On this page, we outline the engineering success of our project. First, we shortly introduce the field of Synthetic Biology (SynBio) and explain why engineering principles are essential in the context of SynBio. After describing the main phases of the engineering cycle, we walk you through the design, build, test and learn stages that our team applied.
Synthetic Biology represents a growing discipline began in the late 20th century and has rapidly progressed since the early 2000s. The first international conference, Synthetic Biology 1.0 (SB 1.0), took place in 2004 at the Massachusetts Institute of Technology where researchers discussed the engineering of biology (MIT News 2004).
From there, the field developed rapidly, starting with signaling circuits and genome synthesis, over metabolic engineering and minimal cells, to xenobiology (ZKBS 2018).
SynBio focuses on the design and construction of new biological parts, devices and systems or the re-design of existing, natural systems for useful applications, making SynBio an engineering science as well as an applied science. It represents an interdisciplinary field integrating concepts from genetic engineering, biochemistry, systems biology, microbiology and computer science.
However, SynBio is not only genetic engineering focusing on targeted DNA manipulation, but also includes an engineering perspective applied to biological systems. Key engineering principles are abstraction, modularity, standardization and characterization.
Abstraction allows biological parts to be decoupled from their natural function for use in another context, leading to completely new systems. Modularity enables the separation and recombination of system components to assemble parts to devices and complex systems. Additionally, standardization of components guarantees compatibility of parts and application of standardized tools and procedures (e.g. BioBrick standard assembly). Finally, parts are characterized, and their features are documented in the Parts Registry, serving as an information system and providing comprehensive information.
The typical engineering cycle for synthetic biologists consists of four stages: design, build, test and learn (Fig. 2).
Scientists first need to define their goal and specify their project idea. They carefully think about a certain challenge and explore possible solutions to the problem. They develop a strategy and design biological parts, devices and systems to realize their idea. This requires biological background knowledge as well as comprehensive literature research. During design, engineers make use of computational tools and ensure compatibility of parts by applying common standards like the BioBrick RFC[10] or Golden Gate Type IIS standard .
In the next phase, the designed parts, devices or systems meant to solve the challenge are built in the laboratory. Common standards like the BioBrick or Golden Gate assembly are followed using construction tools like DNA synthesis and cloning. During implementation in the lab, unexpected difficulties often arise requiring detailed analysis and troubleshooting. After solving these problems, biological devices or systems are constructed and ready for testing.
In the third stage of the engineering cycle, generated devices or systems are validated in the laboratory. They are tested and characterized under various conditions, providing insights towards their capability to solve the current challenge. Like the second phase, unexpected problems might occur during testing which are constantly analyzed and optimized by scientists. All information gained from the testing phase is extensively evaluated.
Ultimately, all results are analyzed and as much information as possible is considered for adjusting the design to optimize the behavior of the biological device or system. The engineering cycle is then repeated starting with a new design phase followed by implementation, testing and learning. Only when the device or system is capable of solving the challenge at hand, the cycle is finished and a novel SynBio application has been created.
In March 2024, we decided to genetically engineer Bacillus subtilis to be used for textile waste degradation. As we divided our project into two main strategies, namely finding functional enzymes via induced expression as well as spore surface display, we visualized small gearwheels representing these strategies next to the main wheel in the middle representing our entire project (Fig. 3).
By going through our project phases, including literature research, subcloning, induced expression and spore surface display see Results page), we managed to turn the gearwheels further and further, coming closer to our final aim of textile degradation.
In the following section, we describe the design of our solution to the textile waste problem, its implementation and testing in the laboratory as well as all information we have gained from evaluating results.
As soon as our team decided on the project topic “ReFiBa – Enzyme based Recycling of Textile Fibers using Bacillus subtilis”, we intensely reviewed literature and looked at the work of previous iGEM teams that addressed the textile waste issue.
For example, Chalmers-Gothenburg 2020 aimed to produce an enzyme cocktail produced by Escherichia coli for elastane degradation, Edinburgh 2021 immobilized cellulases on silica beads and Greatbay SCIE 2022 tested cell surface display systems for textile degradation using PETases and cellulases. Moreover, we were highly influenced by current research in the working group of Prof. Thorsten Mascher focusing on protein display on B. subtilis spores, which is based on the work of the LMU Munich 2012 iGEM team.
Inspired by their work, we were able to meet several experts (see Human Practices page) who significantly shaped our project. We divided the project into two main strategies: 1) Induced expression as a reference approach, in which we aimed to find active cellulases and PETases for working with B. subtilis as an expression host, and 2) Spore surface display as a final strategy, in which chosen enzymes are immobilized on the surface of B. subtilis spores.
