Results and Demonstration

Project Achievements, Overview:

The goal of our project is to develop a simple but accurate test of bioavailable phosphate in soil. We created three new constructs to measure bioavailable phosphate in soil using the Trp repressor as a means to regulate the expression of Beta-lactamase. The amount of Trp repressor present would be regulated by the phoA promoter which is regulated by the concentration of phosphate present. The phoA promoter is activated when there is less phosphate present and turned off when there are higher amounts of phosphate present. Since this phoA promoter is controlling the expression of trpR, there would be less negative regulator controlling the expression of the bla gene (Beta-lactamase) when there is more phosphate present. When the substrate of Beta-lactamase nitrocefin is present, a red color is produced. Therefore, the ideal result in the construction of our constructs is to cause more red color to be produced when more phosphate is presented to the E. coli containing the constructs and less red color when there is less phosphate. Ultimately producing a quantifiable colorimetric nutrient concentration readout for inorganic/bioavailable phosphate.

---

Initial characterization of parts (Do our constructs work?)

After the constructs were designed, synthesized, and arrived from Twist Biosciences the first experiment to perform was one to determine if they worked at all. This experiment is described on the Experiment page. The appropriate solutions of bioavailable phosphate, Beta-lactamase substrate, and nitrocefin were added to each sample and then waited to see if the solutions turned red.

The picture above shows one of the phosphate-inducible constructs. It was very apparent that Beta-lactamase was being produced in most of our strains, although there was not a noticeable difference in red color with the different phosphate concentrations.

We have performed a few repeats of the experiment and eventually all of our constructs have been showing Beta-lactamase expression, but we haven’t been able to determine a quantitative correlation.

It was apparent from this first experiment that Beta-lactamase was being expressed from our construct but there was little difference in the red color between the tubes with the varying concentrations of phosphate.

---

Data analyzed with UV-Vis Spectrophotometry

Below is a procedure that was developed by Alyssa Ebeling last year. It was repeated for all samples run on the UV-Vis.

1. The absorbances were plotted against the wavelength. The wavelengths were narrowed down to 300 to 700 nm to remove some of the deviation and focus on the target absorbance.
   • The height of the bump from the expected curve was the data point we would like to record as our absorbance.



2. We then removed the wavelengths that made up the target bump.
3. An exponential function was graphed to find the baseline corrected absorbance.



4. We used the peak of the graph of the baseline corrected absorbance as the absorbance for that specific concentration.



5. After all the baseline corrections, all of the absorbances were plotted against concentration on the graph below.

We had expected to see a negative correlation between the absorbance and concentration which is not represented in this data. However, we did expect the negative control to be lower than all the absorbances which are represented in this data. This shows us that while our constructs are working, we still have not been able to make a standard curve.

Light Spectrophotometry

Another set of experiments were performed and Light Spectrophotometry was utilized to analyze the results. The first construct is the phoA promoter controlling the trpR gene which codes for the TrpR protein. The TrpR protein, when it interacts with tryptophan, binds to the trpL operator sequence to turn off the expression of the bla gene. We performed an experiment where we grew E. coli containing this construct and then added varying concentrations of phosphate buffer ranging from 0 to 113 ppm phosphate (Figure 1). This experiment was performed in triplicate and the points on the graph are the averages of the three results.

Unfortunately, the results show an opposite trend than what we expected. We were hoping to see an increase in OD510 (red color) as phosphate concentration increased which would suggest that there is more Beta-lactase present when phosphate concentrations increase. Possible reasons for this discrepancy between our results and our expectations is that there is a high amount of background in the experiments. Even when no phosphate is added, there is a relatively large amount of red color in the samples. This suggests there is Beta-lactamase present in the cells even when no phosphate is added. We performed control experiments to show that when there is no bla gene in the E. coli, there is no red color produced so the red color is dependent on the expression of bla. There is also no bla gene in our E. coli strain except for what is provided in our construct on the plasmid. This tells us that there is expression of the bla gene in our construct independent of the phosphate-inducible expression we are trying to perform. In order to address the issue of high background in the samples we created two other constructs. They are both similar to the first except that the new constructs have a protein degradation tag incorporated on either the trpR gene or the bla gene. The first of these new constructs had an ssrA DNA sequence added to the 3’ end of the trpR gene which should cause the TrpR protein to be degraded more quickly after it is translated. The idea behind this is that background TrpR that may have been produced before an experiment would be degraded more quickly which would allow for a greater sensitivity to new TrpR being produced during the experimental conditions. To test this, we performed an experiment similar to the experiment in Figure 1 to see if this construct with the degradation tag on TrpR would give us results where increased phosphate concentration correlated with an increase in red color.

