Week 2 HW: DNA Design

Designing the Code for Sustainable Bioplastics

In the previous assignment, I proposed the biomanufacturing of Polyhydroxybutyrate (PHB), a biodegradable plastic, utilizing the microalgae Chlorella vulgaris and Cupriavidus necator at an industrial scale[cite: 3]. However, as detailed in my final HTGAA project, to prototype and validate the biosynthetic pathway, my experimental chassis of choice for this initial proxy genetic design will be Escherichia coli BL21(DE3)[cite: 3]. Below, I outline the process of designing my DNA construct using Kernel.

Part 0: Set up in Kernel

To keep everything organized within the course workspace, I have created my corresponding repository and folder.

• Kernel location: My project is hosted in the HW2 Folder - Fabrizio Flores repository within the Construct folder.

Part 1: Define your expression target

• What molecule are you expressing? I am expressing a set of three enzymes that make up the complete biosynthetic operon for PHB: beta-ketothiolase (PhaA), NADPH-dependent acetoacetyl-CoA reductase (PhaB), and polyhydroxyalkanoate synthase class I (PhaC) [1]. Additionally, I am including a superfolder green fluorescent protein (sfGFP) reporter gene under an independent promoter to validate the transcription and presence of the construct.
• What is the sequence? The coding enzymatic sequences originate from the natural phaABC operon characterized by Peoples and Sinskey (1989) in Cupriavidus necator H16 (formerly Alcaligenes eutrophus) [1]. These sequences have undergone codon optimization for expression in E. coli. The sfGFP reporter sequence is a standard part derived from Addgene.
• What host or chassis? The host will be E. coli BL21(DE3). This is a widely validated prokaryotic strain ideal for high protein expression due to its carriage of the bacteriophage T7 RNA polymerase gene integrated into its chromosome.
• Where does the product need to go? The final product of this metabolic pathway (the PHB polymer) accumulates as insoluble granules directly in the cytoplasm of the bacterium. The enzymes (PhaA, PhaB, PhaC) also operate at the cytoplasmic level. Therefore, signal peptides or secretion tags are not required; product recovery will necessitate the disruption of the host cell.

Part 2: Design your genetic parts

To build this polycistronic cassette in E. coli, I selected the following parts, ensuring a modular architecture compatible with the assembly:

Part 3: Assemble your construct

The final design was assembled in Kernel following a strict 5’ → 3’ order, generating an insert of 4,725 base pairs:

[pET-28a(+) Backbone] → [T7 Promoter] → [RBS_1] → [phaA_opt] → [RBS_2] → [phaB_opt] → [RBS_3] → [phaC_opt] → [T7 Terminator] → [BBa_J23119 Promoter] → [sfGFP] → [BBa_B0015 Terminator]

During assembly, I verified that the reading frame remained intact, confirming that each coding sequence initiated with its respective start codon (ATG) and finished correctly with its predicted stop codons. I ensured that no internal premature stops existed post-optimization. (Note: An exported schematic from Kernel is included in the attached delivery files, illustrating the exact order and orientation of these parts).

Part 4: Optimize and quality-check

I utilized the integrated design tools to refine and validate the construct’s sequence before proceeding with manufacturing considerations:

• Codon Optimization: The native coding sequences of the phaABC operon from Cupriavidus necator H16 were subjected to codon optimization for heterologous nuclear expression in E. coli BL21 (DE3). This adaptation is critical because C. necator (high GC content) and E. coli possess markedly different codon usage biases. The optimization of the DNA sequence prevents the depletion of the rare tRNA pool in the host, mitigating metabolic stress, avoiding ribosomal pausing that could truncate translation, and ensuring that all three enzymes fold correctly with a high yield. Additional constraints were applied to normalize the GC content and prevent the inadvertent creation of problematic restriction sites.
• Sequence Search (BLAST): Upon executing a sequence search analysis, the conceptual translation of nucleotides to amino acids confirmed 100% homology with the wild-type beta-ketothiolase (phaA), acetoacetyl-CoA reductase (phaB), and PHA synthase (phaC) enzymes from Cupriavidus necator H16 [1]. However, the underlying DNA sequence is reported as “novel” and shows no exact matches in nucleotide databases due to the extensive modifications introduced during the codon optimization step.
• Safety Screen: I ran the biosecurity evaluation on the entire construct, a process analogous to the SecureDNA biosecure screening framework. This tool examines the sequence for homologies with known toxins, virulence factors, regulated pathogens, or sequences associated with dual-use research. The construct passed the evaluation without raising any alerts. This result is expected and consistent with the design, as the phaABC operon exclusively encodes enzymatic pathways related to metabolic carbon storage, and the intended host (E. coli BL21 DE3) is a non-pathogenic laboratory strain operating under Biosafety Level 1 (BSL-1) conditions.

Part 5: From design to build

The translation of an in silico design into a physical DNA molecule requires careful consideration of synthesis and assembly technologies, especially given the dimensions of the construct.

