Team:Athens/Engineering
ENGINEERING SUCCESS
Introduction
This project proposes the manufacturing of a biomaterial that can be created with the tools of synthetic biology and have the following desired properties :
- 1. Preservation of the optical properties of the Flavobacterium species used (i.e. structural coloration).
- 2. Adequate mechanical strength for the various applications it will be used for.
- 3. Ability to be isolated from the rest of the system.
Research was required to find the material that satisfied the above requirements. Additionally, these properties had to be estimated with modelling techniques and subsequently be tested with adequate experimental protocols.
The design process of our project consists of different parts so including them in one page was deemed inefficient. As such, we will briefly describe each design step and provide hyperlinks to their more detailed forms on our Wet Lab and Dry Lab tabs. Therefore, herein we describe engineering success both in our dry lab work and on the work we would have done at the bench, if we had early access to the lab.
Research
According to Taxonomy biologyGluconacetobacter xylinus, Komagataeibacter xylinus and Acetobacter xylinum all refer to the same organism. Also, Gluconacetobacter hansenii was formerly known as Gluconacetobacter xylinus.
In order to design a biomaterial with the desired properties mentioned at the Introduction, our research evolved on different topics. To begin with, we focused on cellulose producing organisms and mainly on Komagataeibacter xylinus and Gluconacetobacter hansenii. We also studied how the genes responsible for cellulose production in those organisms have been transferred to others [1]. In addition, we familiarized ourselves with structural colour in bacterial colonies [2].
After searching through existing literature for ways to simulate the structural colouration of the biomaterial, we opted for COMSOL simulations as found in the literature [3,4]. These research papers propose a way of approximating the interaction of light with a specified periodic geometry and showed potential when it came to comparing the modelling’s output with experimental data. Therefore, we will design and build our system based on this novel technique.
Detailed information about the biomaterial extraction time, as well as the exact amount produced, can provide valuable information about its properties. Based on that, a kinetic model of cellulose production, preferably for a solid culture, was sought after in the literature. Some research articles propose a macroscopic kinetic model based on mass balances which is opted in settings such as bioreactors and big liquid cultures [5]. Other models were created by fitting a considerable amount of experimental data but were also based on a liquid culture and the fundamental differences between our biological systems may have rendered the use of those kinetic equations inaccurate [6,7]. Lastly, existing literature also proposes a kinetic model based on gene expression of the bacterial cellulose operon and growth kinetics. Moreover, the model was describing the secretion of bacterial cellulose from genetically engineered E. coli,which is even closer to our genetically engineered Flavobacteriia system. Based on those observations, a kinetic model that describes both gene expression and growth kinetics was to be designed.
Our proposal
Having read through a sizable amount of existing literature, the steps to be followed had been clearly defined: Since it is possible to transfer cellulose genes to E. coli, our team decided to try to genetically modify Flavobacterium johnsoniae to produce and excrete cellulose, which will retain the properties listed in the Introduction.
Predicting the optical properties of the biomaterial
Design→Build→Test
IIt is established that a medium-biofilm system such as agar-Flavobacteriia can display iridescence. In order to quantify this ability we observed the spatial arrangements of bacterial colonies and established some key observations that can be linked to the optical properties. This design process was based on [3] and can be found on our Background Page.
The observations about the spatial arrangements of the cells can be used to build a simplified geometry and simulate its interaction with light. This was done using COMSOL, a Finite Element Method (FEM) modelling program. We created three different simulations: one with Flavobacteriia and an agar medium, one with Flavobacteriia and a cellulose-based domain and lastly one with only the cellulose based domain where Flavobacteriia were replaced by air. Every simulation will provide different results and insight as to how to improve estimations.
Figure 1. shows the reflectance data for wavelengths between 300 nm to 900 nm. The averaged results have been normalized.
Figure 1. Reflection-wavelength plots from left to right: (a) agar medium with the presence of Flavobacteriia, (b) cellulose domain with the presence of Flavobacteriia and (c) cellulose domain with the absence of Flavobacteriia.
Insight & Possible Refinements
Based on the derived results some interesting comments can be made. Figure 1a. shows the estimated optical response of Flavobacteriia [3]. Firstly, it is apparent that the geometry reflects only certain wavelengths and with a very strong, relative intensity. Secondly, orientation does seem to play a significant role in the reflection of specific wavelengths because the resulting reflectance spectra for each angle (from −30° to + 30°) differed considerably. Johansen et al [3], points that two diffraction spots one at 350 nm and one at 530 nm are observed. The simulation has slightly different peaks one at 380 nm and one at 520 nm while also showing considerable peaks at 540 nm and 830 nm. These differences may be attributed to the assumptions of the geometries.
Figure 1b. also showcases strong reflections for select wavelengths but appears to have a broader range of reflected wavelengths. It should be noted that while the results were normalized, the absolute values of reflection were on average bigger than those from the agar-Flavobacteriia geometry. This may be attributed to the distance travelled by light due to the different refractive index of the domain.
