Team:MichiganState/Engineering
Engineering Success
Gene Engineering
Introduction
In order to accomplish our goal of aiding bees in the breakdown of harmful pesticides like imidacloprid, we decided to introduce a detoxification enzyme into the bee gut microbiome. We needed to first find an organism that would establish in the bee gut and then design a secretion system that could export our enzyme into the bee gut where it could detoxify pesticides.
Research
In our research, we discovered the Bee Microbiome Toolkit [1], which designed and tested vectors and promoters that could be used to genetically engineer bacteria of the bee microbiome. This was done by Nancy Moran and her team at the University of Texas Austin. We found that Moran’s team had success genetically engineering the bee gut symbiont, Snodgrassella alvi. S. alvi is a gram-negative, microaerophilic bacterium located in the hindgut of bees. They were able to express RNAi constructs in S. alvi, which silenced bee genes when the recombinant S. alvi established in the bee gut [2]. This inspired us to use S. alvi as our chassis organism, as we too wanted to be able to express a construct in a microbe that could establish in the bee gut. From the Bee Microbiome Toolkit, we discovered the vectors, promoters, and CRISPR system that was successful in S. alvi. Once we decided we could use S. alvi to produce a detoxification enzyme for the bee, we needed a way for the enzyme to leave the bacterial cell and be available in the bee hindgut. We researched many different secretion systems that would allow us to accomplish this. We decided to use a Type 1 secretion system, as these systems span the entire membrane of gram-negative bacteria and secretes specific, tagged proteins into the extracellular space in one step [3].
We decided to integrate the HlyA secretion system into our plasmid which uses a secretion tag attached to the carboxyl terminus to secrete HlyA, a blood-lysing enzyme. HlyB is an ABC transporter located in the inner membrane, HlyD is a membrane fusion protein anchored in the inner membrane that spans across the periplasm, and TolC is an outer membrane protein. HlyB interacts with the C-terminus attached to the HlyA gene and recruits the other membrane proteins to form a continuous channel that HlyA will be secreted out of [4].
Imagine
We next visualized how we would construct this plasmid to incorporate the three HlyA secretion system proteins, HlyB, HlyD, and TolC. Before deciding on an enzyme to secrete we wanted to imagine what secretion would look like using this system, so we decided to use mRFP, a slow-folding fluorescent protein, in order to visualize secretion and determine the effectiveness of our secretion system. The slow-folding nature of mRFP was important, as slow-folding proteins are typically secreted by Type 1 secretion systems. The mRFP gene linked to the HlyA C-terminus secretion tag, HlyB, HlyD, and TolC would all be added to our vector. We decided to use the vector pBTK510 from the Bee Microbiome Toolkit, as it was found to be the most successful plasmid in S. alvi.
We also researched various biocontainment mechanisms that could be engineered into our construct in the future, if BeeTox were to be administered to bees out in the field. More information about this can be found on our Proposed Implementation page. Since our plans for biocontainment involve using CRISPR to integrate a gene onto the S. alvi chromosome, we knew we wanted to test genome integration of a gene and its expression, as this has not been documented before in S. alvi. We imagined that we would use CRISPR to deliver GFP to the S. alvi chromosome. We could test the expression of this integrated gene by measuring and quantifying fluorescence.
Figure 1: This is a schematic diagram of how we imagined to test the secretion system in S. alvi.
Design
We began designing our constructs using Benchling for use in our experiments. The HlyB, HlyD, and TolC genes combined are very long, so we had to break them up and add them to our vector separately, as opposed to transferring them all at once via a gblock. We first designed a gblock with mRFP (linked to the secretion C-tag) and TolC. A restriction site was introduced at the end of TolC, where the HlyB/HlyD construct will be inserted. We planned to amplify out the HlyB and HlyD genes from the iGEM 2019 Distribution Kit, which had the genes linked together on a construct. To design our chromosomal integration system to introduce GFP into S. alvi, we first found a place on the S. alvi chromosome which would not interrupt the function of other genes so we could integrate the new gene. We chose an open space on the genome that did not contain any genes, and was in between two other genes, so there was less of a chance of an unannotated gene being disrupted by our gene insertion. We designed a gblock that contained the gRNA targeting the PAM site on the chromosome where Cas9 will make a cut, as well as the homology regions that correspond to the chromosome where we will insert the GFP. We added a mutated PAM site to the gblock so the Cas9 will not re-target the mutated genome. In between the homology regions of the gblock, there is a restriction cut site into which we will insert GFP to have a complete replacement cassette of gRNA and homology regions flanking GFP that will be inserted.
