SEED COATING & APPLICATION
Introduction
After customising and producing our bacterial cellulose (BC), we set out to use it for real-world applications. However, this requires us to first develop our own coating protocol as traditional coating equipment was outside of our reach. We did this by using our stakeholders’ expertise. We first contacted Dr. Zohaib Hussain , an expert in bacterial cellulose production, who pointed us to two main methods to achieve a bacterial cellulose (BC) coating: ex situ and in situ application. We quickly realised that the ex situ method may be preferred for larger, upscaled efforts, which is why we based our entrepreneurial plan on this, however the in situ application was favoured for the short-term developments as we quickly had results for this. Additionally, in situ coatings have the benefit of maintaining material properties that may be lost during ex situ processing. We used Komagataeibacter sucrofermentans, which we used to optimise BC production and customise its properties, for these coatings.
After coating our seeds, we aimed at tackling two real-world issues that we identified by talking to stakeholders: seed predation and parasitic weeds. Seed predation occurs largely because of beetle (Coleoptera) larvae, particularly those from the click beetle family (Elateridae). This is where our project comes into play. By developing protein attachment through cellulose-binding domains (CBDs), we attached Bacillus thuringiensis (Bt) protein Cry3Aa. Bt proteins are popular biological alternatives to traditional chemical pesticides1. Upon ingestion, these Bt proteins are activated through proteolytic cleavage and bind to specific receptors belonging to their target taxa - making them highly specific, and reducing the risk of potentially affecting other species (Figure 1). We used this CBD-Cry3Aa on mealworms, also belonging to the order Coleoptera, as a model organism. Here we studied the effect of the fusion protein on mealworm starvation and seed predation.
In our second use-case, we aimed to tackle Striga hermonthica. This is parasitic weed of major agricultural crops in sub-Saharan Africa (sorghum, millet and maize) (Figure 2), and it accounts for losses of up to 8 billion US$ annually, affecting over 100 million people2. Striga can reduce crop yields by up to 80%3, and due to its long seed longevity contaminated fields can no longer be used for cereal crops4. With increasing global temperatures this threat is expected to migrate North5, making it a more pressing issue than ever before. Stakeholders at Wageningen University and Research in collaboration with the Dutch Institute for Ecology (NIOO-KNAW) have identified a substance (henceforth "Compound X") that can reduce the Striga threat. Due to publication issues, we are currently not allowed to specify the exact nature of the compound, nor its mode of action. The main issue with this compound is that currently, exogenous application through spraying is unaffordable for the average sub-Saharan farmer. This is why we sought to increase the efficacy of the compound by extending its duration in the soil through using our seed coat as a slow-release system. To this end, we tested whether the compound could be successfully incorporated and diffused out of the seed coating and have started working on greenhouse pot assays.
Experiments
In situ seed coating protocol development
Before being able to use our functionalised materials for any seed coating application, we first had to create a protocol to get our BC around the seeds. We did this through in situ coating, which implies that seeds are coated in the same reaction as where the BC is produced 6.
We used Sorghum bicolor var. Teshale seeds for our coating experiments going forward. The Teshale cultivar is susceptible to Striga hermonthica 7, the parasitic weed that we want to battle with our seed coating applications.
We established an in situ seed coating method by modifying a protocol from6 , to our needs. However, before we could continue, the seeds had to be free from contamination so the K. sucrofermentans is not hindered in its BC production. We sterilised the seed surface by using a combination of 70% isopropanol-alcohol (IPA) and 10% sodium hypochlorite washing steps.
After we sterilised the seeds, we wanted to increase the surface area for BC and its producer, K. sucrofermentans, to stick to it, in hopes of increasing the efficiency of our in situ coating protocol. We did this by mechanically scarifying our seeds by scratching them with a scalpel.
After this scarification step, we wanted to make sure that seeds were ready to receive a BC coating. One of the ways we ensured that BC was formed around the seeds was by using polydopamine-HCl (PDA). PDA can polymerise into a black polymer that glues the K. sucrofermentans to the seed surface. This ensures that we get optimal BC production as close to the seed as possible (Figure 3). We induced this polymerisation by incubating the sterilised and scarified seeds in 2 mg / ml PDA in 100 mM bicine buffer (pH = 8.5) for 16 hours.
After the seeds were coated with the black PDA, the K. sucrofermentans was added to the seeds where it sticks to this coating. We introduced the K. sucrofermentans by soaking the seeds for 15 minutes in a bacterial suspension with an optical density (OD) at 600 nm of 0.3. By doing this, we ensure that our BC is produced close to the seed surface where the bacteria are fixed6, increasing our chances for an efficient in situ coating protocol.
After we attached the bacteria to the seed surface with the PDA step, we wanted to produce the BC in situ. We coated the seeds in situ by submerging singular seeds in 96-well plates with YPD medium for 24 hours at 30^\circC.
