Introduction

To develop biodegradable seed coatings with customisable properties, we used bacterial cellulose (BC), a natural polymer, as our base material. Owing to its purity, crystallinity, mechanical strength, and biodegradability, BC is highly suitable for agricultural applications. However, to meet the diverse requirements of seed coatings, their properties must be tailored. Therefore, our work focused on functionalising BC through synthetic biology.

We investigated the functionalisation of BC with proteins. In our BC production platform, we co-cultured Saccharomyces cerevisiae with Komagataeibacter sucrofermentans to enhance BC yield. Building on this, we further leveraged the presence of S. cerevisiae and made steps to engineer it to secrete cellulose-binding domain (CBD) fusion proteins that anchor to the BC matrix and extend BC functionalities.

Learn more about these different CBD fusion proteins in our part collection

Learn more about the BC production platform

In parallel, we targeted three key properties: water holding capacity (WHC), biodegradability, and porosity. We chose WHC as it determines the ability of the seed coating to retain water, thus supporting germination, particularly in drought-prone environments. As conventional chemical methods often compromise biocompatibility, we instead employed enzymes to introduce new functional groups that could alter the hydrophilicity of BC and ultimately the WHC. Biodegradability and porosity are also important for the effectiveness of seed coatings as they can influence the loading and controlled release of beneficial compounds from the matrix. To modulate these two properties, we engineered our BC-producing organism, K. sucrofermentans, using synthetic biology, with the aim of expressing genes that alter porosity and degradation rates of BC. We developed a new conjugation protocol to facilitate genetic transformation of K. sucrofermentans, and explored inducible promoter systems to provide finer control over gene expression.

Together, we utilised a variety of strategies to functionalise BC as a versatile multifunctional seed-coating platform. Here we present our preliminary experiments and results, which already show great promise, and we look forward to expanding on them in the coming weeks.

Experiments

Protein functionalisation through secretion by S. cerevisiae

Part of the novelty in our project lies in the attachment of useful proteins to BC. Recent studies attempted, without success, to induce enzyme secretion with K. rhaeticus (a different Komagataeibacter strain). The secretion turned out to be inefficient; it had no measurable effects on the material properties of BC, and the recombinant protein expression system had a high burden on the bacteria. Overall, no successful protein secretion has been reported in BC producing bacteria¹. Therefore, we decided to use S. cerevisiae, as it is highly efficient in protein production and secretion, and genetically accessible. Additionally, it is also part of our production platform, where it plays a role in increasing yield.

An important challenge is ensuring that the proteins secreted by S. cerevisiae bind to the BC matrix. In this context, cellulose binding domains (CBD) can be fused to the proteins, allowing, through non-covalent bonds, a more efficient attachment to the BC. CBDs are short peptides (36-185 AA) at the C- or N-terminus of proteins that bind tightly to cellulose fibrils, thus increasing protein adhesion to BC². In these constructs, proteins can be modularly fused to CBDs via restriction enzyme cloning³.

Using this system, we designed a collection of 10 plasmids in S. cerevisiae for five different proteins and two CBDs (Figure 1A). Originally, we designed 22 plasmids with an additional secretion tag, as literature indicated that the optimal secretion signal must be experimentally determined and cannot be routinely predicted⁴.

We constructed the plasmids using the MoClo YTK kit⁵ for the backbone and the MoClo YSD⁴ for the secretion tag, following the Golden Gate Assembly protocol supplied with the toolkits. We ordered the additional parts for the Spytag003, the proteins of interest (POI) (Figure 1B) and CBDs as compatible synthetic DNA fragments.

We selected the POIs based on their potential to functionalise BC (CBH1 and HFBI), their suitability for detection and quantification of CBD binding (mUkG1, GFP alternative in S. cerevisiae), and their ability to modify colour (AmilCPblue and AmilGFPyellow) . To functionalise BC, we selected CBH1 for its cellulase activity, which can alter BC structure, and HFBI for its hydrophobicity.

By talking to stakeholders, we discovered that colour is an important feature in a seed coating. Therefore, we decided to incorporate blue and yellow chromoproteins. Using the IDT codon optimiser webtool, we codon optimised for S. cerevisiae all non-native protein sequences.

We selected two CBDs: dCBD (BBa_K1321340), belonging to sub family 1, and CipA (BBa_K4380000), belonging to subfamily 3. We decided to use these specific CBDs (Figure 1C) as they demonstrated consistent high binding strength after multiple washes with water, ethanol, and PBS.⁶.

The CEN6/ARS4 origin of replication used in S. cerevisiae results in a low copy plasmid count, typically ranging from 1-8 copies per cell. We chose this low-copy number plasmid over a high-copy number plasmid due to its slightly higher dynamic range and more regular expression patterns⁵.

We used the CuSO4-inducible pCUP1 promoter and a strong tTDH1 terminator to make the protein expression levels tunable. pCUP1 was previously observed to have a 55-fold induction at 50 μM for Venus fluorophore. This promoter is expected to show leaky expression under basal conditions due to the presence of CuSO4 in the yeast extract medium⁵. We chose this promoter over the less leaky alternative within the toolkit because the inducer of the alternative may have been a substrate for K. sucrofermentans, and thus not compatible for the co-culture conditions.

In addition to the collection of plasmids described above, we designed a fusion protein for the detection of non-fluorescent or chromo-proteins and the determination of binding affinity to BC (Figure 1D). This fusion protein consists of a Spycatcher003 domain linked to a sfGFP, and is able to covalently bind to the secreted protein by interaction between the Spytag and Spycatcher domains. This protein is under the expression of a strong T7 promoter in the NiCo21 Escherichia coli production strain. We built this plasmid using Gibson Assembly.

We introduced each plasmid into S. cerevisiae strain CEN.PK113-7D, following the Yeast transformation protocol. We grew the transformed yeast in Yeast Extract Peptone Dextrose (YPD) medium, prepared according to the YPD media protocol, with the addition of the G418 antibiotic. The induction was performed by adding 50 μM of CuSO4 into the YPD medium.

