Background
To transform the electromagnetic signal into thermal signal, we need to set magnetic nanoparticles in E.coli, which can absorb energy and heat, just like receiver to receive electromagnetic signal. We decided to use iron contained ferritin as the endogenous nanoparticle.
Ferritin is composed of 12 or 24 subunits, and in coming together they form a virtually spherical and hollow shell with an internal cavity that can accommodate up to 4500 iron atoms[1]. The iron is stored in the form of ferrihydrite iron cores normally that with superparamagnetic properties [2].
Superparamagnetic nanocrystals have the ability to serve as colloidal mediators for heat generation, so it is the ideal property we want to identify in our bio-mineralized bacteria[3]. Actually, the iron contained ferritin has been reported to generate heat in response to electromagnetic signal. [4]
Besides, ferritin is an outstanding choice because of many advantages such as instability, safety for chassis and easiness of magnetic core synthesis rather than magnetosome and metal binding peptide.
We chose FtnA, ferritin from bacteria, which composed of 24 identical subunits Ftn [1] to serve as mag-receiver. It has a suitable size of 12 nm in diameter and its central cavity is around 7.5 nm in diameter (Fig.1) [5]. We’ve built a model to predict the heat producing efficiency of ferritin FtnA.
Process
To get a FtnA ferritin with superparamagnetic nanocrystals and make this core heat, we need to do following process:
To see more details about each process, please turn to protocol page.
Parts Construction
We asked Berlin for the parts BBa_K1438000, BBa_K1438001, BBa_K1438025, BBa_K1438022, BBa_K1438027, BBa_K1438028, BBa_K1438031. However, we failed to transform the plasmids into E.coli, so we constructed parts to express ferritin on our own. Ferritin ftnA gene was obtained by PCR of genomic DNA using Berlin’s primers. Then, we assemble ftnA gene with PT7 and PT5, name the circuits are PT5-ftnA, PT7-ftnA respetively.
Mag-attraction Test
After the bio-mineralization, we did an intuitive mag-attraction test as a pre-experiment to observe the effect of bio-mineralization.[7]And we found that mineralized E.coli gathered at the edge of the magnet, indicating that we had successfully made our E.coli have magnetism properties, which might be of relevance with the iron contained ferritin as we supposed. To verify such assumption, we did following test.
(a)magnet placement under dish during the test.
(b)mag-attraction result of mineralized bacteria harboring targeted gene.
(c)mag-attraction result of bacteria harboring empty vector without mineralization.
To see more story about this amazing discovery and what measurement method we have developed from this, please view the Captor page
Ferritin Expression Test
We purified the ferritin over expressed in E. coli through Ni-chelating affinity chromatography and highly concentrated it and then do SDS-PAGE. The position of clearly targeted band (19.5kDA[1]) on the gel was consistent with the size of ferritin monomer fused with His tag, explaining that the ftnA expression is successful.
Transmission Electron Microscopy (TEM) Test
After successfully verify ferritin expression in our bacteria, we attempted to observe the result of bio-mineralization with TEM as most literatures do[6].
(a)Induced and mineralized cells harboring PT5-ftnA.
(b)Uninduced and mineralized cells harboring PT5-ftnA.
(c)Mineralized cells harboring empty plasmid back bone.
The image(c) shows that there are no dark blocks in E .coli. But unexpectedly clear dark blocks both appeared in image(a) and (b).
We assumed that the dark blocks in image(b) might be resulted from expression leakage, the modeling proved this assumption. You can see more on modeling page.
Later we did energy dispersion spectrum analysis for dark blocks.
(a)Induced and mineralized cells harboring PT5-ftnA.
(b)uninduced and mineralized cells harboring PT5-ftnA.
(c)Mineralized cells harboring empty plasmid back bone.
The data show that iron exists in all the tested bacteria sample. So we needed another more ideal method to prove that our bacteria was successfully mineralized.
Native PAGE Test
In most occasions, TEM test is the most assurance test, and most literature verify their in vivo mineralization by this test. However, TEM test is quite expensive, and when we doing this test, we found that it’s hard to figure more detailed structure. Consequently, we can’t verify the iron core in ferritin through TEM test. So we explored a new method to verify the iron core synthesized by ferritin after mineralization.
From literature for in vitro [8], we learned that Native PAGE is one of the most powerful techniques for studying the composition and structure of native proteins, since both the conformation and biological activity of proteins remain intact during this technique.
Two native gels will be prepared, one of them is stained by potassium ferrocyanide (K4Fe(CN)6) to determine the mineralized ferritin, the other is regularly processed by Coomassie brilliant blue(CBB), aiming to mark the locus of ferritin as the cross-reference with the previous one.
Ferric iron can readily react with the ferrocyanide in acidic condition, resulting in the formation of a bright blue pigment namely prussian blue. [10]The chemical reaction could indicate the existence of the iron core because the color of bands would be blue when the iron core exist in ferritin cavity.
Gel was stained with (A) potassium ferrocyanide and (B) Coomassie Brilliant Blue R250. lane 1, concentrated ferritin which purified and concentrated from induced and mineralized cells harboring PT5-ftnA (marked as group 1) ; lane 2, bacterium sediment of group 1; lane 3, concentrated ferritin which purified and concentrated from Induced and unmineralized cells harboring PT5-ftnA(markerd as group 2); lane 4, bacterium sediment of group 2.
We did Native PAGE using concentrated ferritin which from 900mL bacterium suspension. The gradual darkening of blue color indicates the increase in iron contents in ferritin FtnA. In brief, we successfully verify the iron core.
Comparable protein-staining bands from lane 1 to lane 4 with native FtnA protein cage indicate that the protein structure of FtnA shells is likely intact after the mineralization process. [11]
After verify the existence of iron core, we did SQUID test to characterize magnetic properties of iron core.
Magnetic Properties of Magthermo coli
(a)Low‐Temperature Magnetization Curves(5K)
(b)Enlarge figure of the left figure, Low‐Field Magnetization Curves
The magnetization curves dosen’t saturate even when the field is 3000mT, indicate that there are antiferromagnetic biomineral in E.coli which is hard to get saturated magnetization curves, such as ferrihydrite.
The coercivity is 7.6 mT at 5K, which is generally similar to the coercivity(9.2 mT) of magnetoferritin reported by Cao et al , 2010,JGR [12], indicating there are nanoparticles similar to magnetoferritin in E.coli.
(a) Magnetization Curves at 300K
(b) Enlarge figure of the left figure
Most proteins in E.coli are antiferromagnetic, therefore, the Magnetization Curves at 300K shows antiferromagnetism in general: magnetization decrease with the increasing of magnetic field. However, there is a saturated magnetization curve between ±200mT which is the feature of ferromagnetic mineral. The curve closed and saturated at 100mT, suggesting that the ferrimagnetic minerals: magnetite or magnetic hematite exist.
(a) Decay curves of saturation isothermal remanent magnetization(IRM) acquired in a 2.5 T field after zero‐field cooling (ZFC) and field‐cooling (FC) treatments.
(b) Normalized IRM acquisition and DC (Direct Current field) demagnetization (DCD), measured at 5 K.
The thermal decay of isothermal remanence acquired in a field of 2.5 T after cooling in ZFC and a 2.5 T FC was measured. The 2.5 T ZFC and FC curves show a repid decrease of remanence with increasing temperature，(Fig.12.) and the remanences are almost zero, suggesting most of the magnetic particles are superparamaneti，c particles. The ZFC and FC curves are nearly superimposed suggesting that ferrimagnetic particles exist in E. coli.
The intersection of saturation isothermal remanent magnetization and DCD is below 0.5, indicating magnetic mineral particals aggregated and formed magnetic interactions in E. coli. IRM and DCD could be divided into two parts: the saturated part and unsaturated part. The saturated part is around 200mT, suggesting that ferrimagnetic minerals: magnetite or maghemite exist.
After verify the superparamagnetism property of iron core, we finally did heating test.
Alternative Magnetic Field Heating Detection
We used Optical Fiber Thermometer to detect the effect of heating mineralized ferritin in vivo and in vitro.
The end of optical fiber was immersed in the sample to measure the temperature change. Mineralized sample and control were in the fixed position of the solenoid of electromagnetic field generator (strength of electromagnetic field is 8 mT the frequency of it was 540 kHz).
The temperature curve didn’t show apparent difference between mineralized sample and control. Dr. Xiaowen Wang explained that it is hard to observe the heated effect of ferritin solution. This protocol would be more promising if ferritin was high-throughout prepared, concentrated and freeze-dry into powder.
Future Work
The circuit PT5-FtnA-RNAT has been constructed and we would try to detect the whole cell heating again.
There are some methods to detect the local heating, such as using fluorophores as molecular thermometers, the fluorescence intensity was recorded with the Electron Multiplying Charge Coupled Device(EMCCD). [13]
Valuing Protein Concentration per cell
To supply our modeling with concentration of ferritin per cell, we explored a method for valuing protein concentration per cell, and tried to make it more convenient.
After biomineralization, we take a little bacterium that resuspended by PBS, and obtain the concentration of bacterium by Hemocytometer Counting, which is marked as C1(cells/ml). The other bacterium suspension is disrupted completely. After centrifugation, the suspernatant containing ferritin is used to do SDS PAGE, and then the optical density of the protein sample is obtained by using the gel imaging analysis system, then we can calculate the concentration of the sample, which is marked as C2(μg/ml).
The protein concentration per cell is marked as C, the formula as follows.
C=C2/C1 (μg per cell)
To obtain C2, we did SDS PAGE as follows. (Note: the SDS PAGE should be repeated for several times to get more accurate data.)
Lane 1, 0.5mg/ml BSA; Lane 2-4, ferritin suspernatant; Lane 5, suspernatant of bacterium containing empty plasmid; Lane 6, 1mg/ml BSA; Lane 7, 0.25mg/ml BSA; Lane 8, 0.1mg/ml BSA.
We use bovine serum albumin (BSA) at differernt concentrations to establish the standard curve that reveals the relation of optical density and protein concentration.
Firstly we used the gel imaging analysis system to establish the standard curve and directly calculated the concentration of samples. But the error was so big, that meant the arithmetic of the software couldn’t meet us need. So we established a new standard curve by simply changing the formula. Two new standard curves calculated from two SDS-PAGE pattern are as follows:
To our surprise, the relation found from new formula doesn’t show the linearity we supposed before. Besides, we found that using gel imaging analysis system to analyze optical density has poor precision that the results have big difference between each other. We speculate it’s probably due to the grey level image processing by the system.
Unfortunately we still can’t conquer the problem that the targeted protein band is hard to measured if the band color is too deep. The stability of the results also troubled us. We decided to try Matlab to analyze our data for future work.
References
[1] Arosio P, Ingrassia R, Cavadini P. Ferritins: a family of molecules for iron storage, antioxidation and more[J]. Biochimica et Biophysica Acta (BBA)-General Subjects, 2009, 1790(7): 589-599.
[2] Papaefthymiou G C, Viescas A J, Devlin E, et al. Electronic and magnetic characterization of in vivo produced vs. in vitro reconstituted horse spleen ferritin[C]//MRS Proceedings. Cambridge University Press, 2007, 1056: 1056-HH03-27.
[3] Fortin J P, Wilhelm C, Servais J, et al. Size-sorted anionic iron oxide nanomagnets as colloidal mediators for magnetic hyperthermia[J]. Journal of the American Chemical Society, 2007, 129(9): 2628-2635.
[4] Stanley S A, Sauer J, Kane R S, et al. Remote regulation of glucose homeostasis in mice using genetically encoded nanoparticles[J]. Nature medicine, 2015, 21(1): 92-98.
[5] Wang Q, Mercogliano C P, Löwe J. A ferritin-based label for cellular electron cryotomography[J]. Structure, 2011, 19(2): 147-154.
[6] Wang Q, Mercogliano C P, Löwe J. A ferritin-based label for cellular electron cryotomography[J]. Structure, 2011, 19(2): 147-154.
[7] Lu Y, Sun M. A simple method for constructing magnetic Escherichia coli[J]. bioRxiv, 2014: 010249.
[8] Cai Y, Cao C, He X, et al. Enhanced magnetic resonance imaging and staining of cancer cells using ferrimagnetic H-ferritin nanoparticles with increasing core size.[J]. International Journal of Nanomedicine, 2015, 10(default):2619-34.
[9] Massé E, Gottesman S. A small RNA regulates the expression of genes involved in iron metabolism in Escherichia coli[J]. Proceedings of the National Academy of Sciences, 2002, 99(7): 4620-4625.
[10] Zhang W, Zhang Y, Chen Y, et al. Prussian blue modified ferritin as peroxidase mimetics and its applications in biological detection[J]. Journal of nanoscience and nanotechnology, 2013, 13(1): 60-67.
[11] Cai Y, Cao C, He X, et al. Enhanced magnetic resonance imaging and staining of cancer cells using ferrimagnetic H-ferritin nanoparticles with increasing core size[J]. International journal of nanomedicine, 2015, 10: 2619.
[12] Cao C, Tian L, Liu Q, et al. Magnetic characterization of noninteracting, randomly oriented, nanometer‐scale ferrimagnetic particles[J]. Journal of Geophysical Research: Solid Earth (1978–2012), 2010, 115(B7).
[13] Huang H, Delikanli S, Zeng H, et al. Remote control of ion channels and neurons through magnetic-field heating of nanoparticles[J]. Nature nanotechnology, 2010, 5(8): 602-606.
