Team:HebrewU/Results

HebrewU HujiGEM 2018



As the enzymes in the pathway come from multiple organisms and biological settings, it would be naïve to expect high specificity to the substrates used in the experiments. Thus monitoring the change of the parent molecule might be challenging. On the other hand, following newly created degradation products could prove feasible. However, these processes occur in the presence of many other yeast enzymes, and as such tracking, their, and only their, activity is extremely difficult. We had to be creative with the experiments we ran to examine the degradation process. As the chemicals we are attempting to degrade, are toxic an option for biological assays arises. To complement the biological assay we engaged in an untargeted whole metabolic comparison through high-resolution chemical analysis. The experiment involved growing transgenic yeast in medium with TCDD like chemicals and then examining two parameters:

1. Growth rate alteration- monitored using O.D. measurements.
2. Changes in chemical composition of the yeast- monitored by Mass Spectrometry.




As with all experiments, the results are only as strong as its control groups. Every experiment run had three control groups for each strain tested (table 1).



Each experiment involved yeast expressing catabolic enzymes treated with a compatible toxic compound (Treatment 1). By growing the yeast with the chemical, we could then analyze how they fair, and if the enzymes are indeed breaking down the compounds. To fully understand the extent of growth inhibition, this treatment was compared to the growth rate of yeast not treated with the chemicals (Treatment 3); as well as yeast which is not armed with the tested enzyme (Treatment 2 and 4) with and without the chemical respectively. This control group gave us a proper understanding of how ‘Wild-Type’ yeast would react to being exposed to these compounds.

A vector with no catabolic enzymes, as opposed to actual WT yeast, was used, so they could be grown in the same medium. As discussed on the yeast design page we used amino acid auxotrophic mutants as selection markers. This means we grow the yeast in a dropout media lacking a specific amino acid. This on its own is enough to alter growth rates, and as such yeast was transformed with the same vector (allowing for growth in absence of amino acids) but without the enzymes for breaking down the compounds. In addition, these treatments also allowed for the measurement of the “expression load” of the enzymes. Clearly, the yeast expressing genes comes at a biological cost. By comparing differences in growth with and without the enzymes, without the toxic compounds, the effect of this cost can be understood.

As both the chemicals were dissolved in methanol, all treatment contained identical concentrations of methanol. This is because similar to the amino acid dropout, methanol in the medium also affects growth rates and metabolism of the yeast.

The team considered to leverage the ability to induce the promoter GAL1/10 as a control group but came to the conclusion the creating different mediums- one with glucose and with galactose- could create other undesired changes in the yeast’s metabolism. As such, this would not give us an accurate comparison; as such empty vectors were used as the control group.

With all of these controls in hand, we could begin our experiments. As using TCDD in the lab was not possible due to safety reasons, we used to separate compounds to simulate the two stages of the pathway. To test the first stage of the pathway-dechlorination- 1,2,4,5-Tetracholobenzen (TCB) was used. After dechlorination, the second stage is the breakdown of carbon backbone, similar in all dioxins. To test this stage, Dibenzofuran (DBF)- a non-chlorinated dioxin-like compound- was utilized.

