Co-culture: It Takes Two

Introduction

Bacterial cellulose (BC) was first discovered in 1838, and since then thousands of potential applications for this material have sprung up, which span from development of prosthetic skins to water purification (1). BC is naturally created by bacteria from the acetobacter family, including the commonly known Komagataeibacter xylinus responsible for fizzy Kombucha drinks. The wide-spread applications of BC are derived not only from the properties of the material, but because of its potential to be functionalized.

However, functionalization of BC through integration of molecules within the fibers of cellulose poses a challenge because of how K. xylinus secretes the cellulose fibers (1). Hence, functionalizing BC must be done as the cellulose fibers are being produced by integrating the desired functional molecules simultaneously. This will be done through combining the bacteria cultures used to create BC and the functional molecules into a single culture to simultaneously synthesize the materials, otherwise known as a co-culture.

Figure 1. K. xylinus secreting cellulose to form fibers of bacterial cellulose.

After the material from the co culture has been synthesized, it must undergo post-production treatments of sterilization, purification, and drying so that it can be safely used as a food packaging material. These stages are what the packaging material industry refers to as post-production treatments, which makes the material more resistant to every-day wear and tear Hence, we set up a series of experiments to determine what method of sterilization, purification and drying of BC would ensure the best customer experience. The degree of customer satisfaction with the packaging would depend on the longevity, aesthetics, transparency, and homogeneity in appearance of the material.

Iterative Design: Feedback From Dr. Burkinshaw on Co-culture

Dr. Brianne Burkinshaw, a professor at the University of Calgary biochemistry department specializing in bacterial secretion systems was a crucial aspect of understanding both the limitations of our protein-producing E. coli, and the need for a co-culture system. When we first approached Dr. Burkinshaw, our initial concern was the inability of E. coli to secrete proteins. Dr. Burkinshaw confirmed that E. coli is a bacteria that has poor secretion abilities, and we would need to move from secreting the recombinant protein as BC is synthesized the protein from E. coli to another method of functionalizing. After this meeting, our team decided to pursue a co-culture model, where E. coli integrates itself within the BC fibers and produces proteins. The proteins will then be released from the cells into the BC when the cells are lysed during sterilization.

In order to lyse E. coli we presented the idea of using an autoclaved further difficulties rose in terms of protein denaturation of nisin. In a subsequent meeting with Dr. Burkinshaw, she advised us to use the pH and thermostability of nisin to its advantage by placing it in a buffer with an ideal pH for Nisin before autoclaving it. This will allow for nisin to retain its activity while being integrated within the BC fibers.

Dr. Burkinshaw further informed our choice of bacterial chassis to produce nisin. Even though nisin primarily targets gram positive bacteria, however it still can affect Gram-negative bacteria if it gets within the periplasmic space and interacts with peptidoglycan, preventing cell replication and promoting cell lysis. So if nisin has the potential to inhibit both the growth of gram positive and negative bacteria, what chassi would our team use? Dr. Burkinshaw informed us that this idea of killing E. coli but the protein remaining can be used to our advantage when functionalizing our bacterial cellulose as we only want our antimicrobial peptide and no bacteria in it.

Co-culture Modelling

To inform how the co culture will be set up, data from a co-culture model will be used to inform how the experiments are set up. Using values derived from research on E. coli and K. xylinus monocultures, the following cellular phenotypes were established using co-culture modeling.

Three-Step Method: Using Results From the Co-culture Model in The Wet-Lab

Three-Step Co-culture Approach: Results

Step 1: Optimizing Media for BC and Recombinant Protein Yield

Figure 2. BC Yield over a period of 3 days in grams at 30oC in a static monoculture. The samples to measure BC mass were derived BC samples that have been blotted with a paper towel and measured on a scale. HS media, LB media, and HS Media enriched with 5 grams of tryptone were tested with BC seeds that all had masses of approximately 0.2 g ± 0.03 g. Data points are the average of one replicate.

Step 2: Impact of Extracellular Secretions on Growth of K. xylinus and E. coli

Figure 3. E. coli OD (595 nm) over a period of 3 days in absorbance 30oC in a static monoculture. The samples to measure OD were derived from media surrounding the BC biofilm. Data points are the average of three biological replicates. Data is represented as an average OD over 3 days.

