Bioreactor Design

Overview of the Bioreactor Design cycles

Cycle 1 – Concept and Initial Design

Goal

Develop a modular, low-cost algal bioreactor with a working volume above 10 L. The reactor should be easily assembled from off-the-shelf components and 3D-printed parts, with basic automation for sensing, control, and data logging.

Design

We began by conceptualizing the overall mechanism and geometry of the reactor. The first 3D parts were designed, including the main base and acrylic tube mounts. Early focus was placed on modularity and printability—parts that could be printed on standard FDM printers and assembled without specialized tools.

Build

The first 3D prints revealed several issues: slicing profiles were not optimized, and tolerances for the acrylic tubes were too tight. The spacing had not been properly calculated, leading to poor fits.

Learn

From this first iteration, we learned that clearance tolerances are critical for printed parts interfacing with non-printed components. Tube slots were redesigned with an additional 2 mm of space inside and outside diameters, and slicing settings were refined for consistent wall thickness.

Cycle 2 – Structural Redesign and Mechanism Development

Goal

Create a stable structure and a reliable attachment mechanism between the base and upper frame.

Design

Following the fit improvements, we designed the tripod stand and its connectors. The first attempt used a bayonet-type locking mechanism to connect the base and upper structure.

Build

Due to an error in the 3D model (a misaligned layer in the CAD file), the bayonet mechanism failed to engage properly. Instead of reprinting the large base, we adapted by designing and printing replacement nodes. The faulty parts were removed, and the new nodes were screwed into the old base, allowing reuse of large prints while restoring functionality.

Learn

This cycle reinforced the importance of designing for repairability and modular fixes. The modular node approach saved time and material while enabling rapid iteration.

Cycle 3 – LED and Lid Assembly

Goal

Integrate lighting and finalize enclosure geometry for a functional photobioreactor.

Design

We designed the LED module, inner lid, and outer lid. The LEDs were essential for algal growth but initially too tight in their mounts. The module was revised with proper spacing, and a holding structure was integrated into the inner lid for better mechanical stability.

Build

During assembly, we discovered that the LED strips were longer than the acrylic tubes, which caused alignment and enclosure issues. To solve this, the lid design was modified to both cover the LED extensions and secure the modules firmly to the top.

Learn

LED housing must allow for thermal expansion and manufacturing variances in strip length. The lid redesign improved accessibility, safety, and alignment.

Cycle 4 – Electronics Integration and Control

Goal

Add sensing, control, and data-logging components to automate the system.

Design

We integrated sensors, control wiring, and a Raspberry Pi for process automation. Initially, there was no allocated space for the Pi or its cabling within the main cap and rod assembly.

Build

To solve this, we designed a dedicated Pi case with a custom mounting mechanism that screws directly onto the main cap. The wiring was routed through the structural rods, and software was configured for temperature control.

Test

Electronics and software were bench-tested successfully. Due to safety rules at the Grand Jamboree, full wet testing was not conducted because the system had not yet been verified as sterilizable for biological use.

Learn

The electronics system functioned as expected. Future work will focus on sterilization compatibility, automated filtration, and improved cable management.

Cycle 5 – Future Development

Insights and Next Steps

Filtration: Nylon filters were found suitable for harvesting algae. The next step is an automated filtration and wash module to streamline biomass collection and cleaning.

Automation: Expanding control to include flow rate and nutrient dosing would move the reactor closer to continuous operation.

Transformation of TFD-EE in C. reinhardtii

Overview of the Plasmid Design cycles

Cycle 1 - Initial Construct

Design

Our aim was to transform the green microalga Chlamydomonas reinhardtii to enhance the bioaccumulation of rare earth elements (REEs) by introducing the TFD-EE gene into the nuclear genome. Although C. reinhardtii is well characterized and its genome has been sequenced [11], engineering the organism remains challenging [3]. Historically, much previous work has focused on chloroplast engineering rather than the nucleus [1,2]. In the nuclear genome, extensive transgene silencing [3] and a very low rate of homologous recombination have led many groups to rely on non-homologous recombination despite the drawbacks of random insertion [4].

