Team:Stanford-Brown-RISD/Results

The results of our substrate tests indicate that the mycelium can grow on any organic material––even Martian regolith mixed with supplemental nutrients. We did find, however that there were factors that impacted the success of mycelium colonization. First the moisture level greatly impacted how quickly the mycelium could grow and whether it entered a dormant state or not(entering it if the material is too dry). Second the particle size also impacted colonization rate; with particles that were too fine or too large being harder to colonize than particles of a medium size.

Temperature Tests

Based off the reuslts of our temperature test (see the tables and graphs below), we concluded that the optimal growth temperature (based on the temperature settings available to us) was 30°C. Based on Our results also allowed us to conclude that at temperatures close to 4°C and 37°C growth is limited. Below are images of the dishes used to grow the dikaryion strain at 24°C, 30°C, and 37°C over the course of 9 days.


Image 1: Time-lapse of dikaryon, Ganoderma Lucidum growth at 4 different temperatures	Image 2: Time-lapse of monokaryon, Ganoderma Lucidum growth at 4 different temperatures.


Table 1: dikaryon, Ganoderma lucidum mycelia area growth at 4 different temperatures over 10 days	Table 2: monokaryon, Ganoderma lucidum surface area growth at 4 different temperatures over 12 days


Graph 1: Surface Area of dikaryon, Ganoderma Lucidum mycelia growth over 4 different temperature conditions	Graph 2: Surface Area of monokaryon, Ganoderma Lucidum mycelia growth over 4 different temperature conditions


Graph 1: Linearization of Surface Area of dikaryon, Ganoderma Lucidum mycelia growth at 24°C and 30°C	Graph 2: Linearization of Surface Area of monokaryon, Ganoderma Lucidum mycelia growth at 24°C and 30°

Growing Mycelia in Molds

Using the procedure listed above combined with the results of the previous tests we were able to produce several bricks and small pieces with different material substrates. Images of these bricks, pieces, and even a stool, can be found below.


Stool after 2 weeks of growth, prior to being baked	Team member Emilia Mann sitting on the stool after it has been baked.	Pile of bricks produced with mycelium and yard waste, wood chips.	Team member Javier Syquia holding the first prick produced by the team!

Mycelium Glue

Figure 1: Lap Shear Adhesive Strength. Different bioadhesive candidates where applied to mycelium samples that were pressed together and allowed to set. Using the Instron 5565, the force necessary to separate the mycelium samples was then measured and converted to kPa accounting for the area of overlap between the mycelium samples.

As shown in Figure 1, all of our purified bioadhesive proteins successfully bound to baked mycelium substrate and produced non-negligible adhesive strengths. The CsgA-CBD and CBD4x fusion proteins were the most successful, with both reaching strengths of 28.2 kPa which is comparable to the tested strength of Elmer’s glue on mycelium at 32.1 kPa. Non-fusion CsgA was slightly weaker at 27.1 kPa on, and fp151 gave the poorest adhesive strength on mycelium at 16.8 kPa.

On a cardboard substrate (Figure 2) there are several notable differences from the mycelium substrate testing. Elmer’s Glue was again the strongest and more than doubled its bonding strength from the mycelium substrate test, to 78.1 kPa. CBD4x was surprisingly stronger than on mycelium, improving to 46.5 kPa. The fp151 protein also more than doubled in strength, reaching 40.6 kPa in the cardboard testing. Csga-CBD was weaker in the cardboard testing, at 22.7 kPa. The sample of csgA bound to cardboard was unfortunately dropped and was not salvageable for lap-shear testing.

Figure 2: Relative Mycelium Adhesive Strength. After testing the adhesive strength of each bioadhesive candidate on mycelium, we tested each candidate again on cardboard. We then took the ratio of adhesive strength on mycelium to adhesive strength on cardboard to determine the relative efficiency of the two substrates.

By taking the ratio of adhesive strength on mycelium to the adhesive strength on cardboard, we can find the relative efficiency of the adhesives on the two substrates. Every adhesive except for csgA-CBD bonded stronger on cardboard substrate than the mycelium substrate. This could be due to cardboard having a more uniform surface and thus allowing better molecular bonding and glue function. This makes it very notable that csgA-CBD was a stronger adhesive on mycelium, and provides evidence that by adding the chitin binding domain to csgA we were able to successfully make the protein a mycelium-specific glue. CBD4x, while still performing better on cardboard, did not decrease in efficiency as much as Elmer’s glue or fp151 which shows, albeit less dramatically, that chitin binding domains can be functionalized as mycelium specific bioadhesives.