Growth Curve Rate Experiment: Though the chemicals used in our experiments were less toxic than TCDD, they are still recognized as harmful molecules. As such, we anticipated the presence of these molecules in the growth medium, would inhibit or completely halt the growth of yeast. As some of the breakdown products of these compounds are also toxic (i.e. pyrocatechol), we were not certain if the enzymes would reduce the inhibition, or actually further decrease the growth rate. By contrast, yeast lines with the proper enzymes would be able to break down the chemical, and growth would not be inhibited, or we would see a significant reduction. Based on this hypothesis, we conducted growth curve experiments to test the enzymatic activity oinr the transgenic yeast. To measure growth, we used a spectrophotometer to measure absorbance at 600nm. By reading the absorbance regularly (5-10 hours during lag phase and every 2 hours during log phase) we could accurately track the growth rates of each yeast line for comparison. What these graphs show, is that the enzymes actually further inhibited growth. Based on the potential growth (in the control group) we created the following equation to show the inhibition examined. To present this data in more intuitive way, we created the following equation to measure the inhibition, caused be the presence of the enzymes. (Strain growth in control media - Strain growth in DibenzofuranCompound spiked media) Full growth Potential This translates to: (Abs Teament2- Abs Treatment1) or (Abs Treatment4 – Abs Treatment2) Abs Treatment4 This calculation was based on equations used for RT-PCR’s, resembling ΔΔCt calculationsallows us to measure the difference cause by the enzymes. By normalizing this difference to the full growth potential, we can compare between the different strains, to understand how the enzymes effect growth in the medium. WhenWe then applied this calculation to growth for both strains, at each measurement during the log phase of growth we get the following graph:.
After the initial growth curve experiment was done, we sent samples for chemical analysis. Using chemical analysis, we could track two things that would allow us to understand the effectiveness of the researched enzymes: 1. Lower levels of the compound we added to the medium 2. Higher levels of expected metabolites in the system Observing either of the phenomena, would point to active enzymes. At first we tried to use HP-LC, but the molecules we were looking for trying to track were not easily tracked found by this method. We then switched to GSGC-MS, specifically using APCI. APCI is a method of ionization which ionizes the sample through the addition of atoms, or small molecules, creating a charge that is read by the machine. The results produced, are masses compounds-represented by molecular mass- of chemical compounds found in the sample. The masses produced se masses are accurate to 5 digits after the decimal- this means the a very small, amount of possible accurately predicted chemical equations can be accurately predicted. The GC-MS analysis was unbiased. This means we did not direct the equipment to look for a specific molecule; rather it processed all the available data from each sample. As we were not sure we would be able to recognize the starting compound easily, we hoped through this method, to find products of the compounds’ breakdown. To sift through the thousands of compound masses delivered by these essays, we first used Seive and Compound Discoverer software to align the sample peaks, and produce primary statistical analysis of the differences between treatments. Even after this help from software, there were thousands compounds to go through. To narrow our field, we looked specifically for compounds that fit the following criteria: 1. Compounds are only existent (or existent in significantly higher concentrations- at least a 60% increase, though usually higher) in samples containing both the compound and the enzymesfrom Treatment1, and not in the control groups. This means this compound is result of the presence of both the enzymes and substrate (compounds). 2. These differences in addition to being significant, had to statistically relevant, using the standard cutoff of P value ≥0.05. 3. Compounds with a molecular weight that was in the range of our metabolites. 4. Chemical equations that corresponded to metabolites we anticipated, with minor variations as a result of the ionization process and/or other insignificant changes resulting from native enzymatic activity.
TCB did not inhibit growth in yeast. We see similar growth curves in the same yeast line both with and without TCB. When we introduced our genes in to the system, we again saw no significant effect in both the TCB and Control Media. As the inhibition effect was weak, we see no significant difference between the two strains tested:   We believe that the LD50 for TCB was not low enough to affect growth at the concentration we were able to add it to the medium; additionally, its metabolic products are not toxic. As such, we did not see a strong biological effect. Despite this, we moved to chemical analysis, an there we found promising results. Though we did not see a strong biological effect, we did find promising results in the chemical analysis. The chemical reactions we expected to see were the removal of chlorine molecules. As the benzene ring is a very stable structure, and the Dehalogenase does not act to dismantle it, the metabolites we were looking for would be a 6 carbon rings with OH groups replacing the Cl molecules. Following the parameters outlined above we found the following masses that we believe represent metabolites of TCB:   M. Weight RT [min] Ratio: T1/T2 Ratio: T1/T3 P-value: T1/T2 P-value: T1/T3 219.11041 4.456 8.313 60.437 0.002 0.0001 165.0763 3.042 16.735 23.84 0.018 0.005 153.04249 5.068 3.268 3.67 0.016 0.015 100.05234 4.531 4.051 22.511 0.002 0.00001 100.05232 4.448 3.969 11.487 0.023 0.0004 RT: Retention Time; T1: TCB, Enzymes; T2: TCB and No Enzyme, T3: No TCB, No enzyme Though we cannot account for all of these masses with relevant chemical equations, due to the endless enzymatic options in yeast, and the range in which they fall, we believe it is likely they are products of the dehalogenation of TCB. Additionally, some of the mass represent molecules we expected to see. A molecular weight of 165.0763 represents the molecular equation: C6H12O5. One of the spatial structures this equation can take is: This molecule could likely represent our original 6 carbon ring, with all 4 of its chlorine molecules replaced by hydroxy groups, and an additional hydroxy group added during ionization process or proton added to an oxygen molecule already present.   Additionally, the molecular weight 153.04249 is of interest to us. Representing the molecular equation C7H7O3N, also possibly representing our original benzene ring, with hydroxy groups. The Though we cannot account for a specific reaction which would introduce an N molecule, nitrogen is abundant in yeast cells. One possible structure could be: Both of these molecules represent promising results. They point at the complete dechlorination of the Tetrachlorobenzene and lead us to believe that this enzyme could help transform 2,3,7,8 TCDD into the non-chlorinated dibenzodioxin. We will further discuss the results in the last section of this page. Dibenzofuran Results:
In contrast to the TCB results, we did find that DBF inhibits yeast growth, as is shown in the graph below. This experiment was run solely with Empty Vector lines of yeast: When the DBF is added to the medium, it takes time for it re-dissolve (changing from a methanol solution to a water based solution). As the readings are done with spectrophotometer, and measure absorbance of light, the first reading shows the different absorbance as a result of a different DBF concentrations in the media, as the yeast concentrations at this point are identical. Using the Control group as the “maximum potential” for growth under the given conditions, we created the following graph measuring inhibition as effected by DBF. This graph is based on measurements after 48 hours of growth. This shows not only that DBF inhibits growth, but that the effect is quantitative. As the concentration grows, so does inhibition. With the above data, we decided to run experiments with the enzymes at a concentration of 0.5 [mM]. This appears to be enough to inhibit growth, but also not such a high concentration that even if the yeast breakdown 50% of the compound they could still be inhibited; such as could be the case at 1.5 [mM]. When we ran a growth curve experiment with D24 yeast line, equipped with appropriate enzymes to breakdown DBF, we expected to see the effect of inhibition reduced. The hypothesis was if DBF inhibits growth, and D24 could break it down, we would see growth similar to that in the control group., but we actually observed the opposite: What these graphs show, is that the enzymes actually further inhibited growth. Based on the potential growth (in the control group) we created the following equation to show the inhibition examined. What these graphs show, is that the enzymes actually further inhibited growth. Based on the potential growth (in the control group) we created the following equation to show the inhibition examined. (Strain growth in control media - Strain growth in Dibenzofuran) Full growth Potential This calculation was based on equations used for RT-PCR’s, resembling ΔΔCt calculations. When applied to growth for both strains, at each measurement during the log phase of growth we get the following graph: Here we clearly see that the metabolic effects of the enzymes in the D24 strain further inhibit the growth of yeast, rather than reducing inhibition. As the log phase begins around 50 hours, we see this effect begin as the yeast become more active. At first the results were confusing, but we found a number of possible explanations. The final product of the metabolic pathway our team designed, is pyrocatechol. This compound is toxic and a known carcinogenic. In plants, this issue is solved by native enzymes, but it possible that yeast don’t have the proper Polyphenol Oxidase enzymes to break this product down. As such, it possible that this, or other products caused by the breakdown of Dibenzofuran are causing the increase in inhibition. It was clear discernible fromto us from the bio-assay that there was metabolic activity at play, we now progressed to chemical analysis to better understand what was going on. Employing a process similar to that described in the TCB section, we identified the compounds of following masses: M. Weight RT [min] Ratio: T1/T2 Ratio: T1/T3 P-value: T1/T2 P-value: T1/T3 206.16661 2.669 1.654 1.647 0.0071 0.0126 173.12369 13.361 1.931 4.483 0.0044 0.0002 150.05405 1.631 23.375 20.906 0.004 0.002 137.03550 1.611 30.85 25.319 0.0013 0.0004 130.02641 3.127 8.383 2.757 0.0025 0.0443 104.04775 9.873 1.983 2.008 0.0016 0.0009 RT: Retention Time; T1: TCB, Enzymes; T2: TCB and No Enzyme, T3: No TCB, No enzyme As these molecules consist of more carbon atoms, anticipating the structures given for each mass become substantially harder, and our guesses would be less accurate. That being said, the masses 206.16661 and 137.03550 correlate closely with molecules in our pathway. 2,2',3-trihydroxybiphenyl and Salicylic Acid (a break down product of Dibenzofuran but not TCDD) weigh in at 202.2 and 138.1 respectively. The replacement of a carbon molecule with oxygen, or changes in the hydrogen make up during ionization could presumably account for these changes in mass. Discussion: The result here are preliminary but substantial. To further understand the chemical makeup of the breakdown products, continued Mass-Spech analysis is required. To do this we can employ a number of methods: Using methods such as 1. Obtain pure chemical standards corresponding to the molecules we suspect as breakdown products. We can run these standards against our sample and see if similar profiles appear. 2. MSMS-, we can breakdown the specific molecules we suspect as metabolites and run them in Mass- Spec, and further strengthen our hypothesis.. These broken down products should have a very specific time signature. If we put together the molecules from this small window of time, we should be able to rebuild the molecule in question. 3. Use flagged molecules such as radioactive isotopes or Hydrogen–deuterium exchange, and see if there are discrepancies in the analysis. These methods would further strengthen our hypothesis of the breakdown molecules we are working with. Additional growth curve experiments with more toxic chemicals, even using 2,3,7,8-TCDD itself could potentially show clearer results in the biological assay. We have completed the first stages of research, and already transformed plants with the Haloacid Dehalogenases. Further examination is required as we continue to develop a genetically engineered solution to TCDD, chlorinated pollutants, and dioxins.