Build
After designing all of our constructs in Benchling we started planning how to assemble all of our constructs into the appropriate vectors. For the secretion plasmid, we planned to first use PCR to amplify the HlyB/HlyD part from the Distribution Kit. We then planned to perform Gibson Assembly to insert the mRFP/TolC gblock into the pBTK510 plasmidt. Next, we would do a restriction digest with the assembled plasmid we had just created, as well as Gibson Assembly to insert the HlyB/HlyD construct. After this, we will have a completed vector with mRFP, TolC, HlyB, and HlyD.
Figure 2: Here is the final secretion system plasmid that will be created, after following the steps above.
To build our vector for the genome integration, we will first perform a Gibson Assembly with our gRNA gblock to add it to the pBTK599, a suicide vector from the Bee Microbiome Toolkit used for genome integration. We will then perform a restriction digest and subsequent Gibson Assembly to insert the GFP gene in between the homology regions. This will result in a complete replacement cassette plasmid with the gRNA and GFP gene with flanking homology on the suicide vector.
Figure 3: This is the assembled sucide vector, pBTK599, with our replacement cassette.
Test
To test our secretion system, we would first transform our plasmid into E.coli donor strain WM6026. We would then perform a conjugation to transfer the plasmid into S. alvi. Once the plasmid is confirmed to have been established in S. alvi we would separate the cells from the supernatant to test the supernatant for the secreted mRFP. This would be done using a plate reader to measure the fluorescence compared to a standard dilution curve. To test our genome integration experiment, we would first transform the suicide vector into the E.coli donor strain, then conjugate it into S. alvi. We would then perform a second round of transformation and conjugation to add the pX2-Cas9 plasmid that contains the Cas9, which will make the cut at our target location of the genome. We will confirm the integration of the GFP gene into S. alvi with PCR. We would then measure and quantify the fluorescence of the cells using a plate reader. To read our detailed protocols on the different aspects of our experiments, see our Experiments page.
Learn and Improve
We were unable to perform experiments in the lab this summer due to COVID- 19 restrictions. Our graduate student mentor, Kati Ford, started the assembly and transformation experiments, but unfortunately, was unable to obtain results from them. One of the challenges she encountered was that the mRFP/TolC gblock was most likely synthesized incorrectly, as it was not able to be integrated into the vector backbone. We would improve this by ordering a new gblock and starting the experiments again. Our next step would be to hopefully complete these experiments we had planned for this summer in the lab. Unexpected results that could come up could be due to S. alvi’s specific growth requirements - 5% CO2 , as well as certain types of media. We were growing S. alvi in LB, but it has been known to grow better in more supplemental media, such as blood agar or heart infusion agar [5]. This is something we could test to improve the results of our experiments. If the secretion was unsuccessful, we could try testing a different Type 1 secretion system in S. alvi, such as the Has secretion system. If the genome integration experiment failed to produce an appreciable level of fluorescence, we could try inserting the GFP gene into another location on the chromosome. If our experiments are successful, we hope to further test our secretion plasmid by inserting a detoxification enzyme into the construct to be secreted. We would then confirm that this enzyme was secreted and active in detoxifying a substrate. Our next step would be to integrate a biocontainment mechanism into our chassis, using CRISPR. More information on our future steps and their experimental design can be found on our Proposed Implementation page.
Bioinformatics
Transcriptomics Experimental Design
Previous research conducted has shown evidence of imidacloprid detoxification by some species of soil bacteria. However, specific genes and proteins involved in the detoxification of imidacloprid in bacteria are not entirely known. We developed protocols for a transcriptomics study to identify the specific genes involved in imidacloprid detoxification in soil microbes that are capable of detoxifying imidacloprid. The protocol we developed is specifically for P. putida EM371. Through our metabolomics study, we have determined that P. putida EM371 is capable of imidacloprid detoxification. The end goal is to use RNA-Sequencing to identify all genes upregulated in the presence of imidacloprid.
Because RNA-seq is very expensive, we decided to use RT-qPCR to perform preliminary gene expression differences in genes that are likely to be involved in imidacloprid detoxification. Information on proteins involved in imidacloprid detoxification in soil bacteria is lacking; however, some proteins involved in neonicotinoid resistance in some insects are well documented. Through literature research, we identified that the three protein families for each phase of drug metabolism: P450s[1], glutathione S-transferases (GSTs)[2], and aldehyde oxidases (AOX)[3]. These proteins are also common enzymes that give insects resistance to neonicotinoids. Therefore, we assumed that soil bacteria that can grow in the presence of imidacloprid may also use these families of proteins.