Once we identified csgA-CBD as our best mycelium specific bioadhesive, we collaborated with the DTU-Denmark iGEM Team to test our glue on their mycelium grown from a different fungus. The DTU-Denmark Team’s mycelium was grown from the fungus Aspergillus oryzae while ours was grown from Ganoderma lucidum. When we performed a lap shear test using our csgA-CBD glue on DTU’s mycelium samples that were shipped to us in the mail, the maximum adhesive strength reached was 6.7 kPa. While this was still significant adhesion, it was lower than the 28.2 kPa reached with Ganoderma lucidum. The difference in adhesive strength, however, provided useful information about the variability in surface composition of mycelium grown from different species of fungus.

Conclusion

In total, from the lap shear testing data it can be concluded that the fusion proteins that were designed for this project, CBD4x and csgA-CBD, successfully act as biological adhesives for mycelium materials. By outperforming fp151, which has been reported to be an extremely strong bioadhesive[9], and almost reaching the adhesive strength of Elmer’s glue both csgA-CBD and CBD4x show promise in future use as structural binders of mycotecture in Mars or other extraplanetary colonization efforts. As Elmer’s glue requires raw materials like natural gas and petroleum that are plentiful on Earth but are often rare to non-existent on other planets, the slight adhesive advantage it shows in comparison is greatly outweighed by the mass producibility of our bacterially synthesized bioadhesives[13]. In further research, the characterization of alternate csgA-CBD fusion proteins implementing different spatial conformations and/or multiples of csgA of CBD polypeptides could yield even better mycelium-specific bioadhesives. The inclusion of a third type of polypeptide in the fusion protein could theoretically create multipurpose mycelium-specific glue proteins, able to bind mycelium together while also performing some other useful task such as molecular sensing.

Biofiltration: From Concept to Prototype

ITC Analysis: Data, Modeling, and Binding Parameters

The results from the ITC experiment to determine binding parameters for peptides 1x-, 2x-, and 3x-HHTC-Re alongside the predicted structure (QUARK Ab Initio program) follows [2]. This program uses Monte Carlo simulations and knowledge of atomic force fields to construct the most probable structural conformation of a protein from just the linear amino acid sequence.

Figure 3: ITC data for the three HHTC-Re peptides. The binding affinity (K_a) was assessed to determine the strength of the interaction, and the results follow: 1x-HHTC-Re K_a = (1.55 ± 0.21) x 10⁶ M^-1; 2x-HHTC-Re K_a = (3.73 ± 0.53) x 10⁵ M^-1; 3x-HHTC-Re K_a = (1.50 ± 0.05) x 10⁵ M^-1. Figure design credits to Jesica Urbina.

The binding affinity values in Figure 3 are largely comparable to one another, and we do see a linear trend between the number of HHTC-Re repeats and injections taken to reach saturation, which means that each peptide can bind a linearly increasing amount of copper proportional to the HHTC-Re repeats. This validates our hypothesis underlying our construct design for our fusion proteins, allowing us to bind multiple metals on the same biomolecule. In Figure 4, we can see how the specific subunits maintain proper folding (predicted by QUARK), supporting the notion that they would retain their original functions.

Figure 4: Quark ab initio model of 2x-HHTC-Re with chitin binding domain. Domains within the fusion protein have been annotated to display the conformation and spatial orientation.

After confirming the copper binding of the individual HHTC-Re x n peptides, we then needed to assess whether our fusion protein could bind copper and chitin, and whether it could do so when already saturated with the other substrate. Figure 5 depicts two experiments a) Raw data and b) isotherm for 2x-HHTC-Re-CBD + NaDg and Cu. In this experiment, N-acetyl D-glucosamine (“NaDg”; analogous to a chitin monomer and widely used in the literature for assessing chitin binding) was first titrated into 2x-HHTC-Re-CBD and no isotherm was calculated because binding sites were not saturated by the ligand. Cu was then titrated into the 2x-HHTC-Re-CBD + NaDg complex and this resulted in a Cu affinity at K_a= 7.61 +/- 1.49x 10⁶ M^-1 that is largely comparable to 2x-HHTC-Re (no CBD) and lowerby an order of magnitude than 2x-HHTC-Re-CBD without bound NaDg. Data show 20 1 µL injections. 2x-HHTC-Re-CBD was selected as the candidate for testing because it displayed the most consistent and strong results during protein purification procedures, and seemed most promising for downstream applications (such as our filter). Work is in progress on optimizing the 3x and 6x fusion proteins and creating prototypes (thus far, protein production as been successful).

Figure 5: Isotherm and data produced by ITC for assessing the binding affinity of our fusion protein (2x-HHTC-Re-CBD) for chitin (represented by N-acetyl D-glucosamine) and Cu.

Tangential Flow Prototype Creation: Material Properties and Design Considerations

The first consideration when thinking about creating a mycelium biofilter had to do with the qualities and characteristics of the mycelium material. Would it be waxy? Would the chitin be exposed? Such questions and more could significantly impact the approach we picked for functionalizing any material produced. The first test we did was comprised of a simple dye being poured on a piece of mycelium to observe the hydrophobicity of the material and whether it could be easily penetrated (as seen in the video below).

It is evident that the mycelium surface is highly hydrophobic. While this may appear to be an undesirable trait for a water filter, we found that subsequent incubation of mycelium in moving water for a period of 48 hours allowed for the material to become permeable. In an interesting twist of synergy, the fusion protein we were attempting to produce was highly hydrophobic as well, and therefore probably showed good affinity for the mycelium material (Figure 6).

Figure 6: Biochemical properties and hydropathy of a CBD-HHTC-Re fusion protein. The fusion protein displays significant hydrophobic tendencies. While this was initially thought to be a problem, it synergized well with the nature of the mycelium surface and most likely allowed the protein to obtain a proximity close enough for chitin binding to occur. The simulation to produce the graphs above was done by LifeTein.

We then also produced RFP-CBD fusion protein following a regular His-Tag purification protocol and analogous transformation/cloning steps as our other fusion proteins to stain mycelial material with, and as a cool visual tool/probe. The following images depict purified protein, and liquid cultures of T7 E. coli cells producing the protein, respectively.

The next factor to contemplate was the manner in which the filter could be implemented. Our first thought was to grow fungus inside a syringe on wood chips and treat the mycelium with our fusion protein, and then run water through this. We discovered that the prototypes created, however, were not very permeable, and applying pressure to push water through washed out the mycelium itself, leaving only substrate behind (as seen in the following two images).

The approach that we selected in the end was to make use of filter-based design in which we could take advantage of tangential flow as opposed to direct head-on flow. The following images show the mycelium strips being incubated with purified CBD-HHTC-Re 2x protein for 72 hours with DTT for cleavage (the protocol we designed for creation of the filters). These were realistically implemented by inserting a thin sheet of glass fiber at the bottom of a syringe to prevent any mycelial material from flowing through. The mycelium material could also be shredded to increase surface area for absorption of metal and display of chitin. The strips were then added on top and water could be flown through this setup. The filter strips were cut into 1 cm x 1 cm squares for prototype testing. Various iterations of this process can be seen in the images that follow.

Tangential Flow Prototype Testing: Bulk Surface Adsorption Using Phen Green

The Cu concentration in the initial copper solution was 325 (+/-25) µM Cu. After 30 minutes of tangential flow, interestingly, the untreated mycelium absorbed about 23% of the copper in solution, revealing a fascinating synergy in that the mycelium possesses some inherent metal sequestration properties. The treated mycelium (filter prototype) was able to sequester ~92% of the available Cu in solution - validating its efficacy and potential utility.

Figure 7: Cu (µM) remaining in solution for n = 3 samples after incubation for 30 minutes while undergoing tangential flow on a flatbed shaker (figure design credits: Jesica Urbina). Experimental conditions were maintained as described previously.

We also tested the filters over a period of 72 hours and measured the copper concentrations at the end in each of the three experimental trials. The filter was able to bind nearly all of the copper in solution to almost undetectable levels, and the plain mycelial material displayed remarkable properties as well.

Figure 8: Cu (µM) remaining in solution for n = 3 samples after incubation for 72 hours while undergoing tangential flow on a flatbed shaker (figure design credits: Jesica Urbina). Experimental conditions were maintained as described previously.

Summary and Conclusion

We were able to rationally design, produce, prototype, and extensively characterize our vision for a mycelial water filtration and biomining system that is fueled by synthetic biology and protein engineering. We created tangible filter strips to demonstrate how our methodology could feasibly be applied in actuality, going beyond just a proof-of-concept phase. These prototypes were able to bind >92% of copper in solution (at >300 μM Cu concentration) within 30 minutes. They are relatively inexpensive, completely biodegradable with ease of metal extraction, and require low, less complex upmass in comparison to bacterial or cellulose-based filters (which require complex sugars for sustenance). A patent disclosure was submitted to NASA on 10/16/18. Our results demonstrate the utility of our end-to-end filtration system in solving water and biomining-related challenges both on earth and in space.

Mission Architecture

Using 425 ml of Anabaena liquid culture (@ end, OD600 is 1.449; OD750 is 1.035), oxygen production went from .8% to ~5% in 60 seconds. Then, oxygen gained 2.7% in 73,560 seconds (with an almost perfectly linear increase).
Given the empirically established rate, it will take ~92 hours (4 days) to achieve 20% oxygen, mimicking Earth's atmosphere. This time-frame is certainly reasonable in the context of our mission architecture. Therefore, we have empirically quantified oxygen production for mycelial growth and astronaut sustenance, demonstrating promising results from cyanobacteria alone!

References

[1] Elmer’s frequently asked questions | how to remove glue. (n.d.). http://elmers.com/about/faqs/general
[2] Zhang et al. (2015). Integration of QUARK and I-TASSER for Ab Initio Protein Structure Prediction in CASP11. Proteins: Structure, Function, and Bioinformatics, 84, 76–86. doi:10.1002/prot.24930