was observed (Figure 2a). Among the three expression units, we selected PampR-RBS as the final candidate, which showed the most linear increase in fluorescence with nitrate concentration among the three, also showed a low basal level expression, and finally showed the second highest fold change compared to the negative control (SNP 0 uM) (Figure 2b).
Figure 2
To further improve the PampR-RBS based nitrate sensor, we tried replacing the RBS with a stronger RBS, which can improve the translation rate of mScarlet-I3. In addition, we also tested the P43-creRBS-based device, which showed the smallest fold change in nitrate-induced fluorescence in the previous experiment, to see if RBS replacement correlatedwith an increase in fluorescence values.
Table 3
As shown in Table 3, although replacement with stronger RBS increased the basal fluorescence level, the P43-based sensor showed a signal increase of up to 250% compared to the original sensor without the RBS replacement, and a larger fold change was also observed compared to the negative control. PampR-RBS based sensor showed up to 200% increase in fluorescence value compared to the previous version. Fold change (F0/F0) increased from 9 to 12-fold (Figure 3).
Figure 3
To determine the range in which our nitrate sensing device can detect NaNO3 and SNPs, we conducted experiments with more refined nitrate concentrations. First, we divided NaNO3 into 12 concentrations ranging from 0 mM to 100 mM and SNP into 12 concentrations ranging from 0 uM to 2 mM. The results showed that NaNO3 could be detected from 2 mM to 10 mM, while the increase from 0.125 mM to less than 2 mM was small but sufficient to detect nitrate. SNPs were also linear from 50 uM to 1 mM, showing the best performance in this range, with sufficient performance to detect from 1 uM to 50 uM (Figure 4b).
Figure 4
Although Luria Bertani media is a rich media, it was observed that treatment with SNP or NaNO3 reduced the growth of E. coli even at low concentrations. To determine if better nutrition could accelerate E. coli growth and improve its sensitivity to nitrate, we tested our nitrate sensing device in brain heart infusion (BHI) media, which is more nutrient-rich. As we expected, E. coli growth was improved by about 10%, with some variation depending on the nitrate concentration (Figure 5a). We checked to see if the increased resistance to nitrate stress translated into absolute fluorescence values. Indeed, we observed an increase in fluorescence intensity that was approximately three times greater in BHI media than in LB (Figure 5b).
Figure 5
Next, we constructed three plasmid DNAs as shown in the figure below by fusing a constitutive promoter to show in green the safety against nitrate under normal conditions without nitrate, a repressor binding element so that transcription can be repressed upon increasing nitrate, and an lva tag recognizable by Mf-lon protease in the GFP coding sequence to accelerate the turnover rate of GFP protein. We checked whether GFP was sufficiently expressed under repressor-free conditions. As shown in Figure 6, all of them expressed GFP, but PJEx-D-vsfGFP-0_lva showed the highest fluorescence.
Figure 6
Based on these results, our team developed Monitro, a nitrate sensor consisting of two plasmids in E. coli (Figure 7).
Figure 7
Finally, we checked whether Monitro, which consists of the two plasmids we made earlier, turns green under normal conditions and changes from green to red in the presence of nitrate. The nitrate donor was SNP, and the cells were harvested 30 minutes, 1 hour, 2 hours, 4 hours, and 6 hours after treatment to measure the intensity of fluorescence. The results showed that the red fluorescence started to be higher than the green fluorescence from 3 hours after SNP treatment (Figure 8).
Figure 8
To apply our team's Monitro in the real world, we next fabricated calcium alginate beads so that we can encapsulate our nitrogen-sensing bacteria in alginate capsules for easier use and distribution.
Preparation of Beads
Fig.9. Calcium Alginate Beads
The first two things to consider when constructing the alginate beads are the size of the beads and the number of bacteria in each capsule. We determined the sizes of the beads to be 4.02~4.18mm in diameter (Figure 9 - A) for efficient fluorescence detection by RGB color sensor. Such size also allowed us to keep the bead in a spherical rather than in a lanceoloid form (Figure 9 - B) which occurs when the bead size falls below our mentioned dimensions. For each of the 4.02~4.18mm beads, we used approximately 20μl of alginate-bacteria mixture, which contains roughly 6.3x108 cells.