To address this, we implemented both approaches and compared them in terms of transformation efficiency, extent of gene silencing, and experimental effort. The TFD-EE gene encodes a de novo–designed REE-binding protein inspired by the calmodulin-related lanthanide-binding protein lanmodulin [5,6]. This protein binds REE cations with very high affinity, making it an attractive candidate for our project [5,7,8]. Moreover, because it is inspired by a naturally occurring protein ultimately related to ubiquitous calmodulin, the likelihood of adverse interference with C. reinhardtii metabolism is reduced [9].

Designing a reliable plasmid for overexpression of our gene of interest in C. reinhardtii was one of the first steps of our iGEM journey. This proved more involved than simply assembling documented building blocks; species-specific constraints had to be overcome. Each design required a trade-off between the simplicity needed for rapid progress and the complexity required for robust expression.

To monitor expression in vivo, we fused a codon-optimized TFD gene to a reporter. The GFP-derived protein mVenus, widely used in transformation experiments, was placed at the C-terminus of TFD and separated by a short GSSSG linker [13,14]. A C-terminal Strep-tag on mVenus was added for possible protein purification. The APHVIII gene conferred paromomycin resistance for selection [12]. To boost expression, RBCS2i and RBCS2i2 introns were inserted at intervals within both genes [14].

Using the commercially available m_Paro_mVenus plasmid as a scaffold, we placed the TFD-mVenus fusion and APHVIII under separate promoters. For TFD-mVenus, we used the well-validated HSP70A-RBCS2 fusion promoter [14]. The expression cassette was flanked by 500-bp homology-directed repair (HDR) arms to enable precise knock-in into the genome of C. reinhardtii.

We selected the SNRK2.2 locus as the homology locus to optimize selection. Knockout of this gene induces expression of a sulfatase that cleaves the sulfate group of an indigo-derived dye supplied in the medium, releasing the dye. As a result, mutant colonies stain blue and are easily distinguished from non-mutant colonies or colonies with insertions elsewhere. Combined with antibiotic selection, this readout allows us to determine transformation efficiency and the proportion of mutants generated via non-homologous versus homologous recombination.

We also considered using an auxotrophic growth marker and inserting our gene of interest into the marker’s sequence. This would have acted as a form of kill switch, preventing engineered algae from surviving outside a designed environment. However, given the convenience of the blue-dye assay and the added complexity of supplying a metabolite continuously in the bioreactor, we set this option aside and relied on containment to minimize escape risk.

Our initial plan, PCR-based amplification of the plasmid, was abandoned after our doctoral advisor highlighted the inherent inaccuracy of cloning large DNA fragments via PCR. We pivoted to bacterial cloning and designed a sixth fragment to enable replication in E. coli.

Given the construct’s size and the practical challenges of handling large DNA molecules, we opted for a minimal bacterial backbone comprising an origin of replication and an ampicillin resistance gene under its promoter. Because the origin of replication downstream of the ampicillin resistance gene contained a long poly-A stretch, we initially assumed a separate terminator would be unnecessary. This auxiliary sequence was inserted into a “safety restriction site,” which we included in our initial design for further modifications, and could be excised using BbsI.

Figure 1: An illustration of version 1 of our designed plasmid.

Build

The assembled design yielded a 5,950-bp plasmid containing all elements needed for expression in C. reinhardtii. Owing to the construct’s complexity, especially the repeated sequences within the introns, we split the plasmid into five fragments for ordering. Each fragment was flanked by BsaI sites for scarless Golden Gate assembly [15]. The ligated plasmid could be linearized at an EcoRV site positioned between the left and right homology arms prior to transformation into C. reinhardtii.

Test

We conducted the Golden Gate assembly and transformed competent E. coli for amplification. No antibiotic-resistant colonies were present, and we concluded either that the Golden Gate assembly had failed, or that our design was faulty.

Learn

When reviewing our design and asking our advisors, the bacterial infrastructure (origin and ampicillin resistance gene) turned out to cut down too much. We wanted to use an existing and proven unit and not tinker with it. Using an existing bacterial backbone to insert our parts into via Golden Gate assembly seemed promising [15]. This required rethinking our HDR strategy: our first design used two fused HDR arms to be separated during a final linearization step; the revised approach better accommodated standard cloning workflows.