We assessed protein production using a microplate reader (BioTek Synergy Neo2) for the detection of chromoproteins and fluorophores. Subsequently, we examined the secretion of the protein of interest by measuring the supernatant and resuspended cell pellet with the Amersham Typhoon at the Cy2 channel and analysed for their mean grey value in ImageJ.

Tuning the water holding capacity of BC through enzymatic modifications

To alter the water holding capacity (WHC) of BC and to broaden its functional potential in seed coatings, enzymes can be used to introduce new functional groups. Consequently, hydrophilicity can be changed, affecting WHC. This is a post-harvest treatment and requires the samples to be dry. Therefore, we first produced BC samples in YPD medium, following the YPD media and BC production protocols. After harvesting, we cleaned the BC according to the downstream processing protocol, froze it overnight, and freeze-dried it (Freeze Drier Christ α 1-4 LD Plus) for 24 hours. We used the freeze-dried samples to perform the treatments.

Laccase/TEMPO-Mediated Oxidation of BC

We used the Laccase/TEMPO catalytic system to target and oxidise the hydroxyl groups in BC chains. This reaction introduces aldehyde or carboxyl groups to enhance the hydrophilicity of BC and, consequently, its WHC. To achieve this we added 15 mg of BC to 5 mL tubes. We then prepared a master mix containing the enzyme laccase from Trametes versicolor (0.4 mg/mL) and 2,2,6,6-Tetramethylpiperidin-1-oxyl (TEMPO) (2 mg/mL) dissolved in 0.1M sodium acetate buffer (pH 4.5). We added 3 mL of the master mix to the tube and incubated the mixtures at 50 °C for 3 hours. After incubation, we filtered the mixture and collected the solid residue. We washed the solid residue with ethanol, froze it overnight, freeze-dried it, and analysed it by Attenuated Total Reflectance Fourier Transform Infrared (ATR-FTIR) Spectroscopy (Bruker TENSOR II). We used wild-type, untreated BC as a control.

Lipase-Catalysed Esterification of BC

We used the enzyme lipase, from Candida rugosa, to esterify the hydroxyl groups in BC chains with palmitic acid, in an effort to introduce more hydrophobic functional groups and reduce the WHC of BC. To do so, we prepared 1.5 mL tubes, each containing 10 mg BC, 50 mg palmitic acid and 500 μL of 1-butanol. We added 10 mg lipase, leaving one tube as an enzyme-free control. We incubated the mixtures at 50 °C for 3 hours in a fume hood. After incubation, we filtered the mixtures, collected the solid residues, washed them with ethanol, froze them overnight, freeze-dried them, and analysed them by ATR-FTIR Spectroscopy.

ATR-FTIR Spectroscopy

We used the analytical technique, ATR-FTIR spectroscopy, to confirm enzymatic modifications in BC and detect any newly introduced functional groups. Before sample analysis, we obtained a background spectrum for baseline correction, using the clean ATR crystal without any sample present. Between each measurement, we cleaned the ATR crystal to prevent cross-contamination by wiping the crystal surface with lint-free, non-abrasive laboratory wipes soaked in ethanol. Subsequently, we dried the crystal for 10 seconds using the inert gas, argon, to effectively displace any residual liquid and prevent damage to the crystal surface. We collected each spectrum with an accumulation of 128 scans, a spectral resolution of 4 cm^-1, and a spectral range from 4000 to 500 cm^-1.

Water Holding Capacity (WHC) Analysis

To assess whether the enzymatic modifications altered the WHC of BC, we measured the WHC of both untreated and enzyme-treated samples following a protocol described in literature⁷. We soaked the BC samples in demineralised water for 48 hours, and WHC was calculated as the ratio of water mass lost during drying to the dry weight of the BC.

Tuning biodegradability by introducing heterologous genes

Biodegradability is one of the key properties that we focus on tailoring. By tuning the biodegradability of BC, we can adjust the decomposition rate and compound release to suit the needs of different seed coatings, which is one of the most important requirements in seed coating designs⁸. In order to modify biodegradability, we inserted heterologous genes into K. sucrofermentans, namely bslA and crdS. Both of these genes were previously demonstrated to modify the surface morphology of BC. In particular, bslA was reported to give a stronger, less brittle BC, while crdS resulted in a curdlan/cellulose bionanocomposite, which reduces the crystallinity of BC^9,10. However, these genes were only tested in Komagataeibacter xylinus, which K. sucrofermentans is a subspecies of, and were not tested for biodegradability specifically. Therefore, we aimed to introduce these heterologous genes into K. sucrofermentans and further analyse how biodegradability changes with their incorporation.

We amplified bslA from Bacillus subtilis and ordered crdS as a synthetic gene. We then used Golden Gate Assembly to insert the construct into the pSEVAb33 plasmid backbone. We chose this backbone as the pSEVA system with pBBR1 ori showed replication in Komagataeibacter in a previous study³ (Figure 2). We placed the genes under the control of the strong constitutive promoter J23104 (BBa_J23104) and a strong ribosome binding site (RBS) which was taken from an expanded synthetic toolbox of Acetobacteraceae, the bacterial family to which the genus Komagataeibacter belongs^11,12.

We first cloned these plasmids into E. coli DH5α and verified them by Sanger sequencing. Subsequently, we transformed the plasmids into K. sucrofermentans via electroporation. To validate the colonies, we performed colony PCR using both template primers and an internal primer set. The template primers amplified the heterologous genes, confirming plasmid insertion, whereas the internal primer set amplified a region of the bcsA gene (a cellulose synthase gene) to confirm that the colonies were K. sucrofermentans and not contaminants.

Because the testing of electroporation protocols brought some difficulties, we also decided to test a different approach, namely conjugation. This transformation technique is standardised for many microorganisms, but the literature on conjugation in Komagataeibacter species is highly limited. Specifically for K. sucrofermentans, no published conjugation strategies could be found. Since literature is limited, we decided to test a conjugation strategy often used for the model organism Pseudomonas putida¹³. This protocol uses biparental mating with E. coli ST18 (5-ALA auxotrophy) to achieve high efficiency conjugation. To test this protocol, we used an empty pSEVAb33 plasmid to transfer into K. sucrofermentans. We used colony PCR to validate the colonies, using the bcsA internal primer set.

