Week 2 HW: DNA Design

Week 2 HW: DNA Design

Recombinant Collagen–Elastin Protein for an Off-the-Shelf Wound Scaffold

My final project asks whether an engineered recombinant fibrous protein could be biomanufactured in advance and later fabricated into an acellular scaffold for large-area wounds such as major burns.

For this homework, I translated that material concept into a DNA expression design.

The production organism is only the manufacturing system. The final wound scaffold would contain purified protein, not living engineered bacteria.

0. Kernel Workspace

I created my work in:

HTBAA 2026 → Hartnell → DanyAitAissa-WoundScaffold

My Kernel construct is named:

DanyWoundScaffold_Expression_Construct

The custom coding sequence is:

Scl2VCL_G4S3_ELP20

1. Define the Expression Target

1.1 What am I expressing?

My target is a recombinant collagen-like / elastin-like fusion protein:

His6 – Scl2 V+CL – (G4S)3 – ELP20

The complete custom CDS is 1299 bp and translates to 432 amino acids followed by one stop codon.

The design contains four main protein elements:

1. N-terminal His6 tag
   Added as a purification handle.

2. Scl2.28 V+CL region
   Derived from the collagen-like protein Scl2.28-1 from Streptococcus pyogenes.

3. (G4S)3 flexible linker
   Protein sequence:

   GGGGSGGGGSGGGGS

4. ELP20 domain
   A de novo elastin-like polypeptide containing 20 repeats of the motif VPGXG.

The ELP20 sequence uses a 4:1 Val:Ala ratio at the X position:

VPGVGVPGVGVPGVGVPGVGVPGAGVPGVGVPGVGVPGVGVPGVGVPGAGVPGVGVPGVGVPGVGVPGVGVPGAGVPGVGVPGVGVPGVGVPGVGVPGAG

1.2 Sequence source

The collagen-like portion was based on:

Scl2.28-1, GenBank AY069936.1

For this educational prototype, I defined residues 35–344 as the Scl2 V+CL segment used in the fusion. This boundary is a design choice for this homework rather than a claim that these are canonical database-defined domain boundaries.

The His6 tag, linker, and ELP20 region were designed as synthetic additions.

1.3 Expression host

The proposed expression host is:

Escherichia coli BL21(DE3)

I selected BL21(DE3) because the construct uses a T7-based expression system.

1.4 Localization

The intended expression is intracellular.

I do not want a secretion signal peptide. After expression, the cells would be harvested and lysed, and the recombinant protein would be purified before scaffold fabrication.

2. Design the Genetic Parts

2.1 Overall expression cassette

The construct assembled in Kernel is:

5’ → T7-lacO promoter → BBa_B0034 RBS → Scl2VCL_G4S3_ELP20 CDS → BBa_B0016 terminator → 3’

2.2 Parts used

The complete CDS begins with a start codon, contains no internal stop codons, and ends with a terminal stop codon.

2.3 Backbone decision

My intended vector context is a pET-28a(+)-type bacterial expression backbone, which is compatible with T7 expression in BL21(DE3) and provides bacterial plasmid replication and kanamycin selection.

However, I could not find a clean empty pET-28a backbone in the Kernel repository search.

One Kernel result, BBa_K4862003, was based on pET28a but already contained a Zika NS3 insert. I did not use it because importing an existing expression construct with an unrelated CDS would not represent my design.

I therefore kept the Kernel object as an expression cassette and documented pET-28a(+) as the intended future vector context rather than adding an inappropriate backbone just to complete the circular map.

3. Assemble the Construct

I manually assembled the expression cassette in Kernel in the following order:

T7-lacO promoter → BBa_B0034 → Scl2VCL_G4S3_ELP20 → BBa_B0016

Each part has a specific role:

• the T7-lacO promoter initiates inducible transcription;
• BBa_B0034 supports translation initiation;
• Scl2VCL_G4S3_ELP20 encodes the fusion protein;
• BBa_B0016 terminates T7 transcription.

The custom CDS was created as a CDS (+start +stop) / Generic protein part in Kernel.

The final protein architecture encoded by the CDS is:

His6 → Scl2 V+CL → (G4S)3 → ELP20

4. Optimize and Quality-Check

4.1 Codon optimization

I ran Kernel’s coding-sequence optimization workflow using:

• Host: Escherichia coli
• Optimizer: Simple Codon Optimizer v0.3
• Preset: None

The original CDS was 1299 bp.

However, Kernel’s output was 1380 bp and was labeled:

SP5-Scl2VCL_G4S3_ELP20

Kernel had added an SP5 signal peptide to the N-terminus. Its signal-peptide analysis also predicted a cleavage site near the junction between SP5 and my original protein.

This was not consistent with my design because I specifically want intracellular expression with no secretion signal.

The additional sequence increased the CDS by 81 bp, equivalent to 27 additional amino acids. Therefore, this output was not simply a synonymous recoding of my original protein.

I rejected the SP5-containing optimization result and retained my original 1299-bp CDS.

This was useful because it showed that an automated optimization output still needs to be checked against the biological design requirements.

4.2 Sequence search / BLAST

I first attempted to use the sequence-search workflow through Kernel, but the BLAST function was not accessible in the interface available to me.

Rather than claim that Kernel had performed a search that I could not actually run, I copied the complete 1299-bp CDS from Kernel and used NCBI BLASTn against the standard nucleotide nr/nt database.

A top specific hit was:

Streptococcus pyogenes collagen-like protein 2 (scl2), scl2.28-1 allele

GenBank: AY069936.1

For this hit:

• Query cover: 72%
• Percent identity: 84.07%
• E-value: 0.0

Other high-ranking results were also S. pyogenes Scl2 sequences and synthetic constructs containing the Scl2 gene.

The result makes sense because only part of my fusion is derived from natural Scl2. The remaining sequence contains the synthetic His6 tag, (G4S)3 linker, and ELP20 domain, so I would not expect the entire 1299-bp fusion to match the natural Scl2 gene.

4.3 Biosafety sequence screening

I screened the complete 1299-bp Scl2VCL_G4S3_ELP20 CDS using the SecureDNA Biosafety Sequence Scanner available through Kernel.

The result was:

No flagged sequences detected.

This means the sequence scanner did not identify a flagged sequence match in the submitted CDS. It does not by itself establish that a future material or manufacturing process is clinically safe.

5. From Design to Build

5.1 How I would build the DNA

For a first physical prototype, I would order the custom fusion CDS as synthetic DNA.

The collagen-like and elastin-like regions are both repetitive, so synthesis of the full sequence may be more difficult than synthesis of a typical non-repetitive gene.

If a synthesis provider flagged the 1299-bp CDS as difficult, I would divide it into two or three overlapping fragments and assemble them into a compatible expression vector.

A logical approach would be Gibson assembly.

The basic workflow would be:

1. synthesize the CDS as one or several DNA fragments;
2. prepare a compatible expression backbone;
3. assemble the overlapping fragments and vector;
4. transform a cloning strain;
5. sequence-verify the complete CDS and all junctions;
6. introduce the verified construct into BL21(DE3);
7. perform small-scale expression testing.

Gibson et al. (2009) demonstrated assembly of overlapping DNA molecules without requiring a restriction site at every junction.

5.2 Sequence errors and repetitive DNA

A correct digital sequence does not guarantee that the synthesized DNA will be error-free.

The repeated collagen-like and ELP motifs could create:

• synthesis difficulty;
• recombination or deletion;
• cloning instability;
• sequence-verification problems.

Carr et al. (2004) demonstrated DNA error-correction approaches, while Tian et al. (2004) and Kosuri et al. (2010) illustrate approaches for scalable synthetic DNA production.

For this project, the practical lesson is that the final plasmid would need complete sequence verification before expression testing.

5.3 Prototype versus manufacturing scale

DNA construction and protein manufacturing are different problems.

For the first prototype, I may need to synthesize and assemble the gene.

At larger scale, I would not synthesize a new gene for every scaffold.

Once the plasmid and expression strain were validated, a qualified cell bank could be used repeatedly for fermentation, protein purification, and scaffold production.

6. What I Do Not Know Yet

A later design cycle could also vary ELP repeat number, linker design, or the relative collagen-like and elastin-like content if the first protein does not have useful scaffold properties.

AI Use Statement

I used ChatGPT (OpenAI) throughout this assignment as a design and troubleshooting assistant.

I used it to:

• translate my wound-scaffold idea into a defined fusion-protein architecture;
• organize the promoter, RBS, CDS, terminator, host, and vector strategy;
• troubleshoot the Kernel interface step by step;
• interpret the unexpected SP5 signal-peptide result from Kernel’s codon optimizer;
• interpret the NCBI BLAST results;
• organize my results into this homework write-up.

Example prompts included:

• "Help me turn my wound scaffold idea into an exact recombinant expression target."
• "What parts do I need for expression in E. coli BL21(DE3)?"
• "Tell me step by step what to click in Kernel."
• "Why did Kernel add SP5 during codon optimization?"
• "Interpret these BLAST results for my construct."
• "Write my Week 2 homework in Markdown based only on what we actually did."

I also attempted to use Kernel AI to search for parts and troubleshoot the design. Kernel AI repeatedly returned errors, including in new conversations, so I stopped relying on it. Final part selection and analysis were performed manually using the Kernel interface, NCBI BLASTn, and SecureDNA.

AI suggestions were treated as design suggestions rather than experimental results. No wet-lab experiment was performed for this homework.

References

1. Gibson DG, et al. Enzymatic assembly of DNA molecules up to several hundred kilobases. Nature Methods. 2009.

2. Carr PA, et al. Protein-mediated error correction for de novo DNA synthesis. Nucleic Acids Research. 2004.

3. Tian J, et al. Accurate multiplex gene synthesis from programmable DNA microchips. Nature. 2004.

4. Kosuri S, et al. Scalable gene synthesis by selective amplification of DNA pools from high-fidelity microchips. Nature Biotechnology. 2010.

5. Streptococcus pyogenes collagen-like protein 2, Scl2.28-1. GenBank AY069936.1.

6. iGEM Registry parts BBa_K2406020, BBa_B0034, and BBa_B0016.

7. NCBI BLASTn, standard nucleotide nr/nt database.

8. SecureDNA Biosafety Sequence Scanner.