In order to make our protocol more efficient, we wanted to verify whether in situ coating was more efficient in a static or a continuous culture. For the continuous cultures, we placed the seeds at 250 rpm in an incubator at 30^\circC (New Brunswick Innova® 44/44R), whereas we placed static cultures at 30^\circC.
We judged the success rate of our in situ coating procedure by counting seeds that had a coating that stayed when we pulled on it (Figure 4). After the seeds were succesfully coated, we washed out excess medium using sterile MilliQ to avoid contaminations later on, when seeds are used for pot experiments. After washing out the medium, we dried the seeds until use so they remain shelf-stable.
In situ protocol optimisations
After we successfully made a protocol that resulted in in situ coated seeds, we set out to increase the coating success rate so we can upscale our process for larger experiments in the future. Our goal was set at 80% coating success rate, as the used cultivar has a relatively low germination rate. In order to account for future experiments, we needed the success rate to be as high as possible so we can sow many more seeds than we need, to account for these germination issues.
Initially, we assessed the effect of volume coverage on coating success rate in 96-well plates when seeds were covered 50% by YPD or entirely submerged.
After we optimised the seed coverage, we set out to further enhance the coating efficiency by investigating the effects of continuous or static cultures, well size, scarification and/or PDA coating on the coating success rate.
Osmopriming of sorghum and germination verification of in situ coating protocol
After optimising the coating efficiency, we set out to include an osmopriming step with two goals: Ensuring homogenous germination and preventing germination during the coating procedure.
A germination assay was performed to assess the effect of different concentrations of PEG6000 creating an osmotic pressure of 0, -0.3, -0.5, -0.8 or -1 MPa during the PDA step. By increasing the water potential, we make it more difficult for seeds to take up water. In situ coating was then performed after which we put the seeds on Whatman filter paper, moistened it with demi water at 27 ^\circC, and monitored germination rate for 3 days.
Use case 1: Controlled delivery of pesticides
Our fist use case focused on tackling seed predation, largely caused by beetle (Coleoptera) larvae. In these experiments we aimed at attaching Bacillus thuringiensis (Bt) proteins,popular biological alternatives to traditional chemical pesticides, to bacterial cellulose (BC) though cellulose-binding domains (CBDs). Additionally, we assessed the effect of CBD-Cry3Aa on mealworms, also belonging to the Coleoptera order.
CBD fusion proteins
In the finalised project, the proteins will be attached to the BC after being excreted by yeast directly in our production platform. However, this was not possible during the project so far due to us all working in parallel. Consequently, as a proof of principle, we produced and isolated the CBD fusion proteins using E. coli. After isolation, the proteins were immobilised on the BC. For these experiments we used an existing CBD part (BBa-K4380000).
We produced two separate proteins: CBD-Green Fluorescent Protein (GFP), to validate the binding affinity of the CBD, and CBD-Cry3Aa, to investigate our research question. The plasmid maps are shown in (Figure 5). We first amplified the fragments using PCR (Q5® High-Fidelity DNA Polymerase, New England Biolabs), and then assembled them using Gibson Assembly (NEBuilder® HiFi DNA Assembly Master Mix, New England Biolabs). After checking the plasmids for correct size using gel electrophoresis, we transformed them into E. coli strain DH10B for amplification. Following this, we sequenced the plasmids to ensure no mutations had occurred, and transformed them into E. coli strain NiCo21(DE3) for protein production. Finally, we isolated the proteins from the cells according to the protein isolation protocol.
CBD binding validation
To determine whether proteins can be robustly attached to BC, even after downstream processing, we used GFP fused to a CBD. We measured the fluorescent signal right after incubation with the protein, after a washing step with deionised water, and after rehydration of the BC following an overnight freeze-drying procedure. We compared this signal to GFP without a CBD, to account for the BC retaining some protein without binding to it. As a negative control, we used water. Using ImageJ v 1.54p, we determined the mean gray value (MGV) of each well. From this, we subtracted the MGV of the background to obtain the corrected MGV.
Pesticide assay
CBD binding affinity to bran
To test the effectiveness of the CBD-Cry3Aa protein, we performed an assay using lesser mealworms (Alphitobius diaperinus) as a test organism. For our initial screening, we decided to use wheat bran, a common feedstock for rearing mealworms, instead of seeds, as this is easier to work with. As wheat bran consists in large part of cellulose, we assumed that the CBD fusion protein would adhere in a similar manner as to BC8. We validated this assumption following a similar protocol to the one that we used to validate the CBD binding to BC. We treated wheat bran with either water or CBD-GFP solution and incubated it for two hours. Then we washed both treatments with deionised water three times and freeze dried them overnight.