As previously mentioned, former iGEM teams have already worked on engineering devices and systems for textile degradation, but none so far have attempted to immobilize textile-degrading enzymes on Bacillus spores as a display platform. Thus, our project represents a unique strategy to tackle the textile waste problem.
We first searched for suitable enzymes for our aim to degrade cellulose and PET using B. subtilis. Even though fungal cellulases are most efficient in breaking down cellulose, we decided to only focus on bacterial cellulases due to difficulties associated with eukaryotic gene expression in bacteria, e.g. lacking glycosylation which may affect enzyme folding and activity.
We chose cellulase candidates out of the phylum Firmicutes (Bacillota), as B. subtilis belongs to this phylum as well. Hence, these enzymes provide the highest chance of functioning well in Bacillus without applying codon harmonization. Consequently, we excluded all fungal candidates as well as genes from other phyla (e.g. Actinomycetota) and focused on cellulases from Bacillota species. For PETase, we chose a candidate that was already codon optimized for Bacillus combined with the signal peptide SPaprE from B. subtilis (Xi et al. 2021).
We defined following criteria:
• biosafety level: S1
• phylum: Bacillota (for cellulases)
• gene size: < 2000 bp (max. 2500 bp)
• active state: monomer
• localization: extracellular (if possible)
• pH range: ≈ 5 - 8
• temperature: ≈ 30 - 60 °C
We decided on ten enzyme candidates (Tab. 1).:
Enzyme | Candidate | Species | Phylum, Class | Abbreviation | Gene size [bp] | Protein size [kDa] | Localization | References |
---|---|---|---|---|---|---|---|---|
Endoglucanase (endo-1,4-β-D-glucanase, EC 3.2.1.4) | EglS | Bacillus subtilis | Bacillota, Bacilli | BsEglS | 1500 | 55 | extracellular | MacKay et al. (1986), Aa et al.(1994), Lindahl et al. (1994) UniProtKB: P10475 |
EglA | Bacillus pumilus | Bacillota, Bacilli | BpEglA | 1980 | 73 | extracellular | Lima et al. (2005) UniProtKB: Q5YLG1 |
|
CelA | Acetivibrio thermocellus | Bacillota, Clostridia | AtCelA | 1434 | 52-60 | extracellular | Béguin et al. (1985), Schmidt et al. (2001 and 2002), van der Veen et al.(2013), Kitago et al. (2006,2007,2011), Weng et al. (2022) UniProtKB: A3DC29 |
|
CelG | Acetivibrio thermocellus | Bacillota, Clostridia | AtCelG | 1701 | 63 | extracellular | Lemaire and Béguin (1993), UniProtKB: Q05332 |
|
Exoglucanase (1,4-β-cellobiosidase, EC 3.2.1.176) | CelO | Acetivibrio thermocellus | Bacillota,Clostridia | AtCelO | 1983 | 75 | unknown | Zverlov et al. (2002), Ahmad et al. (2023), UniProtKB: Q9L3J2 |
CelS | Acetivibrio thermocellus | Bacillota, Clostridia | AtCelS | 2226 | 75-82 | extracellular | Wang et al. (1993), Wang and Wu (1993), Kruus et al. (1995), Guimarães et al. (2002), Wilson et al. (2013), UniProtKB: A3DH67 |
|
β-Glucosidase (1,4-β-glucosidase, EC 3.2.1.21) | BglA | Bacillus halodurans | Bacillota, Bacilli | BhBglA | 1344 | 52 | intracellular | Naz et al. (2010), Takami et al. (2000), UniProtKB: Q9KBK3 |
BglB | Paenibacillus polymyxa | Bacillota, Bacilli | PpBglB | 1347 | 52 | intracellular | González-Candelas et al. (1990), Isorna et al. (2006 and 2007), Tsabitah et al. (2024), UniProtKB: P22505 |
|
BglA | Acetivibrio thermocellus | Bacillota, Clostridia | AtBglA | 1347 | 51-53 | intracellular | Gräbnitz et al. (1991), Ahmed et al. (2019), UniProtKB: P26208 |
|
PETase (Polyethylene terephthalate hydrolase, EC 3.1.1.101) | BhrPETase | Bacterium HR29 | Chloroflexi | BhrPET | 867 | 25 | extracellular | Kato et al. (2018), Wang and Wang (2021), Xi et al. (2021), Wang et al. (2024), UniProtKB: A0A2H5Z9R5 |
After choosing ten enzyme candidates, we started to design the biological parts in the cloud-based design tool Benchling. Based on the literature, we designed translational units containing the optimal ribosome binding site (RBS) “TAAGGAGG” for expression in B. subtilis, a 7 bp spacer “AAAAAAA” (Vellanoweth & Rabinowitz 1992) and the coding sequence (CDS) of each enzyme candidate (Fig. 4). Design of all parts is documented on respective Parts Registry pages (links on Contribution page).
