Results

Summary

Our summer project was designed to improve the process of nitrogen removal in the tertiary sewage treatment stage of raw sewage processing. Currently, a fairly common design is to spatially separate the processes of nitrification and denitrification so that distinct bacterial populations within activated sludge can perform their respective biochemical processes independently. Our project proposes that both processes be temporally separated and housed under a single organism. To achieve this, we attempted to transform the genes encoding for proteins responsible for nitrification into Paracoccus denitrificans, a known denitrifier in active sludge.

By the end of summer experimentation, we were able to transform most, but not all of the nitrification genes into Pc. denitrificans. Our central composite circuit consists of genes for HaoA and HaoB, proteins that catalyze the oxidation of hydroxylamine into nitrite ions. It also includes CycA and CycX, genes encoding for two c-type cytochromes involved in shuttling electrons from the ammonia oxidation reactions to terminal electron acceptors. We were, however, unable to create the full AMO circuit from the G-Blocks for AmoA, B, and C into Pc. denitrificans. We experienced a lot of difficulty in creating circuits through Gibson Assembly. These genes would have encoded for a heterotrimeric protein complex that would have catalyzed the oxidation of ammonia into hydroxylamine.

In addition to the ammonia oxidation circuit we developed, we also designed a series of reporter circuits to test the function of several promoters in P. denitrificans compared to E. coli. We were able to transform Pc. denitrificans with a composite part consisting of Mn-SOD, an oxygen-inducible promoter, and GFP. With this reporter circuit, we attempted to characterize promoter activity in response to different dissolved oxygen levels by measuring relative fluorescence at the absorbance spectra of GFP compared to control untransformed cells.

In addition to Mn-SOD, a new promoter added to iGEM, we also tested the promoter activities of VhB and T7 in transformed E. coli cells. These promoters were connected to the coding sequences for GFP in a composite part. Again, all promoters were tested at various dissolved oxygen levels. These were tested against untransformed E. coli cells. Finally, we accounted for the fact that different oxygen levels may lead to different growth rates, so we divided relative fluorescence units (RFU) values by the OD600 of a given sample that was measured.

Characterization of Chassis

There were three attempts to generate a well-characterized growth curve for P. denitrificans under our laboratory conditions. The first resulted in a culture reaching a growth plateau at an OD600 of around 0.4. The growth plateau was reached between five and six hours after inoculation with a preculture of Paracoccus cells into a 1 L liquid culture in PD media (see protocol/link). Growth was completely plateaued by ten hours. The error bars shown are 95% confidence intervals. The OD600 values of this round of measurement was done with a NanoDrop machine that was potentially not capable of accurately determining the cell density.

The second attempt yielded extremely different results. Again, the culture was grown in a shaking incubator at 33.0 C at 220 rpm. The same PD media was used to grow the culture. The plateau was at an OD600 between 2 and 2.5, Exponential growth ended at 10 hours just as it did in the first trial. Mid-exponential growth was reached at four to five hours just as in the first trial as well. This round of measurement was done with Spectronic Genesys 2, which detected the OD600 of the cell culture more precisely than NanoDrop did.

The last attempt was unsuccessful. The culture reached a plateau at an OD600 of ~0.15 before gradually falling to 0.13. This failure was indicative of the challenge in maintaining a stable P. denitrificans culture.

As a conclusion, the P.D. incubated in 1L of P.D. medium would have exponential growth in the during the first 10 hours of incubation and would then reach the carrying capacity after 20 hours. The OD600 value went up to around 1.8~2.0 at the end of the experiment after the growth curve plateued. For the purpose of P.D. competent cell preparation, 1L of P.D. medium culture will reach the ideal OD600=0.6~0.8 after approximately 5 hours of incubation.

Mn-Sod

While Mn-SOD has been described in literature as an oxygen-inducible promoter element in E. coli, the results of experimentation described under protocols (citation, link) have been inconclusive. Following the previously described protocol, bacterial cell cultures were grown from single colonies for four hours to reach early- to mid-exponential growth phase (shown to be ~4 hours in the literature and our own results). After four hours, colonies were transferred into 14 mL tubes. Because equipment to modulate the dissolved oxygen environment was not available, tubes were covered after a specified amount of time and kept closed until the end of experimentation to simulate differential oxygen levels. Bacteria in tubes that are covered for longer periods of time will consume the available oxygen over time, resulting in lower dissolved oxygen levels. After the experimental window closes, all relevant parameters were measured (DO, OD600, and RFU output from the optical plate reader) simultaneously.

The first trial was conducted on 9/25/2017 and cultures were grown for a total of four hours before being transferred into 14 mL falcon tubes. At four hours, tubes were opened and tested for DO with a Vernier Instruments DO probe and for OD600 with a Genesys 2 spectrophotometer in the order from most hours spent open to least. The reasoning behind this is that the measurement process was time-consuming and exposing cultures that had spent more time open to atmospheric oxygen would minimize the bias caused by re-exposure, as they would be less sensitive than cultures that had more time in an anaerobic environment to changes caused by unsealing tubes. Immediately after all tubes were opened and samples were taken from each strain and time point, the well plate was sealed with an an optical adhesive cover and measured by a BioRad MyiQ optical plate reader.