We planned to assess how bslA and crdS affected the biodegradability of BC. This should be done by producing BC using the transformed K. sucrofermentans. Then, downstream processing of the BC involves treating the sheets with cellulase. After 24 hours of incubation, the slurry should be tested with High Pressure Liquid Chromatography (HPLC) analysis to quantify the presence of glucose and/or cellobiose to determine the biodegradability of the different BC with different plasmids. However, we were unfortunately unable to perform this step due to time limitations and extended trials on the transformation of K. sucrofermentans, which will be discussed further in the results section.

Porosity measurement

Porosity is one of the most important properties in a seed coating. By fine-tuning it, we can regulate the embedding of compounds in the BC matrix and adjust their release rate in the soil. In order to modify porosity, we plan to overexpress three native genes in K. sucrofermentans, namely motA/B and galU. However, before altering this property, a reliable method to assess porosity is needed. In these experiments, we aimed to create a baseline for porosity measurement by using two different approaches: the mass-density method and the SEM imaging, both in wild-type (WT) cellulose.

Porosity measurement with the mass-density method

Before measurement, comparable cellulose samples need to be produced. Consequently, we standardised culture conditions. We inoculated a 10% v/v of OD600 = 0.3 K. sucrofermentans overnight culture in YPD medium and grew it in 1 L Erlenmeyer flasks at 25°C. The medium was prepared according to the YPD media protocol.

After three days, the cellulose sheets were collected and cleaned according to the downstream processing protocol. After cleaning the BC, we followed the protocol for the measurement of porosity using the mass-density method¹⁴. To perform this measurement, we first removed the surface water from the BC samples using lint-free, non-abrasive laboratory wipes. Then, we froze the samples overnight at -20°C, recorded the frozen weight, freeze-dried them (Freeze Drier Christ α 1-4 LD Plus) for 24 hours, and recorded their dried weight. Finally, we assessed the porosity. For the calculation, we assumed the BC nanofiber density to be 1.5 g/cm³, according to previous literature ^{14, 15}.

Pore area measurement with ImageJ

We decided to further analyse two of the freeze-dried BC samples using SEM microscopy. These samples were prepared and imaged by the Wageningen Electron Microscopy Center. Before the analysis, our BC was cut into smaller pieces. Samples of approximately 1 cm² were attached to aluminium stubs with double-sided carbon tape and were sputter coated with approx. 10nm tungsten from above and tilted at +45 degrees and -45 degrees (SCD 500, Leica, Vienna, Austria). The imaging was performed using a Field emission scanning electron microscope (Magellan 400, Thermo-Fischer/FEI, Eindhoven, the Netherlands). SEM specs: 2.00kV, 13pA.

By using ImageJ, we determined the size of the pores and their distributions in our SEM images. To perform these measurements, we first set a scale, by measuring the length of 5 μm in pixels.

Then, we cropped the images to the same size and saved them as bnp files to allow measurement. Binarisation of the images was performed by adjusting the threshold to cover all the pores in red, while leaving in the BC fibrils in gray. The Analyse Particles tool allowed us to determine the areas of the different pores in the pictures.

Inducible promoters in K. sucrofermentans for functionalisation

We investigated the use of inducible promoters as they enable external control of gene expression. This is desirable in our project, as in the future we want to tune the expression of genes involved in porosity and biodegradability based on the need. A genetic toolkit of Komagataeibacter was previously established, which reported two inducible promoters that are AHL-induced (pLux) and ATc-induced (pTet)³. While both inducible promoters showed to drive strong expression with relatively low leakiness, these inducers often are undesirable for large-scale industrial production as they can account for non-negligible costs¹⁶. With the expansion of Komagataeibacter toolbox in mind, we wanted to identify and evaluate alternative inducible systems that are both robust and economically viable for K. sucrofermentans.

We developed and tested the inducible promoter hpdR/phpdH, which was previously tested for tunable control of gene expression in P. putida KT2440¹⁷. We aimed to build and test this inducible promoter in K. sucrofermentans and to evaluate the expression of gfp with different inducer concentrations.

We first constructed the plasmid in silico and amplified the inducible promoters from wild-type P. putida KT2440. Then, we assembled the construct into the pSEVAb33 plasmid backbone using Golden Gate Assembly (Figure 3). To ensure robust expression as well as consistency, we incorporated this system under the same design as the heterologous genes (bslA & crdS).

We introduced the plasmid first into E. coli DH5α, and then into K. sucrofermentans via electroporation. We validated the positive colonies through colony PCR targeting the insert. Additionally, we used a set of internal primers to amplify a part of the bcsA gene, a region of the bacterial cellulose synthase operon that is specific to the Komagataeibacter genus. This dual check confirmed that the colonies were indeed K. sucrofermentans and not contaminants. Following this, we tested the functionality of the inducible promoter by measuring fluorescence under different concentrations of the inducer, Levulinic Acid (LA). We grew verified K. sucrofermentans colonies carrying the plasmid in M9 minimal medium supplemented with 25 g/L glucose and 10mM . We tested four concentrations of LA (0, 0.2, 2, and 20 mM), and measured gfp fluorescence and OD600 at 0, 2, 4, and 6h using a BioTek Synergy H1 Microplate Reader (excitation 485nm, emission 535 nm). Furthermore, we calculated relative fluorescence by normalising gfp intensity to OD600.

Results

Protein functionalisation through secretion by S. cerevisiae

As discussed previously, we designed a collection of 22 plasmids, with five different proteins, two CBDs and an additional secretion tag. Of these, a total of 10 plasmids were successfully transformed into yeast. However, some of the plasmids were missing one or multiple parts. The total composition of each of the successfully transformed plasmids is described in Table 1, including their function and eventually missing parts. We also successfully built and expressed the Spycatcher003 detection protein, but was not used in this work.

We measured the six plasmids expressing either fluorescent or chromo-proteins for growth, induction capability and protein production. All four non-quantified plasmids, four in total, confirmed growth in S. cerevisiae, but were not measured for OD600, absorbance or fluorescence.