For cloning via the BioBrick assembly standard, these parts were flanked with the BioBrick prefix and suffix sequences and 4 bp “GATC” as an overhang to facilitate binding of restriction enzymes. The restriction sites EcoRI, XbaI, SpeI, PstI and NotI were therefore removed from the CDS. To make our parts compatible with the Type IIS standard, BsaI and SapI sites were removed as well. Additionally, HindIII sites were taken out to enable cloning with pET expression vectors in E. coli if needed. This was achieved by codon exchange using the codon usage table of Bacillus subtilis (Codon Usage Database Kazusa). These adjusted biological parts were ordered via gene synthesis from IDT.
Ultimately, these parts were to be cloned into the small vector pSB1C3 (Fig. 5), which is commonly used for subcloning, as it comprises a small vector backbone without promoters or other parts.
In this project phase, we focused on testing the functionality and activity of our chosen enzyme candidates. For that purpose, parts were cloned into xylose-inducible expression vectors to overexpress the target genes. We chose both replicative and integrative vectors, pBS0E-xylR-PxylA and pBS2E-xylR-PxylA, respectively (Fig. 6, Fig. 7), with a xylose-inducible promoter for induced expression and a xylose repressor to decrease basal promoter activity (Popp et al. 2017).
Whereas replicative plasmids provide a high copy number and result in high concentrations of target proteins, genomic integration (in this case into the lacA locus) ensures high stability but results in lower protein concentrations. For comparison, we planned to clone parts in both vectors.
After choosing enzymes for spore surface display, we finally entered the last project phase including completely new construct design. In contrast to our reference approach (induced expression), in which the enzyme candidates were secreted due to a signal peptide (except β-glucosidases), these signal peptides are not required for protein immobilization on B. subtilis spores.
Therefore, each signal peptide was removed from the CDS of the chosen enzymes BsEglS, AtCelO, AtCelS and BhrPET. As the chosen β-glucosidases (BhBglA, PpBglB) do not contain natural signal peptides, the original CDS could be used for construct design. Design of all parts is documented on the respective Parts Registry pages (links on Contribution page).
To express target genes only under sporulation, the sporulation-dependent promoter PcotYZ of B. subtilis was chosen. In previous studies, this promoter has so far provided the highest activity for spore surface display (Bartels et al. 2018, unpublished data of Elif Öztel). Downstream from the promoter, the same RBS with a 7 bp spacer followed as in the construct design of our reference approach.
To anchor the target enzymes on the spore surface, these were fused to the N-terminus of the anchor protein CotY. This anchor is located in the crust, the outermost spore layer, and has been shown to be well suited for protein immobilization (McKenney et al. 2013, Bartels et al. 2018, Lin et al. 2020). We chose to start testing N-terminal fusions because they provided higher reporter signals when working with sfGFP (Bartels et al. 2018). Due to time constraints during the iGEM competition, we were not able to test C-terminal fusions.
Moreover, different linkers between the fused target enzyme and anchor protein were analyzed, as these proteins may affect the folding and stability of each other and, eventually, lead to misfolding and reduced activity. Whereas flexible linkers promote the movement of joined proteins and are usually composed of small amino acids (e.g. Gly, Ser, Thr), rigid linkers are usually applied to maintain a fixed distance between the domains (Chen et al. 2013).
We chose to test three linkers: 1) A short flexible GA linker (L1) encoding the small amino acids Gly and Ala, 2) A long flexible linker (GGGGS)4 (L2) which is one of the most common flexible linkers consisting of Gly and Ser residues and 3) A rigid linker GGGEAAAKGGG (L3) in which the EAAAK motif results in the formation of an alpha helix providing high stability (Chen et al. 2013).
Following the CDS of cotY, we inserted a spacer consisting of 10 bp of the natural genome sequence downstream from the cotYZ operon. This created space before the terminator and ensures that the ribosome is able to read the full length of the CDS. Secondary mRNA structures were predicted by RNAfold WebServer and showed no differences in stem loop formation (Fig. 8). Each construct ends with the terminator B0014, a bidirectional terminator consisting of B0012 and B0011.
