Design

Introduction

Our project started off with the idea of controlling bacterial gene expression using radiofrequency. The initial plan was to express ferritin in E. coli and then expose the bacteria to radio waves which would then heat up the ferritin (Stanley, 2014). However, the idea that ferritin heats up is controversial (Davis, 2020;Hernandez-Morales, 2020). We turned to utilize magnetic nanoparticles, which have proven to respond to alternating magnetic fields (AMF). Our design to enable E. coli to respond to AMF involves three aspects: building a device that produces AMF, attaching the magnetic nanoparticles, and programming a temperature-sensitive response in the bacteria. We planned to experimentally determine the best option for each of the three aspects separately and then put them together.

Overall Process

While optogenetics give us wireless control over gene expression, it cannot penetrate deep within tissue. To improve on optogenetics, we were first inspired by papers describing how radiofrequency (RF) exposure can elicit gene expression changes in E. coli as well as RF which can alter promoter and protein activities (Said-Salman, 2019; Jeong, 2008; Andreeva, 2018). To control the reaction rate of bovine carbonic anhydrase, Andreeva et al. encased the enzyme in a magnetic hydrogel. Upon stimulation with a RF field of 937 A/m, the hydrogel would heat up and stimulate enzyme activity. Our preliminary idea was to design E.coli to secrete enzymes within a magnetic hydrogel; the activity of those enzymes can then be sequently controlled by RF in a functionalized biofilm. However, we quickly moved on from this idea after finding better options which would allow us to use RF to selectively impact a protein inside of the cytoplasm to mimic the application of optogenetics.

We moved on to our next design which aimed to express ferritin in E.coli to allow for RF control. Ferritin was shown to be able to heat up a TRPV1 calcium channel after exposure to a RF in mammalian cells (Stanley, 2014). The ferritin worked by storing iron oxide nanoparticles in its core; RF stimulates the iron oxide nanoparticles which then heats up the ferritin and its surrounding according to the paper. Our idea was to create a construct that can express the ferritin with a SpyTag and a temperature sensitive enzyme fused with a SpyCatcher. We also planned to test different types of ferritin and study their differences. In addition to ferritin, we also looked at using encapsulins, which can store more iron oxide. The benefit of this design is that it would be self-contained since the bacteria would express everything it needs to respond to RF.

One of the issues with attaching ferritin or encapsulin to enzymes was that they are relatively large proteins; this would limit the movement of the enzyme we want to control. The other major concern was that the ability of ferritin and encapsulin to generate heat was not supported by all literature (Davis, 2020). Furthermore, one paper suggested that ferritin wable to control TRPV1 activity through lipid oxidation rather than temperature changes (Hernández-Morales, 2020). The uncertainties with using ferritin encouraged us to seek out another design.

Instead of using ferritin with contains a magnetic core, we looked toward using iron oxide magnetic nanoparticles (MNPs) directly. Although the exact mechanism for how individual MNP heat up is still controversial, the overall ability has been well documented; for example, cancer hyperthermia uses MNPs to heat and kill cancer cells (Giustini, 2010). We initially considered incorporating magnetosomes as the source of iron oxide MNPs within the E.coli before opting to purchase chemically synthesized MNPs to improve the efficiency (Alphandéry, 2014). The benefit of using magnetosomes was that we would create an independent system within the E.coli that can respond to RF. The major drawback was that the reconstitution of magnetosomes would require a significant amount of work. MNPs is a good alternative because it would save us time from expressing magnetosomes inside the E.coli.

Next, we debated whether to conjugate the MNPs to a temperature-sensitive protein or to the surface of the bacteria. The main issue with attaching MNPs to proteins was that it required us to inject the MNPs into the cells; this would take more effort than surface-attachment and we might significantly kill the cells during the process. Thus, we decided to localize the MNPs around the surface of the E.coli. While iron oxide MNPs can interfere with bacterial membrane integrity, the bacterias do not need to be preserved for a long time for us to prove our concept (Arakha, 2015). Upon magnetic stimulation, the magnetic nanoparticle should then heat up the cytoplasm.

Our next goal was to insert a temperature sensor inside our chassis to induce gene expression changes. We looked into using heat shock promoters, RNA thermometers, and transcription repressors. The problem with using heat shock promoters and RNA thermometers was that the induction fold was very low compared to the repressors (Piraner, 2017). However, we purchased with heat shock promoter and transcription repressor plasmid because they came from the same lab. We purchased a set of temperature sensitive parts from Addgene to confirm and test which parts provided the most robust response to temperature change.

After deciding on using magnetic nanoparticles and our temperature sensors, we moved on to designing a device which can stimulate the magnetic nanoparticles to heat up. We realized, with the help of experts in the field, that the magnetic nanoparticles do not technically respond to radiofrequency but rather alternating magnetic field (AMF) (Christiansen, 2017). AMF is generated when an alternating current is passing through a coil. By applying AMF with amplitudes of intensities in the tens range (kA/m) and frequencies in the hundreds range (kHz), we can induce sufficient hysteresis losses in the MNPs to alter local temperature.

Finalized Version

Device

Our device is meant to produce a magnetic field of 100-500 kHz to excite our magnetic nanoparticles (MNPs). It is a simple circuit composed of an amplifier, oscilloscope, function generator, electromagnetic core, capacitors, transformer, shunt resistor, vector board, and wire. While our design works in theory and in the literature, we were unable to get ahold of an amplifier with sufficient power output to produce a strong enough magnetic field for our design (Christiansen, 2017).

The device, when functional, should produce a magnetic field of 100-500 kHz to excite our magnetic nanoparticles (MNPs). We were unable to produce the full circuit due to some shipping delays and inability to access an amplifier of the appropriate power (at least 50 W but preferably 100 W). Once the circuit was built, we would then experiment with the MNP’s to determine their exact rate of heating when exposed to the magnetic frequency. With this, we could then develop an interval system in order to allow for heating while exposed to the frequency and cooling while removed from the frequency to control the temperature of the media between a small range (<5 degrees celcius). Once these parameters had been tweaked and roughly developed, we would then apply it to our MNP-cell mixtures to demonstrate that our cell system does, in fact, respond to a “wireless” electromagnetic frequency.

Attachment

The magnetic nanoparticles (MNPs) we used are iron oxide nanoparticles which have superparamagnetic characteristics. To localize the MNPs to the surface of the bacteria, we came up with three choices including using electrostatic, antibodies, and streptavidin/biotin interactions.

Electrostatic

Through functionalizing the MNPs with positively charged amine or chitosan, the particles would interact with the negatively charged bacterial surface (Huang, 2010; (Le, 2020). This method is simple and efficient to implement. The disadvantage is that it is nonspecific and will bind to any negatively charged surfaces.

Antibodies

The specific antibodies which binds to E. coli surface proteins would be fixed to the MNPs. Compared to using electrostatic interaction, antibodies would allow more specific attachment. However, the antibodies will still bind to other bacteria which also express the epitope.

Streptavidin/biotin

The E. coli would be first transformed to express streptavidin with Lpp-OmpA, allowing streptavidin to localize at the surface. The MNPs would be biotinylated and form an almost covalent bond with the streptavidin. This would allow for the most specific interaction. However, the streptavidin may already be saturated with biotin by the time it is expressed to the surface. In addition, it is more difficult and time-consuimg than the other two methods.

Temperature Sensor

After exposing the MNP-bacteria conjugate to AMF, the cytoplasm temperature of the bacteria should increase as the MNPs heat up. We found three ways to program a temperature sensitive response in E. coli which would regulate the expression of genes. For our case, we decided to control the expression of mWasabi, which is a green fluorescent protein similar to GFP.

RNA thermometers

RNA thermometers prevent translation of mRNA into protein. At high temperature, its hairpin structure “denatures” and exposes the binding site for the small ribosomal subunit (Narberhaus, 2006). We found, however, using RNA thermometers does not produce a desirable high induction fold.

Heat shock promoters

The plasmids we used express heat shock promoters upstream of the mWasabi genes. An increase in temperature induces endogenous heat shock proteins to bind to the heat shock promoters and upregulates mWasabi.

Temperature-sensitive transcription factors

We tested TlpA and TcI, which are repressors that respond to 45°C and 40°C, respectively (Piraner, 2017). At lower temperatures, these transcriptional repressors are in a dimerized state, preventing RNA polymerase from initiating transcription. When the temperature rises, they both undimerize and allow expression. Based on our results, the heat-sensitive transcription factors showed greater induction versus the heat shock promoters. However, the heat shock promoters more consistent induction of gene expression in the cell culture whereas TlpA and TcI worked in scattered cells. For future designs, we would choose between using heat-sensitive dimers and heat shock promoters based on what type of response is needed. For a robust output, heat-sensitive dimers would be used. For outputs that require a higher cell percentage to produce the output, heat-shock promoters would be used.

We also implemented plasmids with constitutive RFP expression to test the relative mWasabi expression levels. For these constructs, each cell will both constitutively express RFP and mWasabi in response to thermal induction. This allows us to measure levels of fluorescence using a plate reader, microscopy, flow cytometry. Experiments were performed in triplicate to verify validity. This can be found in the results section.

Future

We hope to implement our design further as a potential therapeutic. One technological challenge we faced was the concept of localized heating. We hope to improve our design in the future so that our AMF-producing device could selectively heat one bacteria or acutely heat one cytoplasm region so that our technology can match the spatial control of techniques like optogenetics. We also aim to identify and expand the types of temperature-sensitive parts that we could use as well as design bio-circuits using these parts. For instance, we could design a chassis that produces one gene product at a specific temperature and another gene product at another temperature.

Although we moved on from our ferritin design earlier, we still hope to explore genetically encoded MNPs like ferritin. The current drawback is that the mechanisms governing how ferritin responds to magnetic frequencies are unclear. However, this technology is still promising and should be explored. For instance, bacteria that express ferritin are responsive to magnetic fields (Aubry, 2020).

References

Alphandéry, Edouard. “Applications of Magnetosomes Synthesized by Magnetotactic Bacteria in Medicine.” Frontiers in Bioengineering and Biotechnology, vol. 2, 2014. Frontiers, doi:10.3389/fbioe.2014.00005.
Andreeva, Yulia I., et al. “Enzymatic Nanocomposites with Radio Frequency Field-Modulated Activity.” ACS Biomaterials Science & Engineering, vol. 4, no. 12, Dec. 2018, pp. 3962–67. ACS Publications, doi:10.1021/acsbiomaterials.8b00838.
Arakha, Manoranjan, et al. “Antimicrobial Activity of Iron Oxide Nanoparticle upon Modulation of Nanoparticle-Bacteria Interface.” Scientific Reports, vol. 5, no. 1, Oct. 2015, pp. 1–12. www.nature.com, doi:10.1038/srep14813.
Aubry, Mary, et al. “Engineering E. Coli for Magnetic Control and the Spatial Localization of Functions.” ACS Synthetic Biology, Sept. 2020. ACS Publications, doi:10.1021/acssynbio.0c00286.
Christiansen, Michael G., et al. “Practical Methods for Generating Alternating Magnetic Fields for Biomedical Research.” Review of Scientific Instruments, vol. 88, no. 8, Aug. 2017, p. 084301. aip.scitation.org (Atypon), doi:10.1063/1.4999358.
Davis, Hunter C., et al. “Nanoscale Heat Transfer from Magnetic Nanoparticles and Ferritin in an Alternating Magnetic Field.” Biophysical Journal, vol. 118, no. 6, Mar. 2020, pp. 1502–10. ScienceDirect, doi:10.1016/j.bpj.2020.01.028.
Giustini, Andrew J., et al. “MAGNETIC NANOPARTICLE HYPERTHERMIA IN CANCER TREATMENT.” Nano LIFE, vol. 1, no. 01n02, 2010. PubMed Central, doi:10.1142/S1793984410000067.
Hernández-Morales, Miriam, et al. “Lipid Oxidation Induced by RF Waves and Mediated by Ferritin Iron Causes Activation of Ferritin-Tagged Ion Channels.” Cell Reports, vol. 30, no. 10, Mar. 2020, pp. 3250-3260.e7. www.sciencedirect.com, doi:10.1016/j.celrep.2020.02.070.
Huang, Yan-Feng, et al. “Amine-Functionalized Magnetic Nanoparticles for Rapid Capture and Removal of Bacterial Pathogens.” Environmental Science & Technology, vol. 44, no. 20, Oct. 2010, pp. 7908–13. ACS Publications, doi:10.1021/es102285n.
Jeong, Mi-Jeong, et al. “Plant Gene Responses to Frequency-Specific Sound Signals.” Molecular Breeding, vol. 21, no. 2, Feb. 2008, pp. 217–26. Springer Link, doi:10.1007/s11032-007-9122-x.
Le, Thao Nguyen, et al. “A Convenient Colorimetric Bacteria Detection Method Utilizing Chitosan-Coated Magnetic Nanoparticles.” Nanomaterials, vol. 10, no. 1, Jan. 2020. PubMed Central, doi:10.3390/nano10010092.
Narberhaus, Franz, et al. “RNA Thermometers.” FEMS Microbiology Reviews, vol. 30, no. 1, Jan. 2006, pp. 3–16. academic.oup.com, doi:10.1111/j.1574-6976.2005.004.x.
Piraner, Dan I., et al. “Tunable Thermal Bioswitches for in Vivo Control of Microbial Therapeutics.” Nature Chemical Biology, vol. 13, no. 1, Jan. 2017, pp. 75–80. www.nature.com, doi:10.1038/nchembio.2233.
Said-Salman, Ilham H., et al. “Global Gene Expression Analysis of Escherichia Coli K-12 DH5α after Exposure to 2.4 GHz Wireless Fidelity Radiation.” Scientific Reports, vol. 9, no. 1, Oct. 2019, pp. 1–10. www.nature.com, doi:10.1038/s41598-019-51046-7.
Stanley, Sarah A., et al. “Remote Regulation of Glucose Homeostasis in Mice Using Genetically Encoded Nanoparticles.” Nature Medicine, vol. 21, no. 1, Jan. 2015, pp. 92–98. www.nature.com, doi:10.1038/nm.3730.

Team:Pittsburgh/Design