All S. cerevisiae strains showed fast, exponential growth within the first 12-20 hours, after which the diauxic transition occurred (Figure 4A). The diauxic phase facilitated the increase in biomass up to 160 hours, quantified by OD600. Notably, AmilCPblue-dCBD had an early transition and lower final OD600 compared to the other conditions. No significant changes in final OD600 were present between non-induced and induced conditions (Figure 4B). This indicated little to no burden on the cell by induction for recombinant protein expression of UkG1, AmilCPblue and AmilGFPyellow.

The fluorescence-based detection of protein expression and secretion is demonstrated in (Figure 5). Figure 5A presents the fluorescence intensity of mUkG1 (FL475-510) over a 160-hour time period, under both induced and non-induced conditions. UkG1 revealed distinct fluorescence patterns, with constructs containing either mUkG1 or AmilGFPyellow showing fluorescence increased and thus protein production, beginning at around 48 hours, for both non-induced and induced samples.

Compared to the control, containing no protein coding sequence, protein was produced at a fast rate between 48 and 80 hours. After this period, the intensity of fluorescent proteins was stationary when corrected for OD600. Uncorrected data showed a steady increase in relative fluorescent intensity. Protein production per cell continued again when the cells were near the stationary phase, at around 120-130 hours. Induction with CuSO4 increased protein production 2.6-4.2 fold compared to the non-induced condition (Figure 5B). This shows the expected leakiness of the promoter, but also an inducible increase in protein expression.

We also performed measurements on the blue and yellow at 588 nm (Figure 6) and 512 nm (Figure 7), respectively. However, corrected measurements per cell showed no production of either chromoprotein in S. cerevisiae above background. Absorbance on these wavelengths correlated with the increase of OD600, so the production of minimal amounts of chromoprotein may be possible, but not detectable with the plate reader or by eye.

To assess potential secretion of the produced proteins, we measured fluorescence intensity in both the resuspended cell and supernatant fractions of yeast cultures expressing AGA2-UkG1-cipA, AGA2-UkG1-dCBD, and AGA2-AmilGFPyellow-dCBD constructs under induced and non-induced conditions, normalised to a plasmid control set at 100% (Figure 8). Notably, we detected elevated fluorescence levels in culture supernatants for all constructs except for AmilGFPyellow-dCBD. Particularly high fluorescence was recorded for both the UkG1 fusion proteins, with the most pronounced increase observed in induced cultures where cellular fluorescence reached approximately 113.20% for AGA2-UkG1-cipA, 121.52% for AGA2-UkG1-dCBD, and 133.25% for AGA2-AmilGFPyellow-dCBD compared to the control. The presence of detectable fluorescence in supernatants from both induced and non-induced cultures suggests that fluorescent proteins were present in the culture medium, though the exact mechanism remains unclear. The lack of fluorescence in the AmilGFPyellow-dCBD supernatant and the highest intensity in the cells indicates that a lack of secretion and accumulation of intracellular protein might have occurred.

These preliminary data suggest that fluorescent UkG1 proteins accumulate in the culture medium, potentially through a combination of active secretion and potential passive release via cell lysis. However, further investigation is required to distinguish between these mechanisms.

Tuning the water holding capacity of BC through enzymatic modifications

Before testing changes in the water-holding capacity (WHC) of BC after enzymatic treatment, we first employed Attenuated Total Reflectance Fourier Transform Infrared (ATR-FTIR) Spectroscopy. We used this analytical technique to determine whether the enzymes introduced the desired modifications in the surface chemistry of BC, which would in turn affect the WHC.

The structure of BC, composed of repeating D-glucose units and a dense network of hydrogen bonds, produces characteristic vibrational fingerprints detectable by ATR-FTIR Table 2 highlights the most important characteristic vibrations of bonds present in BC and their corresponding absorption regions, used to confirm its identity. In particular, the carbonyl (C=O) stretching vibration near ∼ 1700 cm⁻¹ is absent in natural, unmodified BC. The emergence of this peak indicates the presence of a carbonyl-containing group, making it a crucial region to focus on in the spectrum when detecting laccase/TEMPO-mediated oxidation or lipase-catalysed esterification of BC.

Laccase/TEMPO-Mediated Oxidation of BC:

When analysing ATR-FTIR spectra, we focused on the shape and position of the peaks to confirm the presence of functional groups, as the data are qualitative. The ATR-FTIR spectrum of unmodified BC (Figure 9A) displayed the expected characteristic absorptions, which were annotated with colour-coded arrows corresponding to Table 2. These features confirmed the native polymer fingerprint and the integrity of the BC backbone prior to enzymatic treatment.

Following the 3 hours incubation of BC with the Laccase/TEMPO reaction mixture, the region corresponding to the BC backbone (1000–3000 cm⁻¹) (FIgure 9B) remained largely unchanged, indicating that the polymer remained intact and undegraded during treatment. Furthermore, we observed a distinct peak near ∼ 1700 cm⁻¹ (dark brown arrow), consistent with (C=O) stretching vibrations. This demonstrated that aldehyde or carboxyl groups were introduced into the BC sample, which we expected from TEMPO-mediated oxidation of the hydroxyl groups on BC (FIgure 10).

Effect of Laccase/TEMPO-Mediated Oxidation on WHC of BC

After analysing the BC samples with ATR-FTIR, we went on to calculate its WHC. The BC sample incubated with the Laccase/TEMPO reaction mixture showed a higher WHC than the untreated control (Figure 11). This observation is consistent with previous reports in the literature, which indicate that oxidised BC shows improved WHC. The increased surface hydrophilicity arises from the introduction of more polar carboxyl groups, which can strongly associate with water molecules and trap them²¹.

From a seed-coating perspective, the enhanced WHC of BC ensures that seeds remain hydrated, which is critical for germination and reducing stress under dry conditions. However, we performed no statistical analysis, as WHC values were obtained without replicates. Therefore, the observed increases should be considered preliminary and descriptive rather than conclusive. Nevertheless, the data are promising. In future experiments, we will include replicates to report the mean ± standard deviation and perform appropriate statistical tests to rigorously assess the significance of the results.

Overall, 3 hours of incubation with the Laccase/TEMPO mixture appeared sufficient for oxidation of BC to occur. In future experiments, we plan to reduce incubation time and test it in smaller increments to optimise the reaction. The parameters used in our reaction were based on literature²². We will further refine the conditions by varying the TEMPO concentration and quantifying the effect of reagent stoichiometry on the carboxyl group content of BC. Once the optimal conditions are established, we will upscale the experiment to enable modification of larger quantities of BC. The modified BC will then be processed into a suitable form, e.g., a powder, for coating seeds, where we will evaluate its performance and determine whether enzymatically oxidised BC also improves seed coatings and WHC under dry conditions compared with unmodified BC. Overall, these findings suggest that the Laccase/TEMPO mediated oxidation of BC has strong potential for developing functional seed coatings with increased WHC, to enhance seed resilience under challenging environmental conditions.

Lipase-Catalysed Esterification of BC

In an effort to reduce the WHC of BC, we used the enzyme lipase with palmitic acid as a substrate to catalyse an esterification reaction with the hydroxyl groups on BC (Figure 14). We checked whether esterification had occurred by analysing the ATR-FTIR spectra of the BC samples.

When we compared the ATR-FTIR spectrum of the unmodified BC (Figure 12A) with the second control containing BC incubated with palmitic acid without lipase (Figure 12C), the difference between the spectra is unexpectedly large. The main peaks corresponding to the BC backbone appeared distorted in the second control. Furthermore, the control spectrum (Figure 12C) showed some resemblance with that of pure palmitic acid (Figure 12B). Additional noise and baseline drift, likely due to optical and technical factors could have further contributed to distortion, particularly evident in the second control (Figure 12C).

Ideally, the two controls (Figure 12A&C) and should have looked more similar. The similarity between the control with BC and palmitic acid without lipase (Figure 12C) and pure palmitic acid (Figure 12B) indicated that residual palmitic acid remained associated with the BC even after washing. This observation suggested two possibilities: either palmitic acid was physically absorbed into the BC matrix, making it difficult to remove, or the washing procedure was simply inefficient.

This is important as palmitic acid itself has a C=O bond (Figure 13) that produces a stretching band around ∼1700 cm⁻¹. Since either possibility can produce a peak in this region and be mistakenly interpreted as evidence of ester formation according to the mechanism of lipase-catalysed esterification (Figure 14), the interpretation of covalent ester formation becomes complicated.

Since the washing appeared insufficient, it is likely that the BC sample incubated with palmitic acid and lipase (Figure 12D) also retained excess palmitic acid. In future experiments, we will optimise the washing procedure to completely remove unreacted palmitic acid and improve data reliability. As a starting point, we will wash the solid obtained after filtration with warmed rather than room temperature ethanol under shaking conditions for a longer period of time to see if this is more effective.

To further refine our method, we will test multiple parameters, including testing alternative organic solvents other than 1-butanol, different incubation temperatures (30–50 °C) and duration. In this experiment, we chose the BC:palmitic acid (1:5) ratio, based on previous findings in literature²³. However, in the future, we will also experiment with varied BC:fatty acid ratios. Once the optimal parameters are identified, we will up scale the experiment to modify larger quantities of BC. The modified BC will then be processed into a usable form, such as a powder, for seed coating applications.

In theory, if lipase-catalysed esterification occurs at the hydroxyl groups of BC the surface chemistry will change from more hydrophilic to hydrophobic due to the presence of longer alkyl chains with reduced polarity. With a decrease in the number of hydrogen bonding sites, a reduction in hydrophilicity should be observed. The water affinity and swelling of BC will reduce, in turn decreasing the WHC. Therefore, repeating the lipase-catalysed esterification of BC with other soil compatible long-chain fatty acids like stearic acid would help determine whether increasing the carbon chain length produces a systematic effect on the WHC of BC²⁴.

In the future, once qualitative data proves to be more conclusive, we will measure the WHC of all BC samples treated with lipase in triplicate to also allow for sufficient statistical analysis.

Tuning biodegradability by introducing heterologous genes

As previously stated, engineering the biodegradability of BC could allow controlling the release of compounds from seed coatings. We attempted to introduce the genes, bslA and crdS, into K. sucrofermentans.^9,10.

We tested three different Komagataeibacter electroporation protocols. From these, one proved to be successful. The first protocol was replicated from a previous research on a synthetic biology toolkit for Acetobacteraceae¹¹. The second protocol was replicated from a previous iGEM team, which worked with K. rhaeticus, and the third protocol replicated was from a previous paper on directed evolution working with K. sucrofermentans²⁵. This stage took quite some time for us to test and optimise the protocols, where each transformation trial was done with overnight recovery followed by 48-72h of incubation to see colony formation.

As previously stated, because the testing of electroporation protocols brought some difficulties, we also decided to test conjugation. The conjugation was verified by a colony PCR of 12 colonies (Figure 15). The first 12 bands show the correct amplification length (306bp), and the subsequent 12 bands highlight the verification of the correct transformant species. The bcsA gene is specific to Komagataeibacter and precludes contamination possibilities. In the following 2 bands, the primers specific to the empthy plasmid were tested on a WT K. sucrofermentans to verify correct introduction of the plasmid. And the last two bands are used as a control, where K. sucrofermentans WT is amplified with the primers for bcsA. Based on 12 individual colonies, we achieved a 100% transformation efficiency. However, more extensive testing is required to determine the true efficiency of this method.

After several trials of transformation, we successfully transformed K. sucrofermentans with an empty pSEVAb33 plasmid with electroporation as well, which was confirmed by colony PCR and proved to have a high efficiency, as all of the tested colonies carried the plasmid (Figure 16 & 17).

