Overview

The biological toolkit provides a diverse, adaptable interface for molecular targets. Disease biomarkers are often tractable targets for biosensing due to their natural compatibility with biological recognition elements. Engineering cell-based biosensors that utilize natural receptor-ligand interactions or synthetic circuits can enable detection of disease biomarkers with high specificity and sensitivity [1-3].

While these sensing circuits are rapidly emerging for a wide range of substrates, there are still significant hurdles to overcome to enable deployment in diagnostic devices. One main challenge is preserving whole-cell biosensors for use over time in non-ideal settings. Hydrogels are emerging as a favored platform for keeping biosensors alive and responsive in deployable devices [4] but adapting this to different lab spaces and use-cases remains a challenge.

Building on the work of BostonU iGEM 2024, we sought to join a growing effort to address the challenge of biosensor deployment through suspension and fluorescence detection in agar hydrogels.

Agar is a polysaccharide mixture derived from seaweed cell wall with favorable gelling properties, thermostability and biocompatibility [5]. Important for our consideration was that agar is a common microbiology laboratory resource– allowing our methods to be easily adapted to other labs for other biosensor deployment applications. Additionally, the gelling and biosensor resuspension procedure is simple and does not require any specialized equipment.

The use of agar does pose unique challenges for using whole-cells biosensors. First, agar’s thermostability means that it must be heated to high temperatures to become liquid and allow cell resuspension. Second, optical density is not a usable metric for cell growth due to interference from the 3D polymer structure. And third, we have no standard for measuring fluorescence from a 3D printed device.

To address these challenges we developed a protocol for identifying conditions and calibrating measurements to standardize agar hydrogel use through a three step process of optimization, normalization, and deployment. This process can be replicated for future projects, offering a low-barrier entry point to whole-cell biosensor device integration.

Strain and Preliminary Validation

Our agar testing was carried out with NEB® 5-alpha E. coli constitutively expressing eGFP under the control of the pVeg1 promoter (strain hereafter referred to as M0065). This is a strong promoter that resulted in visibly green colonies and produced a strong signal for testing.

To ensure that heating the agar would not impact our results significantly, we first tested the effect of resuspending in liquid LB media at different starting temperatures. We showed that there were no significant growth differences in liquid culture over the 40 hours before the death phase, and a 22% reduction in peak fluorescence around 30 hours. The reduced fluorescence indicates there is some metabolic effect of heat shock, but only after a long incubation phase in liquid culture which is not ideal for maintaining activity long term.

Optimizing Agar Conditions

The first step in adapting a strain for use in hydrogels is to determine the optimal agar conditions for measuring output.

Overnight cultures of our strain were refreshed 1:100 and incubated to mid-log phase at 37C in a shaking incubator. Solutions of LB agar with increasing concentrations of agar (expressed as weight / volume percentages) were made by diluting a 5% LB agar stock solution with liquid LB media. Solutions were kept at 50C in 50 mL tubes in a heat block to keep them liquid. 5mL aliquots of bacterial culture were centrifuged at 4000 RPM for 5 minutes to pellet the cells and the media was poured off. The pellets were resuspended in LB agar solutions by pipetting up and down with a serological pipette and immediately plated in a 96-well black-walled clear-bottom plate (200 uL working volume was maintained from liquid culture experiments). The plate was kept at room temperature in a plate reader, measuring eGFP fluorescence every hour over a weekend. Our fluorescence measurement settings were kept the same as what we used for liquid media, but increasing the gain may provide better results for higher concentrations of agar.

For M0065, we found that GFP was maintained over the 60 hours after peaking just before 24 hours. The higher percentage agar solutions had even lower fluorescence which may have been due to growth hindrance or interference with light propagation due to the denser structure. In general, the signal was markedly lower than GFP fluorescence in liquid media. In this case the signal is still strong, but the loss is a key consideration for downstream implementation, especially with weak or low dynamic range circuits. A lower concentration 0.5% agar solution was tried in a previous experiment and also performed suboptimally. In another test, 1% and 1.5% solutions had almost no difference in signal strength so 1.5%, typical for agar plates, was used as it was already on hand in the lab [See ‘Making solid and liquid media’ in our experiments page].

Validating Fluorescent Imaging

For biosensor deployment, we adapted the 3D printed disk from BostonU iGEM 2024 for larger agar hydrogels (the STL files can be found in our team’s gitlab). Our disk was 3D printed with the Formlabs 4 grey resin v5. Measuring fluorescence from the disk cannot be done in the plate reader and requires a fluorescence imager. To ensure consistency and enable comparison between our plate reader-obtained values and disk measurements, we compared the two devices with a set of diluted M0065.

M0065 was grown to mid-log phase as before. The culture was transferred to culture tubes in 4, 3, 2, and 1 mL aliquots. The cells were pelleted in a benchtop centrifuge at 4000 RPM for 5 minutes and the LB media supernatant was poured off. At the same time, 1.5% (w/v) LB agar was prepared by microwaving and adding antibiotics, and then the molten gel was held at 50C until use.

Each pellet was resuspended in 5 mL of 1.5% LB agar to create diluted solutions of fluorescent cells. The resuspended cells were carefully pipetted to a 96-well plate in triplicate, covered with the lid and incubated at room temperature overnight.

In the morning, the plate was measured on the plate reader with standard settings and in the iBright FL1500 Gel imager. The 96-well plate was placed upside down in the imager for the best image results. Fluorescent images were taken using the smart exposure setting and the images were exported as TIFF files. The intensity from the triplicate wells was measured by taking the average pixel brightness per well in Fiji, an open source image analysis software. This can be done by circling the well with the ellipse tool then selecting Analyze -> Measure (or cmd + M/ctrl + M).

To ensure the gel imager produced consistent results with the plate reader for future testing and comparison, we performed an orthogonal distance regression to obtain a linear function relating the two datasets. Fitting the function I_GFP = a x P_GFP + b, where I_GFP and P_GFP are the GFP intensities measured on the gel imager and plate reader respectively, yielded a = 9.9 ± 0.15, b = 410 ± 31, and R² = 0.993. The normalized residuals had a mean of 0.008 and a standard deviation of 0.134, indicating this model is highly accurate and even slightly overestimates the error.

These results validate the use of fluorescent measurements with our gel imager, allowing us to calibrate our expected results between measurements and enable comparison in the future.

Deployment

Finally, to demonstrate the measurement for a deployable device we 3D printed a disk with seven wells for hydrogels. Following the same procedure for making culture dilutions as before, we resuspended and plated M0065 to fill the seven wells, spreading across two disks with two wells per dilution. The hydrogels were left covered overnight at room temperature and measured in the gel imager in the morning. The images were exported as TIFF files and analyzed in Fiji.

Due to the top-down images in the wells (as opposed to the flat, bottom view of the 96-well plate) there was light concentration along the well edges likely caused by the meniscus effect. To measure consistently across conditions, the average pixel intensity was taken from an oval encompassing only the top half of the well which was well-illuminated in all conditions.

Despite the lighting effects, the results of the measurements between the 96-well plate and the 3D printed disk were very consistent. The data showed the same behavior and had only a small change in magnitude and standard deviation for which could be attributed to the difference in volume, well size, or oxygen availability. This kind of validation is important for testing out new 3D printed devices to ensure behaviour is consistent with what is expected. Setting up this framework and validation allowed us to move forward with our project

Final Thoughts

We show a simple, reproducible method for adapting strains for use in inexpensive agar hydrogels and provide a detailed protocol for validating measurements across equipment. These resources enable future teams to quickly implement agar hydrogels and build on our work, unlocking new deployment potential for whole-cell biosensors across all field applications without a high barrier to entry.

References

[1] Hicks, M., Bachmann, T. T., & Wang, B. (2020). Synthetic Biology Enables Programmable Cell-Based Biosensors. ChemPhysChem, 21(2), 132–144. https://doi.org/10.1002/cphc.201900739

[2] Wu, Y., Wang, C.-W., Wang, D., & Wei, N. (2021). A Whole-Cell Biosensor for Point-of-Care Detection of Waterborne Bacterial Pathogens. ACS Synthetic Biology, 10(2), 333–344. https://doi.org/10.1021/acssynbio.0c00491

[3] Checa, S. K., Zurbriggen, M. D., & Soncini, F. C. (2012). Bacterial signaling systems as platforms for rational design of new generations of biosensors. Current Opinion in Biotechnology, 23(5), 766–772. https://doi.org/10.1016/j.copbio.2012.05.003

[4] Tang, T.-C., Tham, E., Liu, X., Yehl, K., Rovner, A. J., Yuk, H., de la Fuente-Nunez, C., Isaacs, F. J., Zhao, X., & Lu, T. K. (2021). Hydrogel-based biocontainment of bacteria for continuous sensing and computation. Nature Chemical Biology, 17(6), 724–731. https://doi.org/10.1038/s41589-021-00779-6

[5] Martínez-Sanz, M., Ström, A., Lopez-Sanchez, P., Knutsen, S. H., Ballance, S., Zobel, H. K., Sokolova, A., Gilbert, E. P., & López-Rubio, A. (2020). Advanced structural characterisation of agar-based hydrogels: Rheological and small angle scattering studies. Carbohydrate Polymers, 236, 115655. https://doi.org/10.1016/j.carbpol.2019.115655