Figure 4. BC Yield over a period of 3 days in grams 30oC in a static monoculture. The samples to measure BC mass were derived BC samples that have been blotted with a paper towel and measured on a scale. Data points are the average of three biological replicates. Data is represented as mean ± SD.

Step 3: Co-culture Between K. xylinus and E. coli

Figure 5. E. coli OD (595 nm) over a period of 4 days in absorbance 30oC in a static monoculture. The samples to measure OD were derived from media surrounding the BC biofilm. Data points are the average of two biological replicates. Measurements at hour 23 and 71 have been excluded from the data set because of the high SD, indicating that the high degree of biological variance causes this information to be unreliable to base assumptions on and must be omitted. Data is represented as mean ± SD.

Figure 6. BC Yield over a period of 3 days in grams 30oC in a static monoculture (orange line) and co-culture (yellow line). The samples to measure BC mass were derived BC samples that have been blotted with a paper towel and measured on a scale. Co culture BC samples were cultivated in a culture with K. xylinus and E. coli over a period of 4 days in a static culture at 30oC and given 4 mL of HS media every 24 hours. Monoculture BC samples were cultivated in a culture with only K. xylinus over a period of 4 days in a static culture at 30oC and given 4 mL of HS media every 24 hours. Data points of the co culture samples are the average of three biological replicates. Data is represented as mean ± SD.

Intermittent Feeding

Now that the environmental conditions to establish the co-culture in has been determined using the co-culture model and three-step approach, a way to maintain a static environment with constant glucose levels must be established with a feeding strategy.

The rate at which bacterial cellulose (BC) production is directly dependent on the available surface area of the culture media that it is grown in (3). When K. xylinus is grown in a standing culture, the BC film acts as a means to increase the air interface available for the bacterial growth (3), as K. xylinus is an anaerobe. The idea behind the intermittent feeding strategy is to add fresh culture media onto existing BC pellicles to increase the BC thickness while maintaining stable glucose levels.

There were two intermittent feeding strategies, one of which would put fresh media into the cultures every 24 hours, and the other every 48 hours. According to literature, the 24 hour feeding schedule would produce a single pellicle of BC, while the 48 hour would produce multiple thinner pellicles (1). The goal is to determine which feeding strategy would yield the greatest BC growth.

To determine how much media to add for the 24 hour feeding strategy, the following equation was derived from research done by Hsieh et al (2016) which was based on the surface area vs media ratio of the container that the K. xylinus culture was in:

The following equation was then used for the 48 hour feeding strategy:

Intermittent Feeding: 24 and 48 hour Feeding Schedules

Figure 7. BC Yield over a period of 3 days in grams 30oC in a static monoculture. The samples to measure BC mass were derived BC samples that have been blotted with a paper towel and measured on a scale. The 24 hour intermittent feeding sample was given 4.0 mL of HS media at 8:00 am every 24 hours. The 48 hour intermittent feeding sample was given 20 mL of HS media every 48 hours, or on day 2. Data points are the average of three biological replicates. Data is represented as mean ± SD.

Determining Degree of Protein Distribution in BC

In a co-culture system it has been suggested by literature that E. coli integrates itself within the BC fibers (4). To confirm that integration of functional protein through interwinding the protein-producing chassis within the BC fibers is possible, our team modeled the protein production with green fluorescence protein (GFP) producing E. coli. This recombinant GFP E. coli was grown in a co-culture with K. xylinus to visualize the attachment and the effects of sterilization using an autoclave to the protein and BC.

The co-culture system was set up in a petri dish to maximize surface area. After the samples have grown for four days, the sample was autoclaved and then visualized under a UV light.

Figure 8. The co-culture BC was autoclaved. Most of the GFP E. coli seen before autoclaving was in the surrounding edges, and once it was autoclaved most of the fluorescence was in the surrounding outer edges. Once autoclaved the fluorescence was more of a greenish/brown but still remained on the outer edge, which can be assumed to be thermo lysed bacterial cells.