Once the electroporation protocol was proven to be successful, we continued with the transformation of K. sucrofermentans with the plasmids to deliver the heterologous genes, both bslA and crdS. However, only the bslA plasmid was successfully transformed into K. sucrofermentans, but not the crdS plasmid (Figure 18). We confirmed the bslA plasmid by colony PCR to be successfully inserted into K. sucrofermentans with the expected band (8̃00bp) seen for all the colonies tested, as well as confirmation that these are K. sucrofermentans and not contaminants (Figure 19). Due to time constraints, we were unable to redo the crdS transformation as well as perform the biodegradability assay. We tried to grow the BC with the transformed K. sucrofermentans, however, the incubation time was not sufficient, as the BC was still too thin to be downstream processed.

Porosity measurement

A reliable assessment of porosity is essential to create a baseline for future measurements. The mass-density method and the SEM imaging analysis proved to be important tools that helped our team and could also help other teams in future research.

Porosity measurement with the mass-density method

We used the mass-density method to determine the percentage of porosity in the freeze-dried samples. According to literature, the expected porosity of BC is approximately 92%²⁶. However, in the samples tested within this project, the average porosity was estimated to 98.3% (Figure 20).

A possible reason for this higher porosity measurement is that we used a different BC producing strain or different production conditions. Overall, having an increased porosity allows the embedding of a larger amount of compound and also affects the size of compounds that can be embedded. For this reason, SEM microscopy was used to determine the area of the pores.

Pore area measurement with ImageJ.

For each sample, we analysed 4 images using ImageJ software to determine the area of the pores in μm². We chose two samples of different measured porosity with the mass-density method: 96.9% (Sample 5) and 98.4% (Sample 7). For the first sample, we determined an average pore size of 0.048μm² (Figure 21), while for the second sample the average was 0.063μm² (Figure 22). The higher pore size measurement in the second sample can be related to its higher porosity compared to the first sample. Multiple outliers of higher pore size are present for each image. However, the measurement adopted proved to be significant (p<0.05).

Overall, we selected SEM imaging to determine pore size and to relate it to porosity. This measurement can be useful to determine the size of compounds that can be embedded, therefore our future experiments will focus on trying to add compounds of different sizes and HPLC analysis to quantify compound embedding based on the porosity and average pore area.

SEM imaging, together with the mass-density porosity measurement, will help us also as a baseline to evaluate the impact of overexpressing the genes motA/motB and galU on the overall porosity.

Inducible promoters in K. sucrofermentans for functionalisation

Recent studies have reported that a promoter inducible with 3-hydroxypropionic acid (3-HP) exhibits strong activity, comparable to well-characterised promoters such as PmxaF and PL / O4²⁷. Originally native to P. putida KT2440, HpdR/PhpdH has also been validated in multiple strains of Methylorubrum extorquens AM1, which is an α-proteobacteria which could utilise C1 compounds to produce diverse biomaterials ranging from bioplastics to biopharmaceuticals²⁸. These findings highlight the potential of the 3-HP inducible system as a promising tool for K. sucrofermentans, enabling tunable functionalisation of BC resulting in a modular polymer. This inducible promoter has also been reported to be efficiently induced by levulinic acid (LA)¹⁷. LA itself is a low-cost, renewable, and sustainable platform chemical, which was also one of the reasons that the previous research investigated the use of LA as an inducer. This further enhances the suitability of this promoter system for large-scale applications.

The hpdR/PhpdH inducible system functions through an activator-based mechanism²⁹. HpdR acts as a transcriptional activator that binds to levulinic acid (LA) and subsequently activates the PhpdH promoter to express phpdH (Figure 23). However, in our construct, the native phpdH gene was replaced with gfp, enabling the quantification of promoter activity from fluorescence measurements (Figure 23A). As a negative control, we constructed a variant lacking the hpdR region. In this control, no fluorescence is expected, since the activator is absent even in the presence of the inducer (Figure 23B).

We amplified, assembled, and cloned the plasmid in E. coli DH5α and confirmed the presence of the correct plasmid through colony PCR and Sanger sequencing. With the success of the electroporation protocol, we transformed K. sucrofermentans with the hpdR/phpdH plasmid as well, and we confirmed the hpdR/phpdH plasmid by colony PCR to be successfully inserted with the expected band (1̃800bp) seen for all the colonies tested (Figure 24).

We tested the hpdR/phpdH inducible system using the BioTek Synergy H1 Microplate Reader and induced expression with LA. From the result, we can see that the system does not exhibit a strong increase in fluorescence with increasing concentrations of LA (Figure 25). To account for background signals, we also tested K. sucrofermentans with the empty vector (pSEVAb33) as a blank. These cells displayed autofluorescence during the plate reader assay, suggesting that the fluorescence observed in the negative control may have originated from K. sucrofermentans itself rather than from the phpdH system.

A study reported that G. xylinus was unable to utilise LA within the growth medium, which suggests that it may lack an LA transporter³⁰. This limitation might also apply to K. sucrofermentans , as they belong to the same genus. However, due to the limited availability of genomic information, such possibilities remain speculative. Interestingly, we can see that the inducible system exhibited higher fluorescence, even at 0mM concentration of LA, compared to the negative control, which might suggest that this promoter might be activated by a metabolite naturally present in K. sucrofermentans that shares structural similarity with LA or 3-HP. This might open new doors for the unexplored cross-talk between K. sucrofermentans endogenous metabolites and heterologous regulatory systems, which, if they were verified, could enable self-regulated induction systems.

Conclusions and Outlook

As a team, we enhanced the current functionalisation possibilities of BC, allowing us to change key properties in our seed coatings.

Within protein functionalisation, we demonstrated that the mUkG1/AmilGFPyellow protein exhibits sufficient signal intensity for reliable detection using the plate reader assay. While preliminary evidence suggests that protein secretion is occurring, assay and/or strain optimisation is required to achieve adequate sensitivity and reproducibility for protein quantification and purification. These findings indicate that protein functionalisation can be controlled through this approach and that inducible expression systems enable control over protein production. Further experiments should focus on the incorporation of proteins into the BC matrix during biosynthesis in co-culture conditions. Nevertheless, this shows a proof of concept of the modular protein system in S. cerevisiae, with the option to theoretically replace mUkG1 with any desired POI.