Lastly, Figure 1c. contains the optical results of cellulose without any Flavobacteriia. It is estimated that all wavelengths will be reflected by a considerable amount. This leads to some interesting conclusions: an acellular biomaterial based on cellulose does not seem to retain the structural colour of Flavobacteriia colonies. Additionally the predicted optical result will include all visible wavelengths.
Even though this simulation is generally quite innovative, there is always room for improvement. One important addition to the simulation would be to create different domains with different hexagonal periodicities and averaging the results based on experimental data. Another improvement in the results would be to produce a weighted average of all the domain’s orientations since it is possible that bacteria colonies show a preference in orientation. Unfortunately, no experimental data were available for Flavobacteriia strain. A simple way to tackle this issue would be to generate cross-cut TEM images of our Flavobacteriia colonies and perform image analysis to extract the needed distributions. This was of course not possible due to the health crisis and its implications to our project.
Regarding the cellulose domains, additional assumptions were made. A more detailed geometry could be designed based on TEM images of an intermediate product containing cellulose and bacterial colonies. Moreover, the assumption that Flavobacteriia will be replaced with air, in regards to the acellular biomaterial, is not proven and other unpredictable phenomena might take place that will influence the optical properties of the end-material, such as remainder of bacterial cell walls after decellularization treatment.
Estimating the cellulose production over time
Design→Build→Test
While ensuring whether or not cellulose will be produced is one problem, the rate at which this happens is also something we needed to take into consideration. By having estimations about the production time, we can predict the quantity of the biomaterial and link it to some key mechanical properties. Moreover, it can be the basis for the scaling up process needed in order for the material to be mass-produced. Thus a kinetic model that incorporates both mRNA and protein expression, as well as growth kinetics has been designed. The model was inspired by [8] and can be found on ourBackground Page.
The kinetic model was built by solving the ODEs mentioned above using MATLAB. The values of the parameters have been specified in the literature [8] for a genetically engineered E. coli strain. These values are accrued from fitting experimental data to the ODE system. However, due to the lack of lab access, we were not able to replicate the experiments for our system and as such we did not have data to fit the parameters values to the system and the experimental conditions. The results shown below are based on indicative values for both the initial conditions and the parameters based on what is provided by the literature and should not be used as guidance for our system. In the “insight” section, we will discuss the potential of optimizing the kinetic model to our system.
Figure 2. shows the results of Glucose, Biomass and Cellulose concentrations over a time period.
Figure 2. Glucose, biomass and cellulose concentrations (mg/L) over a time period.
Insight & Possible Refinements
The results indicate that the kinetic model is capable of producing useful time graphs. However, it is evident that in order to produce quantitative estimations the parameters values should be chosen based on experimental data regarding our system.
Based on the experimental values, the system can be optimized by using mathematical tools provided by MATLAB. The end result should better describe the expected outcome when it comes to cellulose production.
Plasmid design and experimental procedure
Design→Build→Test
In order for cellulose genes, consisting of the BC synthase operon (bcsA, bcsB, bcsC, bcsD) and their upstream genes (cmcax, ccpAx), to be inserted into F. johnsoniae, they were designed into BioBricks and underwent codon optimization, as well as removal of the illegal sites for restriction enzymes. Those genes are incorporated in the pHimarEm1 plasmid, using the Type IIS assembly standard, as it is described in the Project Design. The designed plasmid is later inserted into competent Escherichia coli through electroporation and then the cells that have received the plasmid are selectively cultured in petri dishes, selecting for antibiotic resistance.
Cellulose genes are transported from E. coli to F. johnsoniae with conjugation in liquid culture. Cellulose genes are then transposed to the genome of Flavobacterium species, by way of the Mariner transposon. Induction of cellulose genes takes place and structurally colored cellulose is expected to be produced from F. johnsoniae. As a final step, cellulose is isolated, dried and used as colour coating for different surfaces and applications. This process is described in detail at the Project Design pagepage.
The produced cellulose is finally separated from the medium and treated as described in [9]. To dry the final product, the room temperature drying method was followed.
Due to the ongoing global health crisis, our team did not have access to the desired strains, until mid-September 2020, as well as adequate lab supplies. Therefore, our experiments were only designed and could not be carried out. Out of all steps described above, we were only able to culture Flavobacterium johnsoniae according to [1] and order the genes and plasmids we designed. In order to enhance the contrast of the emergent colour, a stain was added to the nutrient solution where the bacteria are grown.
Figure 3. Flavobacterium johnsoniae IR1 M16 culture.
Since we were expecting to be able to get into the lab earlier during the year, we had prepared a series of expected results, possible problems and proposed solutions for each step of our experimental procedure, in order to test and refine our project design.
For the insertion of the designed BioBricks into pHimarEm1 plasmid, the Type IIS assembly standard would have been used. The plasmid would later be inserted into competent Escherichia coli through electroporation and then the cells that would have received the plasmid would be selectively cultured in petri dishes, through Kanamycin resistance that is provided to plasmid bearing cells. In case no E. coli cells grew, it would be an indication that the electroporation had not worked, or that the cells were not competent. Again, the culture conditions and protocols would have to be investigated. In addition, we would have to inspect whether the size of the plasmid would inhibit the plasmid insert and its maintenance.