Viability of E. Coli in Calcium alginate beads
Fig.10. Viability of E. Coli in Calcium Alginate Beads
Another important point to consider when making a biosensor is how long the cells would remain viable in these alginate capsules. The bacteria’s viability is a critical factor in commercializing our project. When we observed the capsules at 4℃ for 7 days, 87% of all the cells were still viable after 7 days, while approximately 73% of the cells remained viable after 9 days (Figure 10). In our experiments, all beads within 3 days of production didn’t show any functional defects or variability in fluorescence in response to nitrogen.
Leaking of E. Coli in Calcium alginate beads
Fig.11. Leaking of E. Coli in Calcium Alginate Beads
One interesting yet concerning phenomenon we observed while measuring the cells’ viability was that we could see bacteria “leaking” from our beads (Figure 11). We still haven’t fully determined the cause, while it is possible that those were simply bacteria that attached to the surface of the beads during the production process. Although the amount of “leaking” bacteria itself wasn’t too concerning, we felt the need to address this problem due to potential biosafety issues. To this end, we changed our production method so that we further encapsulate our alginate bead within another, bigger alginate bead to minimize diffusion out of the beads.
Fig.12. Double-layer Calcium Alginate Beads
Our new “capsule-within-capsule” beads consist of 2 layers of alginate, and none of our double-layer beads produced in a sterile environment showed signs of bacteria leakage (Figure 12). It is important to note that constructing a double layer in a non-sterile environment does not prevent bacteria leakage.
Change in fluorescence over time
Fig.13. Change in Fluorescence Over Time
Now that we’ve spent a lot of effort to produce our capsules, we have to perform functional analysis to see if our hard-earned product actually works like a color biosensor. Although we were able to actually see the beads turning red from green, we used UV light to further amplify the signal and used an RGB color sensor to quantitatively measure the intensity values. Then, an important question arises: “How long after the addition of nitrate do we measure fluorescence?” In the case of mScarlet, the red fluorescence intensity seemed to saturate after some point (~5 hours), perhaps even starting to decrease after a long time (>6 hours), possibly due to repeated UV radiation while measuring the fluorescence intensity at each time point (Figure 13).
On the other hand, green fluorescence coming from mVenus starts to steadily decrease until 5~6 hours post-nitrate addition, while it saturates after 6 hours, similar to mScarlet. Therefore, we determined that some time between the 4~6 hours post-nitrate addition would be the most ideal time point to measure both mScarlet and mVenus fluorescence.
Color change of Biosensor
Fig.14. Color Change of Biosensor Beads
As we saw in Figure 13, the mVenus level is dominant before nitrate is added, making the capsules glow green (Figure 14). However, the addition of nitrate causes the capsules to slowly turn red over time. Since quantitative measurement of the fluorescence using UV radiation causes viability decrease, we qualitatively observed how the color changes over time post-nitrate addition. In a 10mM nitrate solution, the color of the beads changed from green to notably red after 6 hours.
Fig.15. Color Change of Beads in Different Concentrations of Nitrate (2mM, 4mM, 6mM, 8mM)
We also tried varying the concentration of nitrate solution to find out the working range of our sensor. In a variety of nitrate concentrations (2, 4, 6, and 8mM), the same noticeable green-to-red change occurred even after 4 hours (Figure 15). Figure 16 below shows the 4-hour time-lapse image of the alginate bead biosensor changing its color from green to red.
Fig.16. Color Change of Beads Responding to Nitrate
Conversion of RGB value to Nitrate concentration
Fig.17. Relation Between RGB Value and Nitrate Concentration
The most essential step is converting the measured RGB values to nitrate concentration via the hardware. However, due to the nature of biosensors, the results may vary depending on the state of E. coli and the reaction time. To address this disparity, we conducted an additional experiment, obtaining ranges for individual values of R, G, and B. First, the nitrate concentration was divided into 8 equal parts from 0 mM to 10 mM, and the RGB values were measured after 6 hours. The results showed that the value of red ranges from 65-189, green from 104-213, and blue from 176-294 (Figure 17). Since the fluorescence intensity was measured using ultraviolet light, the overall level of blue was high, and the level of green was slightly higher than that of red. In the end, the red value was selected among other values to represent the nitrate concentration because the red value demonstrated an increasing trend along with the contact time with nitrate, and there was no known interference of UV on this value, making it more accurate than green.