The preliminary WHC measurements showed that oxidised BC had higher WHC, indicating that laccase/TEMPO-mediated oxidation effectively increased BC’s capacity to retain water. ATR-FTIR analysis provided qualitative evidence that laccase/TEMPO treatment introduced a new carbonyl functional group, consistent with oxidation of hydroxyl groups in BC. By contrast, lipase-catalysed esterification gave inconclusive results. Baseline drift and spectral noise highlighted the need for improved sample preparation and data processing. Overall, future experiments should incorporate replicates, refine washing protocols, and vary reaction conditions to identify optimal parameters for enzymatic modification of BC surface chemistry to alter the WHC. At present, laccase/TEMPO-mediated oxidation shows promise for increasing the WHC of BC and advancing seed coating formulations. With this improvement, seed coats made from modified BC could support germination under stressful conditions, particularly in drought-prone areas, ultimately enhancing crop resilience.

We have established a foundation for the genetic modifications of K. sucrofermentans, creating opportunities to introduce additional heterologous genes in the future. Additionally, we provided future researchers and iGEM teams with a protocol for conjugation in K. sucrofermentans. Although we successfully introduced a heterologous gene for biodegradability, we have not yet evaluated biodegradability through dedicated assays. Our team also provided a baseline for porosity measurement. This will be used to assess changes in porosity after the overexpression of target genes, and provides future iGEM teams with an easy porosity assessment method.

Finally, we successfully introduced a previously untested inducible promoter candidate in K. sucrofermentans, expanding the potential for tunable expression of heterologous genes, especially in large-scale production with a cost-effective inducer. Further studies are still needed to evaluate the compatibility and efficiency of this promoter system in K. sucrofermentans. To further enhance modularity, different genes should be placed under inducible promoter systems, enabling tunable control of expression and the production of fully customisable BC, tailored for specific needs.

Overall, our results demonstrate that BC can be tailored for specific needs using different approaches, enhancing useful properties for a seed coating.

(1)

Goosens, V. J.; Walker, K. T.; Aragon, S. M.; Singh, A.; Senthivel, V. R.; Dekker, L.; Caro-Astorga, J.; Buat, M. L. A.; Song, W.; Lee, K.-Y.; Ellis, T. Komagataeibacter Tool Kit (KTK): A Modular Cloning System for Multigene Constructs and Programmed Protein Secretion from Cellulose Producing Bacteria. ACS Synthetic Biology 2021, 10 (12), 3422–3434. https://doi.org/10.1021/acssynbio.1c00358.

(2)

Yang, F.; Jin, E.-S.; Zhu, Y.; Wu, S.; Zhu, W.; Jin, Y.; Song, J. A Mini-Review on the Applications of Cellulose-Binding Domains in Lignocellulosic Material Utilizations. BioResources 2015, 10 (3). https://doi.org/10.15376/biores.10.3.Yang.

(3)

Florea, M.; Hagemann, H.; Santosa, G.; Abbott, J.; Micklem, C. N.; Spencer-Milnes, X.; Arroyo Garcia, L. de; Paschou, D.; Lazenbatt, C.; Kong, D.; others. Engineering Control of Bacterial Cellulose Production Using a Genetic Toolkit and a New Cellulose-Producing Strain. Proceedings of the National Academy of Sciences 2016, 113 (24), E3431–E3440.

(4)

O’Riordan, N. M.; Jurić, V.; O’Neill, S. K.; Roche, A. P.; Young, P. W. A Yeast Modular Cloning (MoClo) Toolkit Expansion for Optimization of Heterologous Protein Secretion and Surface Display in Saccharomyces Cerevisiae. ACS Synthetic Biology 2024, 13 (4), 1246–1258. https://doi.org/10.1021/acssynbio.3c00743.

(5)

Lee, M. E.; DeLoache, W. C.; Cervantes, B.; Dueber, J. E. A Highly Characterized Yeast Toolkit for Modular, Multipart Assembly. ACS Synthetic Biology 2015, 4 (9), 975–986. https://doi.org/10.1021/sb500366v.

(6)

Gilbert, C.; Tang, T.-C.; Ott, W.; Dorr, B. A.; Shaw, W. M.; Sun, G. L.; Lu, T. K.; Ellis, T. Living Materials with Programmable Functionalities Grown from Engineered Microbial Co-Cultures. Nature Materials 2021, 20 (5), 691–700. https://doi.org/10.1038/s41563-020-00857-5.

(7)

Schrecker, S. T.; Gostomski, P. A. Determining the Water Holding Capacity of Microbial Cellulose. Biotechnology Letters 2005, 27 (19), 1435–1438. https://doi.org/10.1007/s10529-005-1465-y.

(8)

Skrzypczak, D.; Jarzembowski, Ł.; Izydorczyk, G.; Mikula, K.; Hoppe, V.; Mielko, K. A.; Pudełko-Malik, N.; Młynarz, P.; Chojnacka, K.; Witek-Krowiak, A. Hydrogel Alginate Seed Coating as an Innovative Method for Delivering Nutrients at the Early Stages of Plant Growth. Polymers 2021, 13 (23), 4233.

(9)

Fang, J.; Kawano, S.; Tajima, K.; Kondo, T. In Vivo Curdlan/Cellulose Bionanocomposite Synthesis by Genetically Modified Gluconacetobacter Sucrofermentans. Biomacromolecules 2015, 16, 3154–3160. https://doi.org/10.1021/acs.biomac.5b01075.

(10)

Gilmour, K. A.; Aljannat, M.; Markwell, C.; James, P.; Scott, J.; Jiang, Y.; Torun, H.; Dade-Robertson, M.; Zhang, M. Biofilm Inspired Fabrication of Functional Bacterial Cellulose Through Exsitu and in-Situ Approaches. Carbohydrate Polymers 2023, 304, 120482. https://doi.org/10.1016/j.carbpol.2022.120482.

(11)

Teh, M. Y.; Ooi, K. H.; Danny Teo, S. X.; Bin Mansoor, M. E.; Shaun Lim, W. Z.; Tan, M. H. An Expanded Synthetic Biology Toolkit for Gene Expression Control in Acetobacteraceae. ACS synthetic biology 2019, 8 (4), 708–723.