• How would you build it? The design comprises a substantial insert of 4,725 base pairs (operon + reporter). When integrated into the pET-28a(+) expression vector, the resulting plasmid reaches 10,017 bp. Given the significant length and inherent complexity of a polycistronic operon (with multiple ribosome binding sites), the optimal construction strategy is Whole Plasmid Synthesis via Twist Bioscience. While in vitro assembly techniques, such as the isothermal enzymatic assembly described by Gibson et al. (2009) [2], are powerful for joining fragments up to several hundred kilobases, manually assembling multiple large fragments introduces substantial risks of cloning failure, unwanted homologous recombination, and considerable troubleshooting time. Outsourcing the synthesis of the entire plasmid to a commercial provider ensures sequence accuracy.
• What does synthesis constrain? Commercial de novo DNA synthesis imposes strict constraints, particularly regarding accuracy and yield. As Carr et al. (2004) [3] highlight, oligonucleotide synthesis is intrinsically limited by error rates (deletions, insertions, mutations), making robust protein-mediated error correction methods essential for the accurate assembly of large constructs. Furthermore, synthesis algorithms struggle with highly repetitive sequences, GC-rich extremities, or stable secondary structures. The codon optimization performed in Part 4 not only improved expression in the host but also mitigated these manufacturing risks by normalizing the GC content and eliminating undesirable structural complexities, ensuring that the sequence complies with Twist Bioscience’s synthesis acceptability guidelines without requiring costly fragment assembly modifications.
• Prototype vs. Scale: The synthesis approach differs notably between initial prototyping and large-scale deployment. For a rapid validation of the 10 kb design, I plan to prototype the plasmid using an E. coli Cell-Free Protein Synthesis (CFPS) system. This cell-free environment, dispensed with nanoliter precision using an Echo525 acoustic liquid handler, allows the confirmation of the enzymes’ activity in a matter of hours without the delays associated with stable cell transformation. However, to scale production in controlled 100L bioreactors (Cloud Lab), the dynamics change. At that scale, the physical construction is housed within living bacteria subjected to an immense metabolic burden. Therefore, scaling would rely on continuous kanamycin supplementation to maintain selective pressure and ensure that the bacterial population does not lose (expel) the pET-28a(+) plasmid over prolonged culture cycles.

Part 6: What don’t you know yet?

Despite a rigorous in silico design guided by metabolic models (FBA), significant biological uncertainties persist at the edge of this design that will require empirical resolution:

1. Metabolic Burden and T7 Toxicity: The T7 system provides extremely potent and rapid transcription. Unknown: Will the massive and simultaneous overexpression of the three enzymes PhaA, PhaB, and PhaC provoke a severe metabolic burden, inducing a stress response and cell death in E. coli before a significant amount of the PHB polymer can accumulate? Validation: This uncertainty will be tested by measuring in vivo growth curves (optical density, OD600) at various post-induction time points with different concentrations of the chemical inducer IPTG. If unacceptable toxicity is observed, the next design iteration will explore a weaker and more tunable inducible promoter (e.g., the pBAD system) to balance biomass growth with pathway expression.
2. Granule Formation in the Absence of Phasin-Associated Proteins: E. coli lacks the native evolutionary machinery for bioplastic accumulation. Unknown: Will the intracellularly synthesized PHB polymer chains fold and assemble into discrete organized granules within the E. coli cytoplasm lacking “phasins”, the native structural coating proteins from C. necator that regulate granule size and stability? Or will the polymer precipitate in a disorganized manner, forming toxic inclusion bodies? Validation: This will be evaluated microscopically and via an in vivo stain with the lipophilic dye Nile Red, measured with high resolution on a PHERAstar FSX microplate reader. If polymer accumulation disrupts cell viability due to improper granule formation, subsequent construct designs will require the co-expression of sequences encoding structural phasins.

References

• [1] O. P. Peoples and A. J. Sinskey, “Poly-beta-hydroxybutyrate (PHB) biosynthesis in Alcaligenes eutrophus H16. Identification and characterization of the PHB polymerase gene (phbC),” Journal of Biological Chemistry, vol. 264, no. 26, pp. 15298–15303, Sep. 1989.
• [2] D. G. Gibson et al., “Enzymatic assembly of DNA molecules up to several hundred kilobases,” Nature Methods, vol. 6, no. 5, pp. 343–345, Apr. 2009, doi: 10.1038/nmeth.1318.
• [3] P. A. Carr et al., “Protein-mediated error correction for de novo DNA synthesis,” Nucleic Acids Research, vol. 32, no. 20, p. e162, Nov. 2004, doi: 10.1093/nar/gnh160.

AI and Prompt Citations

• AI Used: Gemini (Google) Assitant, 3.1 Pro vesion
• Prompts Used: The drafting and conceptualization of this genetic design were assisted through an interactive dialogue, utilizing the core logic of my final HTGAA project (Modeling and Automating Polyhydroxybutyrate Bioplastic Production in Microalgae). The AI model was employed as a design review assistant to align codon usage, synthesis constraints, and the context of the homework goals.