Cellulose genes would later be transported from E. coli to F. johnsoniae with conjugation in a liquid culture. Then, plasmid bearing Flavobacterium species would have been selectively grown on a culture medium that contains Erythromycin antibiotic, since the plasmid provides only Flavobacteriia cells with Erm resistance. As a result, a structurally coloured biofilm would be expected to grow on the petri dishes, like in the first step, with more variation in the colours that would have been observed. In case cells would not grow, the reason might be failed conjugation or due to the fact that cells are in a later growth phase than intended.
The procedures would have been improved until the expected optical result occurs. Again, the size of the plasmid might be the reason why the plasmid is not inserted successfully. To inspect the conjugation process, we would check some parameters that Dr Ingham, who regularly performs this process, suggested. For example, he informed us that the process is very sensitive to the growth rate/state of both partners, the presence of calcium, and high quantity of agar. Also, according to Dr Ingham, it is likely that the polymer (carrageenan, starch etc.) in the recovery plates has a big effect on motility/cell organization mutants and it is important to maintain the antibiotic selection pressure (Erm) on all mutants.
Cellulose genes would have been transposed to the genome of the Flavobacterium species and cells that incorporate those genes would have been selectively grown on petri dishes.
IInduction of cellulose genes would then take place and structurally colored cellulose would be expected to grow from F. johnsoniae. In this step it is possible that no cellulose or not enough cellulose is produced. There is also a possibility that the produced cellulose is not colored. In any of those cases, the optimization of cellulose production will be implemented according to literature [11-15]. Also, the reason why cellulose does not grow, might be due to possible mistakes on the designed BioBricks. To examine that case, the DNA would be sent for sequencing.
As a final step, cellulose is isolated, dried and used as colour coating for different surfaces and applications. Different drying methods described in the literature [9,10] (Room Temperature Drying, Freeze Drying, Supercritical CO2 Drying and Oven Drying) are used for drying the final product, in order to test which one interferes less with the structure of cellulose and allows the product to appear coloured.
[1] Buldum, G., Bismarck, A., Mantalaris, A. Recombinant biosynthesis of bacterial cellulose in genetically modified Escherichia coli.
[2] Johansen, V. E., Catón, L., Hamidjaja, R., Oosterink, E., Wilts, B. D., Rasmussen, T. S., Sherlock, M. M., Ingham, C. J., Vignolini, S. Genetic manipulation of structural color in bacterial colonies
[3] Kientz, B., Luke, S., Vukusic, P., Péteri, R., Beaudry, C., Renault, T., Simon, D., Mignot, T., Rosenfeld, E. A unique self-organization of bacterial sub-communities creates iridescence in Cellulophaga lytica colony biofilms
[4] Siddique, Ρ. Η., Diewald, S., Leuthold, J., Hölscher, H. Theoretical and experimental analysis of the structural pattern responsible for the iridescence of Morpho butterflies.
[5] Ismail, F. M., Dissanayake D. M. S. C. Mathematical Modeling Of Bacterial Cellulose Production By Acetobacter Xylinum Using Rotating Biological Fermentor.
[6] Hornung, M., Ludwig, M., Gerrard, A. M., Schmauder, H.P. Optimizing the Production of Bacterial Cellulose in Surface Culture: Evaluation of Substrate Mass Transfer Influences on the Bioreaction (Part 1)
[7] Hornung, M., Ludwig, M., Gerrard, A. M., Schmauder, H.P. Optimizing the Production of Bacterial Cellulose in Surface Culture: Evaluation of Substrate Mass Transfer Influences on the Bioreaction (Part 2)
[8] Buldum, G., Argyro Tsipa, A., Mantalaris, A. Linking Engineered Gene Circuit Kinetic Modeling to Cellulose Biosynthesis Prediction in Escherichia coli: Toward Bioprocessing of Microbial Cell Factories.
[9] Zeng, M., Laromaine, A., Roig, A. Bacterial cellulose films: influence of bacterial strain and drying route on film properties.
[10] Vasconcellos, V. M., Farinas, C.S. The Effect of the Drying Process on the Properties of Bacterial Cellulose Films from Gluconacetobacter hansenii
[11] Hornung, M., Ludwig, M., Gerrard, A. M., Schmauder, H.P. Optimizing the Production of Bacterial Cellulose in Surface Culture: Evaluation of Substrate Mass Transfer Influences on the Bioreaction (Part 3)
[12] Krystynowicz, A., Czaja, W., Wiktorowska-Jezierska, A., Gonçalves-Miśkiewicz, M., Turkiewicz, M., Bielecki, S. Factors affecting the yield and properties of bacterial cellulose
[13] Krasteva, P. V., Bernal-Bayard, J., Travier, L., Fernando Ariel Martin, F. A., Kaminski, P. A., Karimova, G., Rémi Fronzes, R., Ghigo, J.M. Insights into the structure and assembly of a bacterial cellulose secretion system
"Monarch Butterfly Wing Scales" by David Illig is licensed under CC BY-NC-SA 2.0
