Engineering in biology can be summarized in four key steps:
Applying engineering principles to biology, with a focus on developing plans and using parts to create constructs with a desired function. Modeling and previous literature is used to make predictions, design experiments, and make data driven decisions about experimental direction.
Biological parts, such as proteins, DNA, or RNA are acquired, manipulated, and assembled to create parts. In the context of synthetic biology, this often includes cloning, transformation, assemblies, PCR, and various diagnostic procedures such as blotting, imaging, or colony counting. This phase also includes implementation of the construct into a chassis or cell-free system
The constructed parts are tested in various assays for proper function, additional effects, and efficacy. Data is gathered during this phase.
The data gathered during the testing phase is gathered and analyzed. The results can feed back into the models into the design phase to improve accuracy, point out flaws during the build phase, or highlight errors during experimentation. This cycle is repeated, often until success is achieved and there exists no desire for improvement.
Our team went through multiple different design-build-test-learn cycles throughout the course of the summer. Each aspect of the project achieved at least one full iteration of the engineering cycle.
Our colonization engineering efforts began with several key modifications backed by reviewing the literature of stomach bacterial conjugation. Along with an abundance of acid resistance pathways, we determined cell aggregation and production of biofilms to be key abilities in the potential for stomach colonization. The expression of curli amyloid fibers is well known to be a prerequisite for biofilm growth, activated during the stationary phase of bacterial growth. We began with the pBbB8k-csg-amylase, encoding for the csg gene family creating the protein complex required for the production of extracellular curli fibers.
Our first steps in wetlab procedures transformed this plasmid into E. coli Nissle, opening the door for biofilm and colonization experiments. Our first experimental steps involved colorimetric biofilm assays using the Amyloid binding dye Congo Red, protocol here From these experiments, several key results guided the rest of our colonization efforts.
First, we showed that uninduced E. coli Nissle carrying the curli plasmid grew stronger biofilms than DH5α carrying the same plasmid. From these results, we concluded that E. coli Nissle, without the upregulation of csg genes, already holds potential for effective colonization on solid environments. Secondly, we discovered a significant decrease in colony growth as concentrations of the inducer arabinose increased. Although initially surprising, further review of the literature indicated excessive curli fiber expression exhibiting toxicity. With this in mind, our next steps moved towards identifying the ideal concentration of the arabinose inducer to upregulate curli expression without decreasing fitness. We did this by running two separate experiments with a range on arabinose concentrations, one in a liquid culture set up for measuring fluorescence reads of Congo Red bound to curli fibers and a second experiment on solid agar plates using qualitative observations of the color of colony growth through accumulation of Congo Red dye. These experiments indicated an ideal concentration between 0.01% and 0.001% arabinose, overexpressing curli fibers while resulting in minimal to no decrease in survival or biofilm production. Not only this, but additional compelling results from comparing these two experiments arose. On the solid agar plates, colonies in stationary phase for both induced and control groups exhibited little to no observable difference in Congo Red accumulation. In contrast, the fluorescent reads from the liquid phase cultures showed extremely significant contrast between these two groups. This indicated a greater curli expression difference during liquid phase than during the stationary phase, an attribute that could increase a cultures ability to thrive in low pH liquid environments such as stomach acid.
From these results, and our induction metrics established, we decided to move forward with survivability testing in low pH environments. As outlined in the procedures, a series of plate read data yielded interesting results. Without inducing Curli upregulation, we found E. coli Nissle to retain survival, albeit with limited growth, as low as a pH of 4. However, after testing the survival of E. coli Nissle induced for the curli producing plasmid, we found significant growth in pHs as low as 3.5.
Fascinated by these new findings, our next objectives sought to combine curli fiber upregulation with constitutive expression of the gadE regulating gene known to activate key acid resistance pathways.
Our acid resistance engineering cycle began with design of constructs for expression of GadE. We set up a triple assembly pipeline that would allow us to perform all necessary data collection while ordering only one gene fragment from twist. The pipeline was designed with the first assembly inserting the GadE gene into a expression vector in front of a Lac promoter to enable IPTG-inducible expression to quantify ideal induction amounts, then into a second vector for reduction in copy number to mimic the genome, and then a promoter swap with a strong, OrthoRep-designed constitutive promoter for high expression. Each assembly was designed using benchling with golden gate assembly standards. The first golden gate used BsaI to assemble the construct(ordered from Twist Biosciences) into a freegenes vector; we selected a medium copy plasmid from the distribution. However, we were plagued with constant assembly failure. We attempted multiple times but could not get a proper assembly. We tested with multiple other fragments with compatible overhangs and could not achieve success either. After consultation with other iGEM teams who also experienced faults in the distribution, we decided to move to assembly two, which worked on the first try with a BbsI golden gate. Once confirmed successful with correct length on gel electrophoresis, we moved onto assembly three, which involved adding in a strong, constitutive promoter we discovered from recently published literature. Using a PaqCI golden gate assembly, we successfully performed a promoter swap, removing the IPTG-inducible Lac promoter. Once completed, we transformed into Nissle 1917 and ran a plate assay testing survival in different pH ranges, with supplemented glutamate (since one of the major acid resistant pathways activated by gadE requires glutamate) AND with curli to compare the resistance offered by both. We also performed a cotransformation with both of them to see if they exhibited potentially synergistic effects. After the plate read was completed, we analyzed the data and determined that GadE and curli coexpressed exhibited a synergistic effect and allowed survival at a pH of 2.0 for long term. From this data, we learned that supplementation with excess glutamate was required for this acid resistant phenotype. We also learned that GadE in the absence of curli does not exhibit strong acid resistance. Our next data driven step is to test for ability to perform tasks such as conjugation or chemotaxis in high levels of acid with coexpression of curli and GadE.
