Team:EPFL/Deletion Strains


Espress'EAU - Deletion Strains

Abstract

We created S.Cerevisiae reporter strains that produce a fluorescent readout when the cell activates certain transcription factors involved in the stress response pathway. Additionally, we wanted to assess how pesticides would affect the growth behaviour of cells lacking these transcription factors: we selected strains from a knockout library and measured the optical density of the cells in presence of different contaminants.

Methods

Our pesticides selection for the screening of their influence on yeast strains growth behaviour was based on the observations from the latest NAQUA report on groundwater in Switzerland: monitoring results up to 2014 showed that the amount of some pesticides (or their metabolites) exceeded the legal threshold in Switzerland. As a consequence, the pesticides selected were : Atrazine, Bentazone, Simazine, Metolachlor, Diazinon, DDT (insecticide), Desisopropyl-Atrazine (pertinent metabolite). The different knock-out strain cultures were diluted to an OD600 of 0.05. The different pesticides were added with a ratio of 1:100 to these strains and their growth was monitored with a plate reader for around 18 hours. The plates were shaken and then the optical density measured regularly. A 96-wells plate was used for these experiments. Every experiment consisted in one strain and 4 different pesticides Several controls were also placed in the plate: a negative control without pesticide, a control using methanol (in the same concentration than the pesticides) as the stock of commercial pesticides was diluted in methanol and YPD, which was the medium used for these yeast cultures.

Growth curves

As a first step, we wanted to see if there is any difference in the growth curves of wild-type strains of yeasts and the knock-out strains in the presence, or not, of several pesticides. We tested the pesticides listed in the Methods for different concentrations in methanol: at first, 1mg/L with 10% of methanol and 100µg/L with 1% of methanol.

Figure 1

Figure 1: MSN2 and wild-type strains OD600 measurements across time, with and without Bentazon added.The concentration of Bentazon was 1mg/L in a solution of 10% Methanol.

Figure 2

Figure 2: MSN4 and wild-type strains OD600 measurements across time, with and without Bentazon added.The concentration of Bentazon was 1mg/L in a solution of 10% Methanol.

Figure 3

Figure 3: YAP1 and wild-type strains OD600 measurements across time, with and without Bentazon added. The concentration of Bentazon was 1mg/L in a solution of 10% Methanol.




Our first observation is that in presence of Bentazon, even the wild-type strain's growth is affected: indeed, we can see a clear delay in the starting point of growth, and the stationary phase of the growth is not reached at the end of the experiment. Furthermore, deleting the master regulator genes from the yeast genome alone also seems to have an effect on the initiation of the growth of the cells, which might be explained that the fact that they are less resistant to environmental stress than the wild-type. Adding pesticides make this effect even more important: the knock-out cells initiate their growth after approximately 15 hours. As a consequence, knocking-out the master regulator genes makes our yeast strains more sensitive to stress.
However, our solution used for the experiments has a concentration of 10% methanol, so from these experiments, we can not draw any conclusion regarding the effect of pesticides alone on our deletion strains. In order to see if methanol has any effect on our strains, we measured the growth curve of the knock-out strains in presence of a solution of 10% methanol.

Figure 4

Figure 4: MSN2 strain OD600 measurements across time, with Bentazon and a solution of Methanol (10%) added.The concentration of Bentazon was 1mg/L in a solution of 10% Methanol.


Figure 4 above shows the results of our measurements: we are not able to confirm that the effect we are observing on the growth of our deletion strains is entirely due to pesticides or methanol. Decreasing the concentration of methanol to 1% does not induce any effect on the growth of our strains.

Growth factor

As our previous observations were rather qualitative and concerned the delay in the initiation of our yeast strains, we decided to take a more quantitative approach in order to see if there was a difference in the growth rate of the cells. First, we decided to model the growth of our cells by a logistic function, its equation being described in Figure 5 below.

Figure 5.1


Figure 5.2

Figure 5: Mathematical expression of the logistic function used to model the yeast growth behaviour.




In order to quantitatively capture the growth of yeast, we looked at the different knock-out strains growth behaviour (OD600) in presence of the pesticides listed above. For experimental purposes, we reproduce each measure 4 times in order to have 4 replicates. For each measure, we fit a theoretical logistic function to our experimental data and we observe the growth factor k: it indicates the growth rate in the log phase of the growth curve. Once we obtain for each OD600 measurement the growth factor k, we want to assess either or not there is a significant difference in the growth rate of knock-out strains in the presence of pesticides or not. For this, we will perform a Student's t-test on pairs of samples and compute the p-value: a sample is composed of the 4 replicates for an experiment involving a knock-out strain and a pesticide. Student's t-test compares the mean of two samples, the null hypothesis being: "the two sample means are equal". We perform the t-test on pairs of samples composed of one deletion strain in presence of pesticides (1mg/L in solution of 10% methanol) and the same deletion strain in presence of methanol (10%) which represents our control. For each pair o samples, the test statistic t is calculated following the formulas in Figure 6.




Figure 6.1


Figure 6.2

Figure 6: Mathematical expression of the Student's test statistic.




However, as we are performing multiple hypothesis testings, we are in the framework of multiple testing: when we perform multiple tests, rejection of the null hypothesis can occur by chance and thus be a false discovery (incorrect rejection of the null hypothesis). In order to control the false discovery rate, we apply the Benjamini-Hochberg 1 procedure. This procedure corrects the p-values of individual t-tests in order to limit the false discovery rate.

Figure 7

Figure 7: MSN2 strain growth factors in presence of different pesticides (concentration of 1mg/l, 10% of methanol). The Control is a solution of Methanol at 10%




Performing the Benjamini-Hochberg procedure on the growth factors obtained from the different replicates (deletion strains + pesticides) yields the Figure 7 for the MSN2 strain. We found only two significant p-values (level of 5%) when comparing the different pesticides with the control solution of methanol: Diazinon and Metolachlore. For the other deletion strains, no pesticide was found to show a significantly different behaviour when compared to the solution of Methanol.

Conclusion

Our experiments were aimed to assess if pesticides would affect the growth behaviour of yeast strains that are intended to be more sensitive to environmental stresses. Observing the growth curves of deletion strains shows that those strains are indeed more sensitive when in presence of a solution of pesticide (concentration of 1mg/l) and Methanol (10%), but we are unable to know whether this effect is due to pesticides or methanol. Looking at the growth factor of the strains and comparing them to a methanol control, the MSN2 strain showed a significantly different behaviour for 2 pesticides.

References

  1. Yoav Benjamini; Yosef Hochberg. Controlling the False Discovery Rate: A Practical and Powerful Approach to Multiple Testing. Journal of the Royal Statistical Society. Series B (Methodogical), Vol. 57, No. 1 (1995), 289-300.

Special thanks to our sponsors!

Follow us on our social media.

@EPFL iGEM 2020