Cycle 2 - Redesign

Design

The goal of our next step was to move the right HDR arm to the downstream direction of the cassette and Golden Gate assemble the whole cassette into the pTU1-A-RFP_AB (kindly provided to us by Gabriel Cervera Arriaga, who is a Master Thesis Student at the Piel Lab at ETHZ) in one step [15]. For this, we split our HDR arms-containing fragment (one of the five) into two parts via PCR. The primers were designed such that overhangs are added to both HDR arms for the Golden Gate assembly later on [15].

Figure 2: An illustration of version 2 of our designed plasmid.

Build

We ordered the designed primers from Microsynth, conducted the PCR, and then performed the Golden Gate assembly. Competent E. coli were transformed.

Test

The transformation and selection of E. coli were successful. There was a clear distinction between samples transformed with pTU1-A-RFP_AB alone (whose colonies turned red) and the ones allegedly containing our cassette (standard brownish color). A miniprep resulted in a plasmid that was fully sequenced at Microsynth. Sanger sequencing revealed that the cassette had not assembled correctly, and around half of our cassette was missing. This led us to believe that splitting our first plasmid via PCR prior to the Golden Gate did not work correctly. Sequencing the two pieces confirmed our suspicion.

Learn

Continuing to work with our initial DNA fragments seemed tedious and likely to create more problems. The existing sequences did not leave room for flexible primer design to add the overhangs needed in the preceding step. Benchling predicted the primer pairs to be prone to homodimerization and hairpin formation. That is why we decided to set aside the existing design and develop a simplified construct.

Cycle 3 - Minimal Construct

Design

We then adopted a minimalistic redesign to reduce both construct size and sequence complexity. Most introns were removed from TFD-EE (retaining one), and the mVenus reporter was omitted, with the Strep-tag retained to monitor protein production. We also simplified the APHVIII promoter, replacing the fusion promoter with RBCS2p, while keeping HSP70A-RBCS2p for TFD. The resulting 3,940-bp linear insert was directly ordered in the pTwist Amp High Copy plasmid.

Figure 2: An illustration of version 3 of our designed plasmid.

Build

Competent E. coli cells were transformed.

Test

Successful transformation of E. coli was achieved, as evidenced by bacterial growth on ampicillin-containing medium. The plasmid, isolated via miniprep, was fully sequenced at Microsynth. The sequences obtained by Sanger sequencing aligned precisely with the expected sequences, confirming successful cloning of the construct.

Learn

The amplified cassette is now ready for introduction into C. reinhardtii. Following autolysin treatment, the algae can be subjected to electroporation and transformed using CRISPR-Cas. Robust screening will be essential to identify successful transformants and to distinguish targeted homologous recombination events from random insertions.

Conclusion

• A robust expression cassette was designed for targeted gene integration in Chlamydomonas reinhardtii
• An appropriate genomic locus was identified, and the construct was adapted to facilitate site-specific insertion
• The plasmid was successfully assembled in E. coli and verified through iterative sequencing and systematic validation

Key takeaways

• Iterative Design–Build–Test–Learn cycles were essential to identify and address unforeseen technical hurdles in cloning and assembly.
• Simplification of constructs improved feasibility: removing repeated introns, reporters, and fusion promoters reduced sequence complexity and facilitated cloning.
• Using well-characterized backbones and promoters reduces design risk and accelerates troubleshooting.
• A robust selection strategy is essential for reliably identifying successful transformation events.