Purification and Drying

Appearance is one of the first things that consumers use to form judgements (5). Therefore, Cellucoat must not only be structurally durable, but also visually appealing. According to Chris Clark, Category Director at Star Produce, customers are visual shoppers. This means that they not only want aesthetically pleasing packaging, but they want their packaging to be transparent enough to allow them to see the quality of the produce they are purchasing.

Therefore, both a purification and drying method yields a material with the greatest homogeneity in appearance and retained strength must be determined. The best purification method was determined by comparing a common BC purification method, 0.125M NaOH, to distilled water and 0.5M NaHCO3 with a subsequent boiling water bath. The best drying method was determined by comparing vacuum, oven, and air drying.

Figure 9. BC grown in HS media and purified in either 0.5 M NaHCO3, 0.125M NaOH, or H2O for three days on stage 2 on the rocker then all air dried for 3 days were cut into multiple 2 mm x 2 cm strips and the transparency was measured at 595 nm with a spectrophotometer. The more transparent the material is, the lower the OD. Data is represented as a mean of 10 replicates, and the SD of the mean is represented as a graph in Figure 12.

Figure 10. The SD of the OD taken from BC grown in HS media and purified in either 0.5 M NaHCO3, 0.125M NaOH, or H2O for three days on stage 2 on the rocker then all air dried for 3 days were cut into multiple 2 mm x 2 cm strips and the transparency was measured at 595 nm with a spectrophotometer. The better the purification method was, the smaller the SD the method will have. Data is represented as a mean of 10 replicates from Figure 11, and the SD of the mean is represented Figure 12.

Figure 11. Images taken over the course of 3 days of BC purified in either 0.5 M NaHCO3, 0.125M NaOH, or H2O for three days on stage 2 on the rocker then all air dried for 3 days.

Figure 12. Time taken for BC that was originally 3 cm thick to dry using oven, vacuum, or air drying. The BC was grown in HS media and purified in 0.125M NaOH for 3 days. The oven drying occurred at 70 degrees celsius, vacuum drying occurred in a vacufuge at 60 degrees celsius, and the air drying occurred at room temperature at 24 degrees celsius. The oven drying consumed 7500 kW, the vacufuge consumed 4125 kW, and the air drying consumed 0 kW.

Figure 13. Images taken over the course of 3 days of BC that were originally 3 cm thick to dry using oven, vacuum, or air drying. The oven drying occurred at 70 degrees celsius and was lubricated with mineral oil, vacuum drying occurred in a vacufuge at 60 degrees celsius, and the air drying occurred at room temperature at 24 degrees celsius. The oven drying consumed 7500 kW, the vacufuge consumed 4125 kW, and the air drying consumed 0 kW.

HP Interview: Tweaking Properties of BC Material for Longevity and Functionality of Antimicrobial Peptide

Another academic human practices who has helped critic our project design and guide us with his expertises is Dr. Hu. Dr. Hu is a professor here at the University of Calgary working at the Schulich School of Engineering. His expertise is in biomass and bioengineered inspired biomaterials. We reached out to him to talk about our biomaterial made in our co-culture- bacterial cellulose.

Dr. Hu took a chemical approach to evaluate our co-culture design. One of the essential details he pointed out to us was bacterial cellulose functional groups interacting with nisin. Bacterial cellulose is made using repeating chains of glucose. This only exposes the hydroxyl (OH) functional group for any further reaction. Dr. Hu believed Nisin and that OH group could possibly react and this would prevent nisin from binding to lipid II and not form a pore- which could lead to a loss of function for Nisin. Cellucoat’s drylab used molecular dynamics to understand and model the interaction between our protein and bacterial cellulose.

Aside from interactions of the end hydroxyl group cellulose impacting the functionality of our chosen antimicrobial peptide, Dr. Hu also told us to consider that BC is susceptible to cellulases produced by E. coli and any other microorganism that is present on produce surfaces.

Determining Efficacy of NaHCO3 Bath on Longevity of BC Material

In lieu of Dr. Hu’s advice, our team looked into how to optimize Cellucoat to be both less prone to degradation of Cellulases and replacing the hydroxyl group to prevent interference of nisin's functionality. Produce packaging has to undergo various environments to get to the final retail destination.

The main culprit of Cellulose degradation are Cellulases, which are present whenever there is a produce substrate that microorganisms can colonize (6). Therefore, to both reduce the likelihood of microbial colonization on the BC material and promote the longevity of the material, the post treatment was tested against cellulases. According to literature, the NaHCO3 purification and boiling water bath replaces the hydroxyl group at the exposed end of the cellulose chain, which is the end that cellulases recognize and bind to, with a HCO3- group (6). Therefore, the substitution of the functional groups on the end of the cellulose chain not only prevents cellulases from degrading the material, but also prevents nisin from binding to the end of the cellulose chain (7).

This experiment was tested through placing BC that was either purified with NaOH, the conventional method to purify BC, or purified with NaHCO3 and boiled in water.

Figure 14. IPicture of experimental set up. 0.5 x 20 mm slip of BC purified by either NaOH or NaHCO3 submerged in a solution 1.5 ml of 0.1 mg cellulase/0.01 g BC solution. The experiment was conducted with 10 replicates. This experiment intends to determine if NaOH (negative control) or NaHCO3 would hold up best in a commercial environment. The samples were placed at 30 degrees celsius in the incubator and agitated twice a day until the sample appeared dissolved in the solution within the tube and the solution was homogeneous. The NaOH sample began breaking down on day 2, while the NaHCO3 sample did not break down.

Conclusion

The purpose of these series of experiments was to prove that using a co-culture is a viable method to functionalize bacterial cellulose for the purposes of produce packaging. Components of the co-culture that include establishing optimal co-culture conditions, BC feeding schedule, purification and drying method, and distribution of recombinant protein within BC has been demonstrated to be possible to achieve.

A dynamic co-culture model was used to inform how to set up each stage of the experiment. Using results from the co-culture model, the conditions were changed so they favored optimizing the growth of the lagging population (K. xylinus) over the dominant population (E. coli). This was established by increasing initial concentration of K. xylinus in initial co culture by 5x more than E. coli, and using HS media and 30oC growth conditions optimized for K. xylinus growth.

The co-culture experimental process is centered around 1) optimizing the media and growth conditions for BC and recombinant protein production, 2) impacts of extracellular secretions on E. coli and K. xylinus growth, and 3) BC and E. coli yield when synthesized in a co-culture. For step 1, the results indicate that using either HS media or HS media enriched with tryptone, to encourage recombinant protein production, will have similar BC yields as there is no significant difference between the results. For step 2, the extracellular secretions increase BC yield by 51.61%, but have no significant impact on final E. coli cell count. For step 3, the final BC yield from the co culture is 27.05% lower than the BC yield from a monoculture.

Lastly, the post production treatments were focused on how to purify and dry the BC to result in a material that was transparent and homogeneous in appearance. The results indicate that a four day 0.5 M sodium bicarbonate purification treatment, 20 minute boiling water bath, and leaving to air dry for two to three days results in a material with the greatest transparency, homogeneity in appearance, and according to uniaxial testing retained strength compared to a 0.125 M NaOH bath.

Future Directions

The co culture experiments were set up using a single strain of E. coli recombinant expressing GFP. In reality, there would be two E. coli strains, one with the biopolymer PHB plasmid and the other with the antimicrobial peptide nisin plasmid. Future experimentation with these two strains of E. coli and K. xylinus in a co-culture will help determine how byproducts of each population’s growth impacts the growth of the other populations, and how each changes the properties of the final BC product.

During the co-culture phase of experimentation, it became evident that over a couple days the degree of GFP expression in BC lagged while E. coli cell density increased. This indicated that the E. coli cells lost their plasmids in response to the absence of antibiotics producing a selective environment. To ensure that the growth of K. xylinus is not impacted by the addition of antibiotics, the K. xylinus used will also be given antibiotic resistance to chloramphenicol, the same type of antibiotic resistance as the PHB and antimicrobial peptide E. coli strains.

Our team will achieve antibiotic resistance in K. xylinus through using the plasmid identified and experimented with from the 2014 Imperial College of London iGEM Team called IBMc396 (8). Using IBMc396 plasmid and electroporation techniques for transforming K. xylinus, our team can ensure that all species within the co culture will have antibiotic resistance to chloramphenicol (8). Beyond ensuring that proteins are continued to be recombinantly expressed, having antibiotics as a component in the co-culture environment also prevents contamination from other microbes, improving the overall safety of Cellucoat as a food packaging material.

Bioreactor

On a lab level, we implemented an intermittent feeding strategy to optimize the reproduction of bacterial cellulose in our co-culture. As proved by our experiments, the 24-hour feeding strategy caused the greatest bacterial cellulose yield. However, manually curating this process would be labour-intensive and time-consuming when taken to an industrial level. Hence, we hypothesized the production of a bioreactor that mimics this intermittent feeding process to grow the co-culture.

This bioreactor would use the fed-batch cultivation principle, which is similar to our 24-hour intermittent feeding principle, as they involve the periodic addition of new media during the bioprocess. This has been shown to prolong the cultivation time and yield of BC. Horung et al. proposed an aerosol bioreactor that uses an aerosol generator that produces a spray of culture media filled by gravity from a feeding tank. This generator feeds new media to the culture box, where the BC is grown for a certain period of time . A distributor was designed and built to allow for homogeneous distribution of the culture media.

Figure 15. Aerosol bioreactor created by Hornung et al.

The design of our bioreactor would be synonymous with the aerosol bioreactor. We would modify this design to avoid some of the system's drawbacks and, secondly, to account for more optimal conditions for E. coli in the culture. A major drawback of the bioreactor was contamination, as the sterilization methods used in the design seemed insufficient. The problem was said to arise from the aerosol generator or the distributor as they cannot be sterilized during production. Hence, we would modify our design to allow these parts to be sterilized to minimize contamination.

To optimize the conditions, we would include parts that can help us moderate environmental factors such as temperature, oxygen and pH so that the co-culture can grow in optimal conditions.

Lastly, although the bioreactor can increase the production yield of BC, we still have to consider the costs of producing it. The capital costs, however, are relatively low as the process it operates on is fairly simple. The most expensive item in the bioreactor is expected to be the aerosol generator. However, to combat this, it is suggested that having a large aerosol generator connected to several culture boxes could reduce the cost.

References

1. Zhong C. Industrial-scale production and applications of bacterial cellulose. Frontiers in Bioengineering and Biotechnology. 2020;8.
2. Bogdanowicz DR, Lu HH. Studying cell-cell communication in co-culture. Biotechnology Journal. 2013;8(4):395–6.
3. (Hsieh, Wang, M.-J., Lai, J.-T., & Liu, H.-S. (2016). A novel static cultivation of bacterial cellulose production by intermittent feeding strategy. Journal of the Taiwan Institute of Chemical Engineers, 63, 46–51. https://doi.org/10.1016/j.jtice.2016.03.020
4. Drachuk I, Harbaugh S, Geryak R, Kaplan DL, Tsukruk VV, Kelley-Loughnane N. Immobilization of recombinant E. coli cells in a bacterial cellulose silk composite matrix to preserve biological function. ACS Biomater Sci Eng . 2017;3(10):2278–92.
5. Bou-Mitri C, Abdessater M, Zgheib H, Akiki Z. Food packaging design and consumer perception of the product quality, safety, healthiness and preference. Nutrition & Food Science. 2020;51(1):71–86.
6. Atalla RH, Isogai A. Enzymatic degradation. Nucleosides, Nucleotides and Nucleic Acids. 2003;22(12):ix-x.
7. dos Santos GR, Soeiro VS, Talarico CF, Ataide JA, Lopes AM, Mazzola PG, et al. Bacterial cellulose membranes as carriers for Nisin: Incorporation, antimicrobial activity, cytotoxicity and morphology. Polymers. 2022;14(17):3497.
8. Team Aqualose 2014 Results. Team:imperial/results. Available from: https://2014.igem.org/Team:Imperial/Results