Pesticide effectiveness assay
The wheat bran was treated with CBD:Cry3Aa by submerging it in protein solution, followed by a washing step using deionised water. We compared the effectiveness of the protein to deltamethrin, a common, commercially available, synthetic insecticide. After soaking in the pesticide solutions, the wheat bran was freeze-dried overnight. To determine if the process of freeze-drying could affect the experiment, wheat bran treated with only water was also freeze-dried. As a control, untreated wheat bran was used. We exposed the mealworms to the fusion protein following the pesticide effectiveness assay protocol. For each treatment, we performed five replicates.
Use case 2: Parasitic weed suppression
Growth of compound X auxotroph by diffusion from coated seeds
After we optimised our protocol and reached a success rate of 88%, which is high enough to support experiments with large numbers of seeds, we verified that our material is a suitable matrix to deliver active substances.
In order to reduce the Striga threat, we wanted to incorporate compound X into our seed coatings. Prior research has shown that compound X is able to reduce the Striga threat by reducing its seed viability.
One crucial step in making a suitable delivery system is to verify whether we can embed compound X in the BC matrix, and whether it can diffuse gradually.
We first incorporated compound X in the seed coat by introducing it during the washing step, where concentrations of 0, 0.3, 10, 20, 50, 160 and 200 mM were used.
After we put compound X into the seed coat, we looked into whether it was present in the seed coating and whether it could diffuse through solid and liquid media. To do this, we used an auxotrophic E. coli strain reliant on compound X to grow (henceforth Strain A). If the compound is present in the coating and it can diffuse, the growth of strain A is supported, if not, strain A cannot grow. We performed diffusion assays by using liquid M9 medium with glucose, compound X and a kanamycin selection marker.
We set out to confirm the water solubility and diffusion of compound X by using 5 mL of M9 media containing concentrations of 0, 0.01, 10, 10 or 20 mM compound X to set up a reference growth curve. Then to confirm that seeds are capable of carrying and diffusing compound X, seeds carrying 0, 50, 100, 160 or 200 mM were used. Strain A was grown overnight at 37^\circC, after we diluted it to an OD600 = 0.1, in the 5 mL. OD600 was measured every 2 hours for 8 hours using a Synergy H1 plate reader. Three biological and three technical replicates were used.
In addition, we wanted to investigate the diffusion rate through solid medium. For this, we used the disk-diffusion principle, using a protocol by9 modified to our needs. M9 plates with compound concentrations of 0, 0.01, 1, 10, 20 mM were inoculated with Strain A (OD600 = 0.6). Additionally, we inoculated M9 plates without compound X which were instead inoculated with the coated seeds containing 0, 50, 100, 160, or 200 mM. We incubated the plates were overnight at 37^\circC and scored for whether the concentrations in the media allowed for growth.
Results
In situ seed coating protocol development
We first set out to create a protocol to ensure in situ coating occurred in a reliable manner, which required us to optimise the culturing conditions.
One of the first things we looked at was the effect of volume coverage on the coating success rate. When we investigated the effect of half versus completely covered seeds in the wells, we observed that entirely covering the seeds resulted in coating success rates that were almost 5x higher from 9% to 54% from half-covered to fully covered wells (Figure 6).
After verifying that a fuller well results in better coating success rates, we investigated the effect of static over continuous cultures overall. There were no direct differences in coating success rate between agitated and static cultures overall (Figure 7).
We also investigated the effect of well sizes and seed pre-treatments on the coating success rate. We pre-treated seeds by including a PDA step, meant to affix bacteria to the seed surface, and/or mechanical scarification, meant to increase surface area for BC attachment.
Upon closer inspection of the effect of continuous versus static cultures, we observed that continuous cultures produced coated seeds in 100% of well-sizes, whereas static cultures only had results in 75%.
Additionally, we observed that the combination of PDA and scarification produced the highest coating success rates, of at least 67%.
The higher success rates of PDA and scarification in continuous cultures may indicate that this treatment indeed does affix bacteria to the seeds more effectively, as opposed to only having scarification or PDA treatment, leading to detached BC pellicles.
Using what we learned from these experiments, we used fully-covered wells in a continuous culture with PDA and scarification to optimise our protocol. When comparing this between non treated seeds or seeds only having a PDA coating, we found that coating seeds with PDA and scarification resulted in a reproducible 83.3% coating success rate (Figure 8).
Osmopriming of sorghum and germination verification of in situ coating protocol
While our treatment was found to be highly effective in coating seeds, we observed that many seeds germinated during the coating process, leading to germination issues. In order to avoid this premature germination, we implemented an osmopriming step with different PEG6000 concentrations creating different osmotic water potentials.
As before, without osmopriming, seeds did not germinate efficiently within the 3 day period. However, by using an osmopriming step of -0.3 MPa, we achieved a germination efficiency of 73.3% (Figure 9). We observed that the higher the osmotic potential was after -0.3 MPa, the more reduced the germination speed was. The germination speed after osmopriming was even faster than non-treated sterilised seeds.
Use case 1: Controlled delivery of pesticides
CBD binding validation
Based on the fluorescence measurements, the fusion to the CBD allowed the proteins to bind to the BC. Even before the washing step, we observed a 1.4 fold difference in fluorescence, and this difference approached two-fold after the washing step (Figure 10 & 11). Because the imager measurements were normalised to the autofluorescence of the plate, we were also able to draw some conclusions about the decrease in signal over time. Unfortunately, after the freeze-drying and subsequent rehydration, we lost roughly half of fluorescent signal, although the difference between CBD-GFP and GFP remained significant. Based on these results we can confidently say that the CBD allows for the attachment of proteins to our BC.
Pesticide assay
CBD binding affinity to bran
Based on the fluorescence measurement, we can conclude that the CBD is able to bind to wheat bran as well, showing a three-fold increase in fluorescence compared to the control. (Figure 12 & 13). A Welch Two Sample t-test revealed that the mean values of the treatment and the control differed significantly (p-value = 0.02697).
Pesticide effectiveness assay
After one week of feeding, it appeared that the majority of the mealworms treated with the fusion protein survived, with an average mortality of 6.67% (Figure 14). When we corrected this for the mortality in the control groups, this dropped to 6.20%. This is very low compared to the mortality of 96.15% of the groups treated with deltamethrin. Although the observed effect size was significant from the control group it was not practically relevant, as a pesticide that killed such a small fraction of pests would be difficult to sell.
However, when we looked at the effect of the treatments on the weights of the mealworms, we got a different picture. We saw that after one week of feeding, the mealworms lost weight (Figure 15). We can see that the group treated with Cry3Aa differed significantly from the control group, but not from the Deltamethrin group. This is supported by the individual weight measurements, where we observed that the variance in the Control and Water treatments had increased, but not in the Deltamethrin and Cry3Aa treatments (Figure 16). All treatments were significantly different from the initial measurements; with the exception of Deltamethrin, which we can attribute to the reduced sample size. Here, Cry3Aa again differed significantly from the control, but not from the Deltamethrin treatment after one week of feeding.
Use case 2: Parasitic weed suppression
Growth of compound X auxotroph by diffusion from coated seeds
After we created the in situ coated seeds, we equipped them with increasing concentrations of compound X. Then, we investigated whether the seeds carrying compound X could support growth of an auxotroph, indicating ability to diffuse and successful compound embedment.
We observed that growth increased along with increasing compound X concentrations until 1 mM was achieved (Figure 17A). Below 1 mM, growth did not persist past 4 hours. When the compound was delivered via seed coatings, concentrations \geq50 mM were observed to support similar growth patterns (Figure 17B). This growth rate from the seed diffused compound X appears to correspond to a known concentration of 1 mM in broth. We again confirmed these results in from plate assays, where only concentrations \geq1 mM supported visible growth within 24 hours (Figure 17C&D).
Conclusion
We managed to create a successful in situ coating protocol that is able to not only encapsulate seeds, but also equip them with biological pesticides, such as Cry3Aa and deliver active compounds to the soil.
For Cry3Aa, our results showed that, although the fusion protein was not effective at killing mealworms, it resulted in starvation effects, indicative of reduced feeding behaviour. This proves that our seed coating is able to deliver active proteins to the seed, to protect it from being eaten by pests. In the end, it does not matter to the consumer if the pests die or not, as long as the seeds are not damaged. Thus, future research should investigate whether the fusion protein can prevent seed damage, or if the feeding was only stopped after the seeds had already been damaged. Furthermore, the LD50 of the fusion protein, as well as the shelf-stability should be investigated in future studies. Still, our results are a major step towards a novel, more environmentally friendly, pesticide delivery system.
As for the Striga, the compound was successfully embedded in the BC and was diffusible across liquid and solid medium. This is a big step in reducing the threat imposed by Striga, and research at Wageningen University and Research and the NIOO-KNAW will continue to further weed out this threat.
Overall, we increased BC functionalisation possibilities, by fine tuning different genes, adjusting physical growth condition and by attaching useful proteins using yeast. Our production platform played a key role in allowing in situ protein functionalisation, and will be an essential tool to increase yield. We demonstrated, with the use cases, that our BC can be used to coat seeds, and act as a platform for protein delivery and compound release. Cry3Aa could be attached to cellulose and cause starvation in mealworms, while compound X was embedded in our seed coating and showed diffusion. Overall, our wet lab results put a step forward in obtaining a modular, sustainable seed coating, that will help tomorrow crops surviving different stressors.