In Figure 2, we see an increase in OD510 (red color) when phosphate concentration increases which is what was expected. Also, in this graph, we attempted to account for the background red color by subtracting the OD reading for the sample that did not receive any phosphate from all of the other samples. The line produced by the linear regression had an equation of y=0.0006x + 0.0432 which we could use to determine the concentration of phosphate in a dirt sample after exposing the dirt to E. coli containing the construct and nitrocefin. The third construct that we made was the same as the first except that it had the ssrA degradation tag attached to the bla gene. Again, the goal of this was to reduce background color. If Beta-lactamase was being produced in the cells prior to the experiment, the degradation tag could reduce how long the enzyme remained active by degrading it more quickly. When E. coli containing this construct were mixed with the various phosphate concentrations, the overall amount of red color was reduced in these cells but the trend line showed a decrease in red color as phosphate levels increased similar to what was seen with the first construct (Figure 3). This demonstrated that this construct likely was also not going to be useful for measuring bioavailable phosphate concentrations in the soil.

Given these results, we found a construct (Figure 2) that appeared to give us increased red color with increased phosphate concentrations using phosphate buffers prepared in the lab.

---

Soil test

Our next goal was to determine if we could determine phosphate concentrations in soil. We tested two soil samples and two samples that came from biosolids. Biosolids are treated byproducts of waste water treatment plants than can be used as amendments to fields. To test these samples, we weighed two grams of each dried soil or biosolid sample and added it to 10 mM Tris buffer and allowed it to incubate at room temperature for 10 minutes. E. coli containing our functional construct and nitrocefin were then added to each sample and all samples were allowed to incubate for one hour. After one hour, the samples were filtered through 0.45 micron syringe filters and absorbance was measured on a spectrophotometer. Unfortunately, the biosolid samples did not filter as cleanly as the soil samples as there was significant brown coloration or cloudiness after filtration. Thus, we did not try to calculate phosphate concentrations with these samples. The soil samples provided samples that could be measured so we continued calculations with these. While we performed the experiments with soil and biosolid samples, we also performed control experiments with the phosphate buffers again except mostly following the conditions we used to test the soil and biosolid samples. We did not filter the samples since there were no solids. Unfortunately, the graph that resulted (Figure 4) from the control experiments with the phosphate buffers produced a negative correlation similar to the original construct (Figure 1) and the construct with ssrA-tagged bla gene (Figure 3).

It is possible that the greater volumes (roughly 10 times greater) we were using in the experiments involving the soil samples affected the experiment so it would be worth going back and trying the control experiments with all of the constructs in these conditions to see if they all reacted similarly. However, if we took the absorbance values from the soil experiments and calculated phosphate concentrations using the equation determined for the linear regression line in Figure 2, we would get the values we see in Table 1 below.

The values we determined for phosphate concentration using the E. coli test that we developed were very close to the known values determined previously using chemical tests. However, we cannot claim this as completely accurate since we are using the standard curve from one experiment with the results from another. However, if time permitted we would certainly like to repeat the soil and control experiments to see if we could get valid control and sample results that would allow us to make sound scientific conclusions. That being said, it is encouraging that the results in this table are somewhere in the range of previously known results and that they have a similar ratio relative to each other. This suggests that future experiments could demonstrate that our E. coli testing method is a viable way of testing soil for bioavailable phosphate concentrations.

---

Future work

Last year we strived to test multiple nutrient sensors and we had said we would have liked to test more nutrient types. This year we have modified the phosphate contrasts we had made last year and got more promising results. We were able to demonstrate that Beta-lactamase is being produced from the construct when exposed to phosphate. However, more experiments need to be performed to fine-tune the amount of Beta-lactamase being produced to create a standard curve so we can use our sensor to provide an analytical output.