When plotted against hours spent open and exposed to open atmosphere, a steady increase in the relative fluorescence can be observed. When compared against the DO content (in mg/L), however, the opposite trend is observed. For the relationship between DO and RFU, the coefficient of determination is 0.445, corresponding to a correlation coefficient of -0.667. This indicates that the observed RFU of bacteria containing the Mn-SOD reporter biobrick is negatively correlated with DO with a linear relationship of moderate strength. The relationship between the RFU and time spent exposed to atmospheric oxygen, while positive, was much weaker. The coefficient of determination (R2) was only 0.023, corresponding to a correlation coefficient (R) of 0.152. This is indicative of a weak linear relationship.

Figure 1: OD600 normalized RFU values between all measurement replicates over time and DO saturation/content (mg/L).

The same pattern is also observed in results taken from the second trial on 10/22/17. This time, bacterial cell cultures were grown at 37 C in a shaking incubator (220 rpm) for four hours before being transferred into smaller 3.5 mL culture volumes. All aliquots were grown and a new tube was covered every hour as described above, except cultures were grown for a total of six hours. As in the DH5a strain tested on 9/25/17, the strain transformed with the Mn-SOD reporter plasmid showed increasing fluorescence over time but a decreasing RFU value as the observed DO content increased. The coefficient of determination and correlation coefficient for the relationship between RFU value and time spend open were 0.503 and 0.709 each. This indicates that the relationship is positive and moderately strong. The relationship between DO and RFU values was negative and the coefficient of determination and correlation coefficient were 0.374 and -0.612. This is indicative of a moderately strong negative linear relationship.

Figure 2: Results from the trial performed over 10/22/17

Again, the observation is made that as the duration spent opened increases, the level of relative fluorescence picked up by the optical reader does as well. It is also evident in Figure 2 that as the dissolved oxygen content increases, the relative fluorescence decreases.These results both contradict and support our original hypothesis and indicate that an unmeasured effect is causing the DO content to decrease in cultures that spend more time exposed to atmospheric oxygen.

Figure 3: Aggregated data for Mn-SOD in Pc. denitrificans across both trials

Combining both trials yields stronger evidence for the trends described above, as a positive linear trend is also observed when plotting relative fluorescence against time spent exposed to atmospheric oxygen. In addition, the plot of RFU versus DO shows weaker linear relationship with coefficient of determination and correlation coefficient of only 0.085 and 0.292 when both trials are aggregated into a single dataset. This is weaker than the observed relationship between RFU value and time spent open, which had corresponding R2 and R values of 0.341 and 0.584. This indicates a moderate linear relationship between RFU value and time spent open.

In addition, Mn-SOD was tested in Pc. denitrificans and the results corroborated claims in the literature and results found by experimentation. The R and R2 values for the relationship between RFU value and time spent open were 0.198 and 0.445. The relationship between DO content and RFU value had R and R2 values of 0 and 0. These results were mildly positive and indicate to us that Mn-SOD generally behaves the same way in Pc. denitrificans as it is reported to in E. coli.

Figure 4: Scatterplots of relative fluorescence versus time and dissolved oxygen content

VhB

VhB has been described in literature and previous iGEM projects as being maximally active in hypoxic conditions, ideally at 2% dissolved oxygen saturation. When tested with the above describe protocol, results are inconclusive. The first trial was conducted on 9/25/17 and yielded positive data that corroborated previous evidence of maximal induction under hypoxic conditions. As time the tube spent open increased, measurements of RFU decreased, indicating inhibition of GFP expression. This followed a linear relationship with R and R2 values of 0.044 and -0.210. There was almost no observed relationship between relative fluorescence and dissolved oxygen content, which had R and R2 values of 0.022 and 0.148. Both relationships were very weak and indicate little relationship between RFU value and either DO or time.

Figure 5: Scatterplots of relative fluorescence versus time and dissolved oxygen content DH5a E. coli transformed with the VhB reporter biobrick.

When tested on 10/22/17, there is a clear positive relationship between time spent exposed to atmospheric oxygen and the intensity of the relative fluorescence. These data are not only contradictory to the other data collected on 9/25/17, but it is also contradictory to the results found in literature and by other iGEM teams. Because of this, it is reasonable to conclude that there may be a bias in the measurement of fluorescence emitted by cells expressing the VhB reporter biobrick. The relationship between time and RFU value had R and R2 values of 0.369 and 0.607. This is a moderately strong relationship. The relationship between DO and RFU was almost negligible and its R and R2 values were only 0.007 and -0.0837.

Figure 6: Scatterplots of relative fluorescence versus time and dissolved oxygen content DH5a E. coli transformed with the VhB reporter biobrick.

When all data is aggregated between trials, the R and R2 values are much stronger. For the relationship between RFU and time, they are 0.393 and 0.627. For RFU and DO they are 0.236 and 0.486. It must be stressed, however, that these results are contradictory to previous findings and there is strong likelihood of bias in the results. One potential source may be due to the actual versus observed DO level. Volumes used for testing were smaller than the volume recommended for use by the Vernier DO probe. There was also the issue of the time taken to complete all required measurements. The total time between removal of samples from the shaking incubator and the start of the qPCR protocol to measure relative fluorescence was between one and one and a half hours. In that time, gene transcription and translation could have changed significantly.

DH5a Untransformed

In order to compare expression against a control baseline, we also measured the fluorescence of an untransformed culture of DH5a E. coli cells. When a line of best fit was calculated for the data, there was almost no significant relationship between observed relative fluorescence and either the dissolved oxygen content and the time spent open and exposed to the atmosphere. This was most evident when the data from both replications were aggregated. The coefficient of determination for relative fluorescence versus dissolved oxygen content was 0.008. The coefficient of determination for relative fluorescence versus time spent open is 0.041. These values correspond to correlation coefficients of 0.0894 and 0.202 respectively. These are both very weak linear relationships.

Figure 7: Aggregated data from both trials for fluorescence measurements of untransformed DH5a E. coli.

When the data was split by trial run, however, there were also contradictory findings. Data from the first trial conducted on 9/25/17 suggest a positive relationship between relative fluorescence and hours spent open and exposed to atmospheric oxygen, while the relationship between dissolved oxygen and relative fluorescence was close to nonexistent. The coefficients of determination for each were 0.198 and 0.094. These correspond to correlation coefficients of 0.444 and 0.306 respectively. This means the relationship between relative fluorescence and hours spent open to the atmosphere is moderately strong while the relationship between relative fluorescence and dissolved oxygen content is weak.

Figure 8: Scatterplots with trends fit to the data for untransformed DH5a bacteria.

Results from the second trial conducted on 10/22/17 indicate that time spent exposed to atmospheric oxygen has an opposite effect on relative fluorescence. There is a moderate to weak relationship between time and relative fluorescence. The relationship between relative fluorescence and dissolved oxygen content, however, is non-existent just as it is in the first trial. The coefficients of determination were found to be 0.207 and 0.017. These correspond to correlation coefficients of 0.455 and 0.130. This means a weak to moderate negative relationship between dissolved oxygen and relative fluorescence and a weak relationship between relative fluorescence and dissolved oxygen.

Figure 9: Scatterplots with trends fit to the data for untransformed DH5a bacteria.

Testing Device Efficacy by Monitoring Ammonium Removal and Nitrate Production

Results from ammonium flux indicate that the Paracoccus denitrificans (PD) transformed with HAO/cytochrome oxidizes ammonium more effectively than untransformed PD, but less effectively than Nitrosomonas europaea (NE), which is a native nitrifier [NE -2.79 <-1.86 PD (transformed) < -1.33 PD (untransformed)].

Figure 10: Illustrates change in ammonium concentration following spike to 50mM ammonium over a 60 minute interval. Depicts average change for three trials of each bacteria, PD, NE, and PD (transformed). Data supports the hypothesis that the integration of HAO/cyt gene complexes into PD increases its nitrification capacity.

Figure 11: Illustrates change in nitrate concentration following spike to 50mM ammonium over a 60 minute interval. Noticeable increase in nitrate production by transformed PD with HAO/cyt gene complexes indicates that the HAO/cyt oxidize hydroxylamine and nitrite effectively within the Paracoccus chassis.

These results were anticipated due to the kinematic properties of the reaction sequence.The first step in nitrification is oxidation of ammonium to hydroxylamine. While the monooxygenase enzyme is not present in the PD (transformed), Paracoccus does contain limited nitrifying capacity, as outlined in project design. The spike in ammonium, in addition to the high ammonium culturing conditions, is expected to induce this forward oxidation. Once hydroxylamine forms, the Hydroxylamine oxidoreductase (HAO) and associated cytochromes present in the PD (transformed) oxidize hydroxylamine to nitrate.

This brick is maximally induced in the presence of 1mM arabinose, so we expect any hydroxylamine formed to be further oxidized to nitrate. This result is demonstrated in Figure 11. Due to Le Chatelier equilibrium principles, we anticipate that the immediate removal of the hydroxylamine intermediate increases the rate of ammonium oxidation. This prediction is reflected in Figure 10, as the rate of ammonium oxidation is greater in the PD (transformed) than that of the PD (untransformed).

The observed increase in nitrate production observed in Figure 11 for the PD (transformed) indicates that the HAO/cyt BioBrick is functioning in the cell. As compared to PD (untransformed) and NE (untransformed), the PD (transformed) is expected to produce more nitrates because it is being maximally induced in the presence of arabinose.