Sources

1. [1] Boynton, J. E.; Gillham, N. W.; Harris, E. H.; Hosler, J. P.; Johnson, A. M.; Jones, A. R.; et al. Chloroplast Transformation in Chlamydomonas with High-Velocity Microprojectiles. Science 1988, 240 (4858), 1534–1538. https://doi.org/10.1126/science.2897716.
2. [2] Esland, L.; Larrea-Álvarez, M.; Purton, S. Selectable Markers and Reporter Genes for Engineering the Chloroplast of Chlamydomonas reinhardtii. Biology 2018, 7 (4), 46. https://doi.org/10.3390/biology7040046.
3. [3] Schroda, M. Good News for Nuclear Transgene Expression in Chlamydomonas. Cells 2019, 8 (12), 1534. https://doi.org/10.3390/cells8121534.
4. [4] Sodeinde, O. A.; Kindle, K. L. Homologous Recombination in the Nuclear Genome of Chlamydomonas reinhardtii. Proceedings of the National Academy of Sciences USA 1993, 90 (19), 9199–9203. https://doi.org/10.1073/pnas.90.19.9199.
5. [5] Cotruvo, J. A., Jr.; Featherston, E. R.; Mattocks, J. A.; Ho, J. V.; Laremore, T. N. Lanmodulin: A Highly Selective Lanthanide-Binding Protein from a Lanthanide-Utilizing Bacterium. J. Am. Chem. Soc. 2018, 140 (44), 15056–15061. https://doi.org/10.1021/jacs.8b09842.
6. [6] Cook, E. C.; Showalter, S. A.; Cotruvo, J. A., Jr. Structural Basis for Rare-Earth Element Recognition by Lanmodulin. Biochemistry 2018, 57 (40), 5748–5758. https://doi.org/10.1021/acs.biochem.8b01019.
7. [7] Gutenthaler, S. M.; Tsushima, S.; Steudtner, R.; Gailer, J.; Hoffmann-Röder, A.; Drobot, B.; Daumann, L. J. Lanmodulin Peptides—Unravelling the Binding of the EF-Hand Loop Sequences Stripped from the Structural Corset. Inorg. Chem. Front. 2022, 9, 4009–4021. https://doi.org/10.1039/D2QI00933A.
8. [8] Prejanò, M.; Marino, T.; Russo, N.; Toscano, M. Periodicity of the Affinity of Lanmodulin for Trivalent Lanthanides and Actinides. Inorg. Chem. 2023, 62 (25), 9831–9842. https://doi.org/10.1021/acs.inorgchem.3c00754.
9. [9] Kawasaki, H.; Kretsinger, R. H. Structural and Functional Diversity of EF-Hand Proteins. Cold Spring Harb. Perspect. Biol. 2017, 9 (12), a035659. https://doi.org/10.1101/cshperspect.a035659.
10. [10] Caldwell, S. J.; Haydon, I. C.; Piperidou, N.; et al. Tight and Specific Lanthanide Binding in a De Novo TIM-Barrel Protein. Proc. Natl. Acad. Sci. U.S.A. 2020, 117 (49), 31149–31159. https://doi.org/10.1073/pnas.2005412117.
11. [11] Merchant, S. S.; Prochnik, S. E.; Vallon, O.; Harris, E. H.; Karpowicz, S. J.; Witman, G. B.; et al. The Chlamydomonas Genome Reveals the Evolution of Key Animal and Plant Functions. Science 2007, 318 (5848), 245–250. https://doi.org/10.1126/science.1143609.
12. [12] Sizova, I.; Fuhrmann, M.; Hegemann, P. Chlamydomonas reinhardtii aphVIII Gene Confers Resistance to the Antibiotic Paromomycin. Protist 2001, 152 (3), 185–195. https://doi.org/10.1078/1434-4610-00068.
13. [13] Nagai, T.; Ibata, K.; Park, E. S.; Kubota, M.; Mikoshiba, K.; Miyawaki, A. A Variant of Yellow Fluorescent Protein with Fast and Efficient Maturation for Cell-Biological Applications. Nat. Biotechnol. 2002, 20 (1), 87–90. https://doi.org/10.1038/nbt0102-87.
14. [14] Crozet, P.; Navarro, F. J.; Willmund, F.; Mehrshahi, P.; Bakowski, K.; Lauersen, K. J.; et al. Birth of a Photosynthetic Chassis: A MoClo Toolkit for Chlamydomonas reinhardtii. The Plant Journal 2018, 95 (2), 304–318. https://doi.org/10.1111/tpj.13948.
15. [15] Engler, C.; Kandzia, R.; Marillonnet, S. A One Pot, One Step, Precision Cloning Method with High Throughput Capability. PLoS One 2008, 3 (11), e3647. https://doi.org/10.1371/journal.pone.0003647.