We spent significant time researching the best possible molecular gradient to elicit a chemotactic response. After extensive consideration of our target pathogens, the environment of the stomach and gut, and our potential methods of detection, we settled on using a small molecule induction due to permeability of the membrane, specifically urea, which forms a gradient in the stomach that leads to sites of H. pylori colonization. This allows us to completely bypass any form of engineering or modification of receptors or receptor pathways, which, while incredibly interesting, poses a significant challenge. Our design bypasses this obstacle by using intracellular detection via a transcription factor, UreR, that binds to Urea and activates a promoter. This engineered circuit allows us to express proteins at different rates that correspond to urea concentration within the cytoplasm. Simultaneous to this design process, our computational members began designing a protein binder to EfliG, a small protein that serves the role as a “switch” in chemotaxis flagellar motor rotation. Using mutants of the protein found in literature, a set of binders were designed to inhibit FliG and act as an inducer for tumbling in chemotaxis, which biases movement and inhibits straight movement. Our strategy was to combine this with our urea inducible expression circuit and induce tumbling chemotaxis in the presence of urea, which would bias motion of the bacteria to move down a urea gradient and towards sites of H. pylori colonies.
This was quite a lofty goal due to the precise regulation required for correct communication between urea detection and flagellar motion, so we decided to pursue a highly combinatorial, mass screening approach for expression and degradation optimization. We designed a library generating Golden Gate assembly that synthesized a massive number of combinations of RBS of differing strengths, different FliG binders designed in-silico by our computational team, different N-term and C-term degradation tags, and different copy number backbones. The total number of combinations successfully created was 384 via plate scraping after the assembly transformation. All variants were then cloned into Nissle and screened on our custom designed chemotaxis assay plate, which consisted of a absorbent paper soaked in urea solution embedded in soft agar to form a gradient as it diffused through the liquid. We then streaked our transformation in parallel lines and grew overnight to observe growth patterns. Unfortunately, we could not observe any patterns that suggested chemotactic mobility. After analysis of our construct design and testing method, we believe there were limitations to our assay method. We had issues with horizontal growth, and our plan is to add lanes to divide cells up. Furthermore, our current set up includes a rectangular plate, but we have ideas for using circular plates with the soaked urea paper plated in the middle or on the sides. Altering urea concentration or gel percentage is also of interest to us, to determine whether those variables could be responsible for our lack of success. Finally, using a reporter to determine the extent of variability or even functional expression of our FliG binder would be a useful diagnostic.
Our engineering cycle for this sector of the project was by far the most extensive in comparison to acid resistance and chemotaxis. The large nature of RP4 posed significant difficulties at almost every single phase and experimentation step as most common kits and protocols fail to account for large plasmids. Every single step needed to be planned out ahead of time, even from the smallest transformation to gel electrophoresis to modification. We sought out new protocols, bought new ladders, and researched nuances of procedures to produce the highest accuracy during each procedure. Due to transformation difficulties, we sought to take a unique approach involving complete in-vitro replication and modification. This included three rounds of primer design, with the first round not meeting qualifications for purchase, the second round getting purchased but resulting in significant failed amplification and extensive off target effects, and the third finally achieving success. During the first two rounds, fragments were selected based on gene similarity, i.e. keeping all genes involving the T4SS apparatus on the same segment and all of the partitioning proteins together, but flanking regions were not optimal for primer annealing. Primers were designed using genetic design tools (benchling and snapgene) automatic primer generation and analyzed and improved by hand. Important note is that these primers included extension tails (regions that didn’t anneal to the template). Primers were ordered and arrived, but did not work well, resulting in 1/14 perfect PCRs (no off targets, high yield of target amplicon). Primers were redesigned 100% manually, targeting ideal annealing sites throughout the plasmid. All amplification was done to create homologous ends between each fragment, and primer extensions were minimized to only four fragments and had optimal, long annealing regions with GC clamping. All primers were screened via alignments to check for off target annealing. All primers were further analyzed using IDT’s oligoanalyzer for hairpin/secondary structure formation, homodimer, and heterodimer, with a maximum ΔG of -4 Kcal/mol-1 for annealing regions. Once all primers met these qualifications they were ordered and tested. Optimization of these primers included multiple gradient and touchdown PCR methods that spanned a variety of temperatures, times, cycle numbers, and additives. Each reaction was screened with an agarose gel at 0.7% agarose. Additives tested were ETSSB, betaine, and 1,2-propanediol. Touchdown PCR was performed to one degree below the calculated Q5 annealing temp using NEB ™ calculator. Gradient PCR was performed from 61°C to 72°C. After analyzing results and getting successful yields and limited off target replication, ideal cycling conditions were recorded and repeated to generate higher yield of fragments. Fragments were then assembled using NEB HIFI assembly, which yielded successful lengths on gel electrophoresis. However, the yield from reactions have been far too small to allow any downstream applications such as sequencing or even transformation, which is difficult with such a large construct. We need to run more assemblies and combine them to reach a yield significant for transformation or sequencing.