(12)

Karničnik, B.; Accetto, T.; Fanedl, L.; Jugović, I.; Trček, J. Isolation and Characterization of Komagataeibacter Piraceti Sp. Nov. And Novacetimonas Labruscae Sp. Nov.: Two Novel Microaerobic Cellulose-Producing Acetic Acid Bacteria from Vinegars. Microorganisms 2025, 13 (2), 456.

(13)

Thoma, S.; Schobert, M. An improvedEscherichia Colidonor Strain for Diparental Mating. FEMS Microbiol. Lett. 2009, 294 (2), 127–132.

(14)

Hossen, M. R.; Talbot, M. W.; Kennard, R.; Bousfield, D. W.; Mason, M. D. A Comparative Study of Methods for Porosity Determination of Cellulose Based Porous Materials. Cellulose 2020, 27 (12), 6849–6860. https://doi.org/10.1007/s10570-020-03257-9.

(15)

Malm, C.; Genung, L.; Fleckenstein, J. Densities of Cellulose Esters. Industrial & Engineering Chemistry 1947, 39 (12), 1499–1504. https://doi.org/10.1021/ie50455a022.

(16)

Politi, N.; Pasotti, L.; Zucca, S.; Casanova, M.; Micoli, G.; Cusella De Angelis, M. G.; Magni, P. Half-Life Measurements of Chemical Inducers for Recombinant Gene Expression. Journal of biological engineering 2014, 8 (1), 5.

(17)

Sathesh-Prabu, C.; Tiwari, R.; Kim, D.; Lee, S. K. Inducible and Tunable Gene Expression Systems for Pseudomonas Putida KT2440. Scientific reports 2021, 11 (1), 18079.

(18)

Hosakun, Y.; Halász, K.; Horváth, M.; Csóka, L.; Djoković, V. ATR-FTIR Study of the Interaction of CO2 with Bacterial Cellulose-Based Membranes. Chemical Engineering Journal 2017, 324, 83–92. https://doi.org/10.1016/j.cej.2017.05.029.

(19)

Yuan, Y.; Teja, A. S. Quantification of Specific Interactions Between CO2 and the Carbonyl Group in Polymers via ATR-FTIR Measurements. The Journal of Supercritical Fluids 2010, 56 (2), 208–212. https://doi.org/10.1016/j.supflu.2010.12.010.

(20)

Kaczmarek, M.; Białkowska, A. M. Enzymatic Functionalization of Bacterial Nanocellulose: Current Approaches and Future Prospects. Journal of Nanobiotechnology 2025, 23 (1). https://doi.org/10.1186/s12951-025-03163-x.

(21)

Piao, X.; Li, J.; Zhao, Y.; Guo, L.; Zheng, B.; Zhou, R.; Ostrikov, K. Oxidized Cellulose Nanofibrils-Based Surimi Gel Enhancing Additives: Interactions, Performance and Mechanisms. Food Hydrocolloids 2022, 133, 107893. https://doi.org/10.1016/j.foodhyd.2022.107893.

(22)

Zhou, Q.; Zhang, Q.; Wang, P.; Deng, C.; Wang, Q.; Fan, X. Enhancement Biocompatibility of Bacterial Cellulose Membrane via Laccase/TEMPO Mediated Grafting of Silk Fibroins. Fibers and Polymers 2017, 18 (8), 1478–1485. https://doi.org/10.1007/s12221-017-7306-5.

(23)

Amaliyah, S. Synthesis and Characterization of Cellulose Stearate Ester from Bacterial Cellulose and Stearic Acid Enzymatically Using Immobilized Lipase. Molekul 2016, 11 (2), 256. https://doi.org/10.20884/1.jm.2016.11.2.225.

(24)

Luo, T.; Koirala, S.; He, S.; Jellicoe, M.; Truong, V. K.; Alhaji, J. H.; Ibrahim, M. Efficacy of Stearic Acid, Potato Starch, and Polycaprolactone in the Development of pH and Temperature-Responsive Fertilizer for Eco-Resilient Agriculture System. International Journal of Biological Macromolecules 2025, 142566. https://doi.org/10.1016/j.ijbiomac.2025.142566.

(25)

Laurent, J. M.; Jain, A.; Kan, A.; Steinacher, M.; Enrriquez Casimiro, N.; Stavrakis, S.; deMello, A. J.; Studart, A. R. Directed Evolution of Material-Producing Microorganisms. Proceedings of the National Academy of Sciences 2024, 121 (31), e2403585121.

(26)

Tang, W.; Jia, S.; Jia, Y.; others. The Influence of Fermentation Conditions and Post-Treatment Methods on Porosity of Bacterial Cellulose Membrane. World Journal of Microbiology and Biotechnology 2010, 26, 125–131. https://doi.org/10.1007/s11274-009-0151-y.

(27)

Sathesh-Prabu, C.; Ryu, Y. S.; Lee, S. K. Levulinic Acid-Inducible and Tunable Gene Expression System for Methylorubrum Extorquens. Frontiers in Bioengineering and Biotechnology 2021, 9, 797020.

(28)

Phan, U. T.; Jeon, B. W.; Kim, Y. H. Microbial Engineering of Methylorubrum Extorquens AM1 to Enhance CO2 Conversion into Formate. Enzyme and Microbial Technology 2023, 168, 110264.

(29)

Sathesh-Prabu, C.; Tiwari, R.; Lee, S. K. Substrate-Inducible and Antibiotic-Free High-Level 4-Hydroxyvaleric Acid Production in Engineered Escherichia Coli. Frontiers in Bioengineering and Biotechnology 2022, 10, 960907.

(30)

Zhang, S.; Winestrand, S.; Chen, L.; Li, D.; Jonsson, L. J.; Hong, F. Tolerance of the Nanocellulose-Producing Bacterium Gluconacetobacter Xylinus to Lignocellulose-Derived Acids and Aldehydes. Journal of agricultural and food chemistry 2014, 62 (40), 9792–9799.