Bioremediation Database
While we focused on the bioremediation of 1,4-dioxane using the tetrahydrofuran monooxygenase (THMFO) enzyme, there are a number of other environmental pollutants and enzymes that may be good candidates for similar bioremediation approaches. We compiled a database of selected pollutants and corresponding enzymes that have been found to degrade each pollutant. Future teams building on our work may choose to follow our experimental workflow and use the pollutant and enzyme information in our database to design bioremediation systems for other pollutants.
Optimizing Protocols
Our team established a degradation pathway describing the breakdown of 1,4-dioxane via THFMO, which became the basis of our protocols. With literature on plasmid construction and metabolic engineering of the host bacterium, Pseudomonas putida, the degradation pathway of 1,4-dioxane established the basis for our work. During our project, we adapted this information into concise protocols. We continuously optimized the protocols to fit our needs based on accuracy and efficiency. For example, we noticed that P. putida was able to grow in harsher conditions that are considered difficult for E. coli to grow in, which could aid future iGEM teams running experiments with bacteria similar to P. putida in the future. We also optimized the transformation of P. putida and tested the degradation of 1,4-dioxane.
An important component of this process included keeping detailed notes on any experimental errors we faced. We hope our work can help other iGEM teams anticipate and navigate issues that we encountered by providing resources on our troubleshooting experience and solutions to common errors or issues. Information about our troubleshooting process can be found on our wiki page, with detailed description in our protocols, results, lab notebooks, and experiments. In addition, we developed models for predicting the biodegradability of a given compound.
Software and Open Source Design
The best way to prevent the release of hazardous compounds is to not use them in the first place. The tests to determine biodegradability involves time-consuming wet lab experimentation. Computational modeling could possibly expedite biodegradability testing and encourage actors to rigorously evaluate their compounds’ viabilities. To that end we developed a number of machine learning models capable of predicting biodegradability. We achieved results of over 80% accuracy and around .9 AUC. This was accomplished by testing a litany of feature selection methods, models, and training schemes. Additionally, we confirm the physical properties of water with dissolved 1,4-dioxane using molecular dynamics simulations.
Our hope is that our research will assist future work into computational biodegredation models. First, we provided a framework for completing molecular dynamics simulations using the LAMMPS software which can be replicated with higher levels of 1,4-dioxane concentration or other molecules. Second, we believe our rigorous exploration of machine learning models provides valuable insights into why certain models behave the way they do with biodegradability data. We hope this guides other researchers in the intelligent design of machine learning algorithms in the biomolecular space.
All code, data, and molecular dynamics simulation files can be found on our gitlab.
Bioreactor Design
Our bioreactor design was created with the goal of targeting chemical pollutants in drinking water in contrast to traditional wastewater treatment bioreactors, which are mainly utilized for sewage treatments. We aimed for our bioreactor to be materialistically and energetically conservative, implementable in existing water treatment infrastructures, and effective in degrading 1-4 dioxane.
We began our research process by defining what a bioreactor is and what types exist and are being used. We first conducted a literature review regarding the most common types of bioreactors: membrane, activated sludge, and moving bed biofilm reactors (MBBR). When it came to water treatment, we found that the membrane bioreactors may run into a fouling problem, where the filters get clogged with debris and bacteria, while the activated sludge bioreactors were limited in that they needed frequent maintenance and large areas of operation.
Another bioreactor option we considered to host our remediation strategy is the MBBR. This design features a collection of small, high-surface area plastic carriers that provide a surface for beneficial bacteria to grow. These plastic carriers are circulated throughout the contaminated medium, maintaining interaction between the bacterial biofilm and the water contaminant. Due to the simplicity and scalability of this design, we developed a 3D model demonstrating its usage in the context of our remediation strategy. The computer-aided design (CAD) model includes two input centrifugal pumps, capable of providing a continuous supply of nutrients and waste water to the bioreactor, respectively. One output centrifugal pump is employed to extract clean water from the bioreactor following the remediation process. The centrifugal pump model was inspired by online documentation produced by Jorge Eduardo Ortiz1. An elaborate pneumatic system consisting of a framework of pipes and an associated air pump was devised to circulate these carriers throughout the water.
While MBBR offers many advantages, certain considerations bar its usage in our context. Particularly, MBBR requires a bacterial biofilm that is both adherent to the plastic carriers as well as capable of independent regeneration in the water. These criteria are not compatible with our bioremediation strategy. To circumvent the latter, we designed a companion system consisting of side-mounted water jets and modified bar screens to facilitate the periodic collection and replacement of the plastic carriers to replenish their bacterial biofilms. Furthermore, in order to ensure the force of the air pump does not dislodge the biofilm, we designed for a metal mesh to be placed several feet above the air pipes at the bottom of the bioreactor to diffuse the force before hitting the carriers. These adjustments, all of which are demonstrated in the CAD model, were deemed excessive, so we continued our search for a more suitable bioreactor design for our bioremediation strategy. We have made this exploratory CAD model inspired by MBBRs available on our gitlab for any teams who may find MBBRs more aligned with their project.
While researching other types of bioreactors, we drafted a list of important requirements for the bioreactor, including low maintenance, biofilm compatibility (for self-sustaining degradation), and consistent performance. Upon further research, we came across the concept of rotating bioreactors (RBCs), which would have biofilms growing on disks attached to a rotating rod in the bioreactor2. After analyzing several schematics of RBCs and talking with members of Dr. Lin’s lab, we chose to direct our attention to these types of bioreactors and tailor them to our needs. We thoroughly documented our design process with the intention of assisting future teams to direct their development of bioreactors that will be used in similar conditions or environments.
During this design process, we prioritized stakeholder feedback. We reached out to our local community and treatment plant to gauge the interest and compatibility of our model. Surveying community members in Ann Arbor informed us of a preference for a larger-scale bioreactor model integrated with existing treatment measures, as opposed to a small scale design. When visiting the treatment plant in our area, we also learned about how such a model would be realistically integrated into a larger infrastructure, beyond local wells.
Focusing on large-scale implementation alongside existing infrastructure, we completed the design of our RBC, which features a multi-purpose modified screw-shaped axle that is capable of holding the bacterial biofilm, aerating the biofilm, as well as constantly circulating the water in the bioreactor to maintain its interaction of remaining 1,4-dioxane contamination with our custom-engineering bacterial strain. The design, similar to the MBBR design, also includes 2 input and 1 output centrifugal pumps, for the purpose of facilitating the passage of nutrients and contaminated water into the bioreactor, as well as clean water out of the bioreactor. This model was deemed ideal for our setup, since it ensures scalability, and offers a solution for the periodic replacement of bacterial biofilm. Furthermore, the large surface area of the screw-shaped axle allows for greater adhesion of the bacteria, preventing dislodgement during agitation of the contaminated medium. Our design process is reflected in our CAD models, which are available for open-source download for all teams on our gitlab.
Screen capture of two separate orientations of our CAD implementation of a rotating bioreactor
