Results | ASU - iGEM 2024

FUexp In-Vivo Characterization

New Composite part FU1

Successful Cloning of FUexp

Below are the results of our successfully transformed construct, FUexp, which was assembled into a backbone containing Kan resistance, T7 terminator, sfGFP, and an origin of replication.

Verifying the Construction of FUexp with Sequencing

We sequenced the plasmid to verify that our constructed plasmid was the same as what we had designed it to be. Our sequencing results can be found in Figure 2. They indicate no mismatches between the expected FUexp plasmid and our constructed plasmid. Because our cloning was successful, we were able to move forward with optogenetic experiments to further test this construct.

Figure 2. Results of FUexp (ideal) and FUexp (sequenced) show that there are no mismatches.

Testing FUexp in Optogenetic Experiments

After confirming the sequencing results, our next step was to test the UirS/UirRs optogenetic controls under inducing and repressing conditions: UV violet and green, respectively. We conducted experiments using UV violet light, green light, and darkness and measured the gene expression of sfGFP after each.

FUexp Fluorescence Measurements

Figure 3: FUexp construct under green light (repressing), UV light (inducing), and darkness. The dotted horizontal line represents the baseline darkness condition. All measurements were adjusted for fluorescence/OD600 and normalized to darkness. * Paired t-test p-val < 0.05

Successful Fluorescence Results: UV Violet and Green Light as Activation and Repression Wavelengths

Fluorescence readings in this experiment showed differing levels of activation and repression. UV violet light yielded the most sfGFP gene expression. Total darkness yielded the second most sfGFP gene expression. Green light yielded the least sfGFP gene expression. In the scope of our experiment, the data shows green light is a more effective repressor than darkness.

New Findings for UirS/UirR: Green Light is a More Effective Repressor than Darkness

Based on the literature, green light was cited to be an effective repressive condition, (Ramakrishnan, Tabor, 2016), where sfGFP levels from green light to UV violet light increased 4.41 fold. However, in our latest experiment, sfGFP levels from green light to UV violet light increased only 1.15 fold. Possible experimental setup choices such as light contamination may have skewed these results.

Future Plans for FUexp: Building Multi Control Systems for Gene Expression

In future experiments, we hope to better characterize the activity of UirS/UirR over a longer period of time and determine the switch-off time between UV violet to green. To do this, we will monitor expression time more closely, such as 10-minute intervals, to get the most precise results on the switch-off time. We will also construct sturdier, light-insulated boxes for our incubator instead of using cardboard boxes. We will also explore constructs with promoters similar to those used in published results for improved off switching of gene expression in green light using PACE.
Despite some of the unpredicted results, we maintain that UirS/UirR has the potential to be a powerful optogenetic tool for gene monitoring. For one, experimentation with the UirS/UirR construct revealed two distinct repression conditions: darkness and green light. This allows the genetic circuit to offer scientists a wider range of tunability, with three expression conditions, which can’t be replicated with the chemical inducers. Secondly, because cyanobacteriochromes offer such a broad range of wavelengths, this project can be built upon in the future by combination with other optogenetic switches like CcaS/CcaR for greater control over gene expression in a multiplexed manner. By refining our optogenetic control, we hope scientists can use this tool for protein engineering and PACE experiments.

Validation of HIV-RT functionality interference with AlphaFold3

For this aspect of the project, we utilized experimentally validated interactions between HIV-RT thumb domains and endogenous human protein eEF1A as a basis to demonstrate the efficacy of our models. Having achieved this, we were able to quickly screen and generate hundreds of viable peptides with our workflow, and we chose some examples below to demonstrate the success of our project.

Figure 4: AlphaFold3 predicted structure of the full HIV-RT heteroligomer fused with sfGFP, interacting with some of our designed peptides. (A) pepMDM interacts with the thumb domains of both p51 and p66 (B) pepQ30 interacts with the thumb domains of only the p66 subunit, and it forms a homoligomer (C) pepM8P interacts with both thumb domains (D) AlphaFold 3 correctly predicts interactions between eEF1A and thumb domain of HIV-RT (E) Introducing pepMDM into the system perturbs and removes eEF1A from both thumb domains (F) pep Q30 binds to the p51 thumb domain, and eEF1A no longer binds to either thumb domain (G) pepM8P binds to both thumb domains, and eEF1A is disjoint from the complex

In the above images, we show the peptides (in green) interacting with the thumb domains of HIV-RT p51 and p66 (Fig 4A-C). Furthermore, we show the correctly predicted interaction of HIV-RT thumb domains and eEF1a (Fig 4D), but the introduction of our peptides prevents this interaction (Fig 4E-G).

The above structural predictions utilize the recommended method from the authors of AlphaFold3, which outlines the steps they took to achieve higher prediction accuracy and resolution than AlphaFold2, used in the generation of peptide candidates. AlphaFold3 has recently been used to generate protein binders for novel therapeutic targets. Because AlphaFold3 is separate from AlphaFold2, while still recapitulating its predictions, we can say with some confidence that this structural prediction is a separate validation of our workflow.

In future steps, we hope to perform experimental assays to validate the results of our modeling and give further credence to the work that we have accomplished this year.

Validation of Hardware Functionality

Figure 5: This is a graph of optical density data taken from our turbidostat during an experiment that was run using the LCD and rotary encoder to initiate the Run PACE experiment. The graph starts increasing exponentially, showing bacterial growth; when it hits the set point (mid-log growth), the values level off and remain constant. After the media ran out at approximately 13 hours into the experiment, the optical density values became erratic; however, as this is after the media ran out, it is not relevant to the system’s functionality but was included for transparency.

Experimental Validation of Turbidostat

Our final graph and several previous experiments demonstrate our system's ability to maintain optical density at a specific set point. These results are explained in detail along with the remaining experiments we performed on our Hardware page. The ability to maintain a bacterial culture at a specific set point means the user is able to move bacteria into the lagoon when they are the most healthy, at mid-log growth which is essential for ensuring successful PACE experiments.

Documenting Hardware for Reproducibility

All of the physical hardware and software necessary to run these experiments is extensively documented for potential users to construct, run experiments, and build upon our work. We created modifiable 3D models; annotated versions of our code; shared instructions for system assembly, circuit diagrams, and video demonstration; and designed a printed circuit board. Not only are the physical hardware and software reproducible but so are the experiments as we have step-by-step instructions on how to set up each experiment and helpful tips along with information for potential troubleshooting.

Contributing to the Accessibility of Protein Engineering

The goal of our hardware, to create a functional turbidostat that can maintain an optical density setpoint, did not change throughout the course of the project. These results demonstrate our success in achieving this goal. Additionally, they prove the possibility of making a functioning turbidostat for under $200, that is user-friendly and easy to assemble. Ultimately, our hardware contributes to the overall goal of our project to make protein engineering more accessible.