The entire construct was flanked with the BioBrick prefix and suffix, allowing for cloning via BioBrick assembly standard (). The vector pBS1C (Fig. 10) from the Bacillus BioBrickBox was used as an integrative plasmid backbone (Radeck et al. 2013), thereby enabling genomic integration into the amyE locus of B. subtilis.
As soon as the experimental plan was finalized, we started with the subcloning of ordered biological parts into the small vector pSB1C3, which was provided by the laboratory collection. After restriction digest and ligation, the plasmids were transformed into chemically competent Escherichia coli DH10β cells. Successful transformations were verified by Colony PCR and sequencing. Respective cloning results can be found on the Results page and respective Parts Registry pages (links on Contribution page).
In our reference approach, genes of interest were cloned into inducible expression vectors to verify enzyme functionality in B. subtilis as a host organism. Parts were amplified and cloned into xylose inducible vectors pBS0E-xylR-PxylA (replicative) and pBS2E-xylR-PxylA (integrative). After restriction digest and ligation, expression plasmids were transformed into chemically competent E. coli DH10β cells and transformants were verified by Colony PCR and sequencing.
Due to the limited time available within the framework of the iGEM competition, our reference approach was essential to test the functionality of the enzyme candidates in B. subtilis before subjecting immobilized proteins to testing.
Subsequently, these generated expression plasmids were transformed into B. subtilis WB800N, a genetically engineered variant of W168 in which eight extracellular proteases are deleted (Jeong et al. 2018). As we seek to secrete most of our enzyme candidates, this strain is perfectly suited for our reference approach. Successful transformants were verified by Colony PCR.
All results related to cloning as well as transformations of E. coli and B. subtilis can be found on the Results page and the respective Parts Registry pages (links on Contribution page).
In our final spore display strategy, we aimed to fuse genes of interest to an anchor gene out of the spore crust of B. subtilis. In sporulation-dependent expression, enzymes will be immobilized on the spore surface. All biological parts, including enzyme candidates with varying linkers, the promoter PcotYZ, the terminator B0014 and the anchor gene cotY were amplified, assembled and ligated into pBS1C.
After restriction digest and ligation, plasmids were transformed into chemically competent E. coli DH10β cells and transformants were verified by Colony PCR and sequencing. The generated spore display plasmids were linearized and transformed into B. subtilis W168. Successful transformants were verified by starch assay, as the vector backbone provides genomic integration into the amyE locus encoding an amylase.
All results related to cloning, transformation of E. coli and B. subtilis can be found on the Results page and the respective Parts Registry pages (links on Contribution page).
abcdefg
In both of our strategies, induced expression and spore surface display, we faced difficulties during cloning. Especially, the ppurification of the digested vector backbone appears to be crucial for subsequent ligation. We would therefore directly purify digested vectors by PCR clean up instead of gel extraction in future experiments.
Using this approach, a re-ligation control during E. coli transformation must be included, as the original insert containing RFP is not removed. Colonies with re-ligated plasmids, however, appear pink on the plate due to the RFP insert and can therefore be distinguished from correct transformants via pink-white selection.
During the construction process of our spore display constructs, cloning issues occurred as well, particularly concerning oOverlap PCR. Parts were most likely misassembled due to complementary sequences, as bands of wrong sizes were detected after agarose gel electrophoresis. In the future, we would adjust the design regarding overhangs added by oligonucleotides to avoid unspecific binding and misassembly.
After having discovered background activity of the WB800N strain in initial quality assays regarding endoglucanase and β-glucosidase activity, we planned to generate WB800N deletion strains (ΔeglS, ΔbglH) to remove the background signal for our reference approach.
This idea was ultimately discarded, though, as multiple attempts to transform WB800N with the knockout constructs including antibiotic resistance cassettes failed. Consequently, we always included WB800N as a control in activity assays. We learned that deletion strains are not necessary for testing enzyme activity, as a visible difference between some expression strains (e.g. endoglucanase BsEglS) and the control could be detected.
Regarding the B. subtilis transformation itself, we discovered that the protocol has to be adjusted when using the strain WB800N. To test the secretory expression of additional enzyme candidates, we would recommend earlier addition of DNA (OD600 ≈ 0.7) to not miss the timepoint of competence. Cells should then be grown until OD600 ≈ 1.1-1.3, following the same procedure as in the initial protocol (see Experiments).
As the earlier addition of DNA has no negative effect on either DNA or cells, this approach could also be applied to the transformation of W168. In transformation experiments for the spore display strategy, we implemented this adjustment and obtained successful transformants.
BsEglS:
BpEglA:
AtCelA:
AtCelG:
AtCelO:
AtCelS:
BhBglA:
PpBglB:
AtBglA:
BhrPET: