Team:Chalmers-Gothenburg/Results

Page Content Induction intro One Induction System Two Induction Systems Three Induction Systems FAS system

RESULTS

Induction intro

One of the crucial parts of our project in achieving the programmability in fatty acid production, is to be able to control the expression of thioesterases. Our approach to this issue is to construct three different induction systems linked to three different thioesterases. As a confirmation that our induction systems are able to function properly without having to measure fatty acid profiles, we used fluorescent proteins (GFP, RFP, BFP) in place of the thioesterases in parallel. This would provide us with a faster, easier and cheaper way to test the induction system and the fluorescence intensity dependent on added inducer concentrations.

All three fluorescent induction plasmids were transformed into our native yeast strain (CENPK 102-5B) in all combinations, meaning that a total of seven different strains were tested, three containing each individual plasmid, three containing all dual combinations and one strain containing all three plasmids.

All results were obtained from three different experiments using two different machines. The first two experiments were performed with the Guava flow cytometer, a machine that measures samples automatically from a microplate. The Guava machine was unable to measure BFP, so Fluorescence Activated Cell Sorting (FACS) had to be used in this instance instead.

The number of samples that we were able to run using the FACS was limited due to time constraints. Therefore, we had to prioritize the number of data points for the TetON-BFP (while more data is available for the CUP1-RFP and Estra-GFP systems).

Single benchmarking (one induction system)

Single plasmid experiments were conducted, where fluorescence from each individual plasmid was tested with different inducer concentrations.

Single Copper plasmid

RFP expression is dependent on the copper inducer concentration. Figure 1 displays violin plots of the logarithmic values for red fluorescent expression when adding the copper induction agent in concentrations of 1, 10, 50, 100 µM for experiment A and 0, 100, 500 and 1000 µM for experiment B. The right side of figure1 provides control values when different concentrations of inducers are added to wild-type yeast.

There is a clear indication that our induction plasmid provides more red fluorescence than the native yeast (control), meaning that the system would provide an output. However, it is clear that the copper induction system is leaky due to the high expression that both experiment A and B shows for the test without any induction agent, located to the far left. Ideally, there would be no signal from the uninduced sample above the signal from the negative control. Still, there is an increase in expression for the higher concentrations of inducer, the difference being especially clear in experiment B for the 500 µM case.

Figure 1: Violin plots of RFP expression for single plasmid in yeast strain CENPK102-5B, from two experiments, A and B. Measured with Guava. Pairwise statistical tests between controls (0uM added inducer) and remaining samples in each strain were performed using unpaired two-sided Wilcoxon-rank-sum tests (ns: not significant, *p < 0.05, **p<0.01,***p<0.001,****p<0.0001)

To further clarify the differences in fluorescent expression, the RFP intensity was plotted against density of measurement points for different doses in figure 2. These density plots provide a similar result to figure 1 with very similar RFP intensity. However, the density and intensity provided in the two bottom graphs displays more clearly that the concentration 100 µM for experiment A and 500 µM for experiment B has a higher expression. The key aspect is shown by comparing the position of the curve peak, where the more right it is, the stronger the expression.

Figure 2: Overlapping density plots of RFP expression for all concentrations of copper inducer. Measured with Guava.

Single Estradiol plasmid

GFP gene expression was coupled to the estradiol induction system, where the intensities from different inducer concentrations with a single plasmid is provided in figure 3. In this case it is rather clear that for both experiments, the two highest expressions are achieved for concentrations 0.01 and 0.001 µM of estradiol. For experiment B the correlation between increased intensity and increased concentration is also clearly visible. When comparing the control group with the non-induced sample, there is a slight level of expression observed for the non-induced control, for both experiment A and B, indicating that the estadiol system is also leaky, though the difference is still far from any of the induced ones, meaning that the inducer affects the system expression significantly.

Something that is for clear for experiment A in figure 3, is that all induced sample groups are expressing some intensity on the same level as all of the control groups, meaning that they do not express the fluorescent plasmid at all, indicating that the plasmid has dropped out of the cell, which is especially clear for experiment A where there was lag time from adding the cells to the microwells and the actual measurements, described in greater detail in the notebook. However, there are still some similar results for experiment B, indicating that the plasmid itself drops out of the cell more easily.

Figure 3: Violin plot of GFP expression from single induction plasmid in yeast strain (CENPK 10-5B, from two experiments, A and B. Measured with Guava. Pairwise statistical tests between controls (0uM added inducer) and remaining samples in each strain were performed using unpaired two-sided Wilcoxon-rank-sum tests (ns: not significant, *p < 0.05, **p<0.01,***p<0.001,****p<0.0001)

The results provided by the density plots in figure 4 shows the same thing as in figure 3 but in a different way. The peaks for concentrations 0.01 and 0.001 µM in the bottom two graphs display a clearer difference in expression than the other samples, illustrated by a light colour gradient, confirming together with figure 3 that the estradiol system has a high dynamic range with less leakage.

Figure 4: Overlapping density plots of GFP expression for all concentrations of estradiol inducer. Measured with Guava.

Single Tetracycline plasmid

Our tetracycline induction system is connected to the expression of BFP. The measured expression from the single plasmid analysis is given in figure 5 as violin plots. The data shown in this graph point to a great difference in intensity compared to the negative control, making it rather clear that the tetracycline system is leaking, given the high expression with the inducer concentration at 0 µm. It can still be seen that there is a slight increase in expression of BFP when the inducer concentration increases.

Figure 5: Violin plot of BFP expression from single induction plasmid in yeast strain (CENPK 102-5B). Measured with FACS. Pairwise statistical tests between controls (0uM added inducer) and remaining samples in each strain were performed using unpaired two-sided Wilcoxon-rank-sum tests (ns: not significant, *p < 0.05, **p<0.01,***p<0.001,****p<0.0001)

The results from figure 5 show that the induction system is to some extent concentration-dependent, illustrated more clearly with the density plots in figure 6. Both the top and bottom graphs show a slight increase in expression when the concentrations of inducer increase. It is more visible in the bottom graph that, without a complete overlap, the higher concentration of tetracycline, the more to the right the curve is shifted.

Figure 6: Overlapping density plots of BFP expression for all concentrations of estradiol inducer. Measured with Guava.

Conclusion

All three induction systems show an inducer concentration-dependent behaviour. Both the copper system and the tetracycline system show very high leakage levels, but they also show an increase in expression with higher concentrations of inducer. The estradiol system also shows indications of leakage, but compared to the actual output signal when inducers are added, the leakage is relatively small.

Combinatorial benchmarking (two induction systems)

Although the single plasmid results indicate that the induction systems are functional, it does not explore the potential effect they could have on each other. For this reason, we also tested dual plasmid combinations in the same fluorescence manner as the single plasmid systems. The GFP and RFP combination was tested using the Guava flow cytometer while the others with BFP were tested using the FACS machine.

Dual plasmids of Copper and Estradiol induction

The results from the yeast strain with the copper and estradiol dual plasmid system are displayed in figure 7. The left violin plot shows the GFP expression for different estradiol inducer concentrations (Est) and the right plot shows the RFP expression for different copper inducer concentrations (Cup). The violin plot for GFP shows clearly that the expression increases when estradiol is added, with or without the copper inducer. There also does not seem to be a change in expression between the samples with 0/0.01 µM (0 µM copper inducer and 0.01 µM estradiol inducer) and 100/0.01 µM, indicating that the system is not affected by the copper. The right plot for the RFP expression is, however, not as clear as the estradiol one. The RFP expression does not seem to change much for either of the inducers, but there is a slight increase when comparing the 100/0.01 sample with the 0/0 sample, but it is not as distinct as in the estradiol case.

Figure 7: Violin plots of GFP (GRN.B.HLog) and RFP (RED.R.HLog) from yeast strain containing copper and estradiol plasmids. Measured with Guava. Pairwise statistical tests between controls (0uM added inducer) and remaining samples in each strain were performed using unpaired two-sided Wilcoxon-rank-sum tests (ns: not significant, *p < 0.05, **p<0.01,***p<0.001,****p<0.0001)

For additional clarifications there are also density plots for green and red fluorescent expression provided in figure 8. Both of the plots on the left show the GFP expression, and it is rather clear in both graphs that the expression increased as the estradiol inducer is added. The bottom one also shows that the expression does not change as the copper inducer is added, just as the violin plots also did. The RFP on the right side of figure8, is on the other side of the spectra, and is not as clear. However, the bottom graph displays more clearly that the RFP expression is higher for the sample with both copper and estradiol, which indicates that the copper system increases in expression when estradiol is added.

Figure 8: Overlapping density plots for BFP and RFP expressions from yeast strain containing both copper and estradiol plasmids. Measured with Guava.

'

A comparison of the given fluorescent expression for GFP compared to controls is given in figure 9. The graph shows a large difference in expression between native yeast and transformed one, corroborating the notion that the system leaks. The most interesting aspect is, however, that there is no difference in expression between case and control with 0.01 µM estradiol and/or 100 µM of copper, meaning that the estradiol system is not affected by the copper inducer.

Figure 9: Violin plots of GFP expression from three yeast strains, one case and two controls. Measured with Guava. Pairwise statistical tests between controls (0uM added inducer) and remaining samples in each strain were performed using unpaired two-sided Wilcoxon-rank-sum tests (ns: not significant, *p < 0.05, **p<0.01,***p<0.001,****p<0.0001)

A similar violin plot for RFP with controls, is displayed in figure 10. The result from this figure shows a great difference in expression if a plasmid is integrated (case) or not (CENPK control) into the yeast, showing the leakage of the system. The most important aspect in this figure is the difference in expression between the 100/0.01µM samples. For the control with only the copper plasmid integrated, the expression is much lower than for the yeast with both estradiol and copper plasmids, providing additional results that support the notion that the copper system is affected by the presence of the estradiol plasmid.

Figure 10: Violin plot of RFP expression from three yeast strains, one case and two controls. Measured with Guava. Pairwise statistical tests between controls (0uM added inducer) and remaining samples in each strain were performed using unpaired two-sided Wilcoxon-rank-sum tests (ns: not significant, *p < 0.05, **p<0.01,***p<0.001,****p<0.0001)

Dual plasmids of Tetracycline and estradiol induction

The results from the yeast strain containing both tetracycline and estradiol plasmids are given in figure 11. The left violin plot shows the logarithmic value of TetON-BFP (log_BFP), for different tetracycline inducer concentrations, while the left shows the Estra-GFP expression (log_EGFP), for different estradiol inducer concentrations. The provided BFP results shows, as discussed for the single plasmid results, that the tetracycline system is leaking due to the similar expression both with and without inducer. The GFP expression for estradiol is also rather similar for non-induced sample and induced sample, though the expression is higher in intensity for the induced sample.

Figure 11: Violin plot of BFP and GFP from yeast strain with tetracycline and estradiol plasmids. Measured with FACS. Pairwise statistical tests between controls (0uM added inducer) and remaining samples in each strain were performed using unpaired two-sided Wilcoxon-rank-sum tests (ns: not significant, *p < 0.05, **p<0.01,***p<0.001,****p<0.0001)

The density plots of the same samples are provided in figure 12, where the expression for both BFP and GFP is clearer than in figure 11. The bottom plot on the left for BFP shows a higher expression than the system is induced then when it is not and thus, that the inducers change the expression slightly when added. For the estradiol system the highest GFP expressions are only shown for the induced sample, showing a change when inducers are added. There is still a large portion of the expression overlapping between the samples, indicating leakage.

Figure 12: Overlapping density plots for BFP and GFP expressions from yeast strain containing both copper and estradiol plasmids. Measured with FACS.

Dual plasmids of Tetracycline and copper induction

The results given for the dual plasmids of tetracycline and copper are displayed in the violin plot in figure 13. The left plot shows the TetON-BFP expression (log_BFP), where there is a slightly higher expression for the induced sample than the non-induced one. Similar results are displayed in the right plot for Copper-RFP (log_miRFP670), where the expression is higher for the induced sample than the non-induced. The differences in expression between induced and non-induced are not large.

Figure 13: Violin plot of BFP and GFP from yeast strain with tetracycline and estradiol plasmids. Measured with FACS. Pairwise statistical tests between controls (0uM added inducer) and remaining samples in each strain were performed using unpaired two-sided Wilcoxon-rank-sum tests (ns: not significant, *p < 0.05, **p<0.01,***p<0.001,****p<0.0001)

Density plots for BFP and RFP from the same samples as in figure 13 are given in figure 14. In both cases of BFP and RFP the expression from induced and non-induced samples is overlapping, but there is a slight difference where the induced one is shifted slightly more to the right. This is especially clear for the bottom graphs, indicating a slightly higher expression for the induced samples.

Figure 14: Overlapping density plots for BFP and RFP expressions from yeast strain containing both copper and estradiol plasmids. Measured with FACS.

Conclusion

The results from the dual plasmids displays a somewhat clear increase in fluorescent protein expression in an induced sample compared to a non-induced one, indicating that the system is functional to some degree. There still is, however, a noticeable expression from the non-induced sample for more or less all strains, showing that all three systems leak to some degree. Additionally, the estradiol was not affected by the presence of the copper plasmid, as it provided the same signal with or without copper inducer, while the opposite could be said for copper, where the overall output signal was higher for strains with the estradiol plasmid even when there was no copper inducer present.

Overall, it was observed that the system works for dual plasmids reasonably well despite leakage, providing data that the plasmids and systems work together.

Combinatorial benchmarking (three induction systems)

Fluorescent expression analysed for the strain containing all three induction plasmids are given as violin plots in figure 15. The plots show the three fluorescent expressions separately for three different inducer concentrations. The TetON-BFP (log_BFP) plots shows an increase in expression when the tetracycline inducer is added compared to when it is not added, though the difference between the two concentrations of inducers does not seem to affect the system greatly, but the increase in expression with the added inducer shows that the system is still functional to some degree with all three plasmids. The plot in the centre for Estra-GFP (log_EGFP) expression does not change between the non-induced sample and the induced ones, indicating that either the system is inhibited by the other plasmids or that the plasmid has dropped out of the cell. For the right and last plot, the expression of Copper-RFP (log_miRFP670) is displayed. The expression increases as the copper inducer is added, indicating that the system is still functional. As previously shown, the RFP induction did increase as the estradiol inducer was added, but this is not the case in this experiment, which means that the copper system might only be affected by the other plasmids rather than other inducers.

Figure 15: Violin plot of BFP, GFP and RFP from yeast strain with tetracycline, estradiol and copper plasmids. Measured with FACS. Pairwise statistical tests between controls (0uM added inducer) and remaining samples in each strain were performed using unpaired two-sided Wilcoxon-rank-sum tests (ns: not significant, *p < 0.05, **p<0.01,***p<0.001,****p<0.0001)

Density plots of the same data as in figure 15 is shown in figure 16, where the close expression overlap shown in the violin plots from figure 15 becomes extra clear. The slight increase in expression for both BFP and RFP is especially clear in the bottom graphs, where the induced samples are located more to the right or higher up on the expression axis. The GFP density plot in the middle is almost identical for all samples.

Figure 16: Overlapping density plots for BFP,GFP and RFP expressions from yeast strain containing both Tetracycline, estradiol and copper plasmids. Measured with FACS.

Conclusion

The overall results from the triple plasmid system shows a slight increase for TetON-BFP and Copper-RFP when inducers are added, but not for the Estra-GFP, meaning that the system works for BFP and RFP, but not for GFP. The results from the dual and single plasmid systems shows that the estradiol system functions the best out of the three systems, and it should translate to the triple plasmid system. The single plasmid systems also showed that the estradiol plasmids most likely has a faster plasmid drop-out rate, which could be the reason for the estradiol system results.

The combinations of all experimental data for the plasmid induction systems shows that all three systems should work rather well, with some leakage, and also shows that the plasmids are affected when more plasmids are present in the same cell. Interestingly enough the results provided by the guava machine showed a much clearer expression separation for different inducers than the FACS, which is something that could be taken into consideration when additional testing is performed in the future. Another beneficial part of using the guava is the automatic sampling from a microplate, which helps increase the amount of different inducer concentrations when needed. Therefore, for a future perspective, when benchmarking a precise amount of inducer for a specific output signal a wellplate reader such as the guava will be the better choice and the BFP gene should be replaced to ensure compatibility.

Another important aspect that needs to be tested in the future, is in which growth phase each inducer is the most efficient and has the best function. Because as of now each inducer were tested using the same procedure, expect for experiment A for single plasmids of copper and estradiol which suffered from some time lag, further described in the notebook, though essentially, the cells were cultivated for a longer period of time before being measured in the guava, which might be a reason for the higher numeric value of expression of experiment A compared to experiment B in figure 1 and 3.

To summarize the results, all systems were proven to work to some degree, and to be dependent on concentration of added induction agent, which are the key results for the future of our project.

FAS system integration

To be able to regulate and control the thioesterase production and thus the fatty acid profiles, one crucial step is that the native FAS system should be replaced with a bacterial one. The bacterial FAS system that we have integrated consist of 8 different genes, which we have combined into integration fragments with two genes each, resulting in 4 DNA fragments to be integrated. To simplify the prosses, each DNA fragment was inserted into a EasyClone marker free integration plasmid from the lab, which matches a specific integration site in the yeast genome.

To confirm the success of a transformation a colony PCR was performed with primers surrounding the integration site, meaning that if the transformation was successful, a fragment around the size of 4500bp would be visible on the agarose gel when running the PCR product through it. If the integration was not successful, a fragment the size of around 1500bp would be found instead.

Three out of the four gene fragments were successfully integrated into the yeast genome, and we confirmed the integration through sequencing. The gel image in figure 17 shows the results from a sapphireAMP PCR after transformation of fragment number 2, where 16 colonies were tested. The top band present for 13 out of 16 colonies is located around 4500bp, thus, proving that the integration was successful.

Figure 18: Sapphire PCR for fragment 2, in integration site XII-4

The PCR results run on a gel shown in figure 19 shows that the transformation of fragment 1 was successful. 10 out of 16 colonies were successful, with some bands more intense than others. It is clear that the top band for example from colony 5, 6 and 8 show a band around 4500bp.

Figure 19: Sapphire PCR for fragment 1, in integration site XII-1

Figure 20 displays a clearer gel image of a successful transformation of both fragment 1 and fragment 3, with a band around 4500bp.

Figure 20: Phusion PCR for fragment 1 and 3, in integration sites XII-1 and XII-5.

Conclusion

Three out of four fragments containing six out of the eight genes required to establish a bacterial FAS system were successfully integrated into the genome.

Unfortunately, our team ran out of time to finish the fourth and final transformation for a complete bacterial FAS system. Therefore, the next step would be to complete the fourth integration and then use the thioesterase linked induction systems for fatty acid production. Additionally, a fifth fragment could also be transformed after the bacterial FAS system had been integrated. The fifth fragment that our team planned to integrate was an antisense RNA coding fragment, to downregulate the yeast’s native FAS system, however this step was not implemented due to time constraints.

Data Availability

All code can be found on GitHub – iGEM Judging release.