We designed primers for the GST, p450, and AOX gene in P. putida EM371. Then, we designed primers for a housekeeping gene, rpoD, and a cellular response control, ahpC. Housekeeping genes are used as controls since their expression levels should not change when growth conditions change. The housekeeping gene, rpoD, is the sigma factor for RNA polymerase and has been used as a reference gene in previous studies involving Pseudomonas[4]. The cellular response gene, ahpC, is an oxidative response gene that has been found to be upregulated in Pseudomonas in the presence of ampicillin and MTBE[5]. When conducting the experiment, a control culture (without imidacloprid, ampicillin, or MTBE), cultures with one of the three chemicals, cultures with two of the three chemicals, and a culture with all three chemicals should be grown; and RT-qPCR conducted on each sample condition. The control culture with no chemicals provides insight into gene regulation in ideal conditions without toxins. This will serve as a baseline for rpoD, ahpC, GST, p450, and AOX gene regulation. By running RT-qPCR for each chemical, we can discern between chemicals in which genes are up and down-regulated. Recording changes in the regulation of all genes for all conditions is extremely important for later statistical analysis. The combination of ampicillin and MTBE will provide important insight into the regulation changes of ahpC, a gene known to be upregulated when oxidative stressors are present. RT-qPCR should be run for a culture grown with all three chemicals because we know that ahpC, our cellular response gene, is upregulated when an oxidative stressor is present. If ahpC does not show signs of upregulation, then we know there is something wrong with the sample or the execution of the procedure. If the regulation of rpoD changes drastically from the baseline, then we know there is something wrong with the sample or the execution of the procedure. Troubleshooting the procedure will be required when unexpected results arise from the housekeeping gene and cellular response gene. If the GST, p450, or the AOX gene show no significant regulation changes with the addition of imidacloprid, more literature research will be needed. If rhpC significantly changes in the presence of imidacloprid, then some research efforts will be shifted towards rhpC.
If RT-qPCR provides statistically significant data, then we would proceed with RNA-seq. We would conduct RNA-Seq on samples grown without imidacloprid and samples with imidacloprid. Because of the price of RNA-seq, we are limited in the number of samples we would be able to submit.
To determine the concentration of ampicillin, MTBE, and imidacloprid to add to the cultures for RT-qPCR and RNA-seq, we would record growth curves of cultures with the various combinations of chemical stressors. A good starting point for deciding concentrations of each chemical would be to read the methods section of scientific journals that researched rhpC using ampicillin and MTBE and detoxification of imidacloprid by Pseudomonas. To evaluate more conditions, concentrations from the literature review can be concentrated and diluted by factors of two, five, and ten. The concentrations of each chemical must remain consistent for all the cultures (e.g. 100μM of imidacloprid is picked, then all cultures that are to have imidacloprid added to it should contain 100μM of imidacloprid). The growth of the culture containing all three chemicals will most likely determine the concentration of each chemical because the cells have to metabolize and defend themselves against three toxins.
Metabolomics
Using data collected by LC-MS/MS and analyzed through the GNPS environment, we hope to develop a detailed snapshot of the metabolites generated by P. putida EM371 when exposed to imidacloprid. Using an untargeted approach, we are able to study both the metabolism of imidacloprid into various derivatives and any changes to other metabolic processes when P. putida is exposed to xenobiotic stress.
While we have generated promising preliminary results indicating that we are able to identify imidacloprid and imidacloprid metabolites within the extracellular environment of P. putida after a 96-hour biotransformation period, several parameters still must be optimized to take full advantage of the data generated by our metabolomics experiments.
A key limitation in the interpretation of our preliminary results was the carry-over of imidacloprid across numerous samples. During our first trial, we developed a protocol that exposed P. putida to 1.76 mM of imidacloprid which resulted in the saturation of our chromatography column. This over-saturation of imidacloprid caused imidacloprid to be detected in our negative control samples thereby limiting the interpretation of our trial. While we were not able to analyze the differences in metabolic profiles between the samples treated with and without imidacloprid, we did confirm the ability to detect imidacloprid in concentrations as low as 1 uM.
In future experiments we designed (from which we are still waiting on data), we have scaled-up the total number of imidacloprid concentrations to four concentrations (8.8 uM, 17.6 uM, 35.2 uM and 70.4 uM) to optimize the concentration of imidacloprid such that we will not induce imidacloprid carryover across our samples. Additionally, we have added a concentration gradient of imidacloprid to develop a standard curve based on the relationship of the chromatographic peak area and concentration of imidacloprid.
While the use of chromatographic information to quantify the concentration of a target compound is not a novel idea within iGEM, new computational methods have been released that allow for the use of this chromatographic information to improve the process of generating molecular networks. This new approach, called feature-based molecular networking, is explained in more detail in our Modeling page.
Multi-omics Approach
For most of our wiki, we have discussed metabolomics and transcriptomics as two distinct and separate analytical techniques, but the final desired outcome of this “pathway cracking” approach is the merging of information generated from both thrusts.
Through our metabolomics approaches (including both our classical and feature-based molecule networks) we are able to generate networks consisting of nodes representing each unique metabolite and edges representing the relationship between each metabolite. In general, these edges represent biochemical reactions catalyzed by specific enzymes. Through our transcriptomics approaches, we hope to develop a list of genes that are potentially involved in the metabolism of imidacloprid.
By understanding the relationships between these metabolites and the genes involved in imidacloprid metabolism, we hope to develop a more detailed understanding of imidacloprid metabolism within simpler microbes that have a higher degree for engineering success in the future.
In addition to aiding our process of better understanding imidacloprid metabolism in soil microbes, we have envisioned this multi-omics approach as a potentially powerful tool for measuring the effectiveness of our Bee-tox probiotic within the complex gut environment of the bee.
References
Gene Engineering
- Leonard, S. P., et al. “Genetic engineering of bee gut microbiome bacteria with a toolkit for modular assembly of broad-host-range plasmids”. 2018, ACS synthetic biology, vol. 7, no. 5, 1279-1290, doi: 10.1126/science.aax9039
- Leonard, S. P., et al. “Engineered symbionts activate honey bee immunity and limit pathogens.” Science, vol. 367, no 6477, Jan. 2020, pp. 573-576
- Delepelaire, P. “Type I secretion in gram-negative bacteria.” Biochimica et Biophysica Acta - Molecular Cell Research, vol. 1694, May 2004, pp. 149-161, doi: 10.1016/j.bbamcr.2004.05.001
- Gentschev, I., et al. “The E. coli α-hemolysin secretion system and its use in vaccine development.” Trends in microbiology, vol 10, no. 1, Jan. 2002, pp 39-45, doi: 10.1016/S0966-842X(01)02259-4
- Kwong, W, and Moran, N. “Cultivation and characterization of the gut symbionts of honey bees and bumble bees: description of Snodgrassella alvi gen. nov., sp. nov., a member of the family Neisseriaceae of the Betaproteobacteria, and Gilliamella apicola gen. nov., sp. nov., a member of Orbaceae fam. nov., Orbales ord. nov., a sister taxon to the order ‘Enterobacteriales’ of the Gammaproteobacteria.” International Journal of Systematic and Evolutionary Microbiology, vol. 63, no. 6, 2013, pp.2008-2018., doi: 10.1099/ijs.0.044875-0
Bioinformatics
- Roditakis, E., Morou, E., Tsagkarakou, A., Riga, M., Nauen, R., Paine, M., Morin, S., & Vontas, J. (2011). Assessment of the Bemisia tabaci CYP6CM1vQ transcript and protein levels in laboratory and field-derived imidacloprid-resistant insects and cross-metabolism potential of the recombinant enzyme. Insect Science, 18(1), 23–29.
- Yang, X., He, C., Xie, W., Liu, Y., Xia, J., Yang, Z., ... & Yang, F. (2016). Glutathione S-transferases are involved in thiamethoxam resistance in the field whitefly Bemisia tabaci Q (Hemiptera: Aleyrodidae). Pesticide biochemistry and physiology, 134, 73-78.
- Dick, R. A., Kanne, D. B., & Casida, J. E. (2005). Identification of aldehyde oxidase as the neonicotinoid nitroreductase. Chemical research in toxicology, 18(2), 317-323.
- Savli, H., Karadenizli, A., Kolayli, F., Gundes, S., Ozbek, U., & Vahaboglu, H. (2003). Expression stability of six housekeeping genes: a proposal for resistance gene quantification studies of Pseudomonas aeruginosa by real-time quantitative RT-PCR. Journal of medical microbiology, 52(5), 403-408.
- Yeom, J., Imlay, J. A., & Park, W. (2010). Iron homeostasis affects antibiotic-mediated cell death in Pseudomonas species. Journal of Biological Chemistry, 285(29), 22689-22695.