Week 2 HW: DNA Design

B2 · Homework · Class 2 · 27 Sep 2026 · Sergio O. Vazquez-Franco

A secreted hydrophobin, HFBII, as a “moisture dial” for my desert mycelium product family.

AI assistance: Claude (Anthropic). My prompts are listed at the end. In Homework #1 my intro line says “bacterial cellulose”, but the living system I proposed, and the one designed for here, is mycelium.

Part 0Set up in Kernel

My work is in the Kernel repository VazquezFranco-DesertMycelium, in the folder of the same name.

Part 1Define your expression target

Molecule

One protein, hydrophobin II (HFBII). Mycelium is built by many enzymes, so there is no single “mycelium protein”. I chose the protein that controls the property my research question is about: how the surface behaves toward moisture. Hydrophobins are small proteins fungi secrete to make surfaces water-repellent, so HFBII acts as a moisture dial for the five desert pieces.

Sequence

Trichoderma reesei Hydrophobin-2, UniProt P790731. The precursor is 86 amino acids: a 15-residue signal peptide (MQFFAVALFATSALA)4 and a 71-residue mature protein with 8 cysteines2:

AVCPTGLFSNPLCCATNVLDLIGVDCKTPTIAVDTGAIFQAHCASKGSKPLCCVAPVADQALLCQKAIGTF

It is a natural protein, not de novo. The DNA in my construct is a designed back-translation, so it encodes the natural protein but is not a copy of the natural gene. Structure: Hakanpää et al.3

Host

Trichoderma reesei, a filamentous fungus, as a stand-in for the mycelium from Homework #1. This is a guess: it is well characterized and already makes this protein, but the mycelium I would grow into 20 × 20 × 5 cm pieces is probably a different species, so the parts would need checking there.

Destination

Secreted. Hydrophobin works at the surface, where it self-assembles into a water-repellent layer, so the construct needs a signal peptide. I use the protein’s own leader, removed to leave the mature protein (residues 16–86).

Part 2Design your genetic parts

LayerKernel typePartSourceWhy
Transcription startPromoterpcbh1T. reesei cbh1 promoterStrong and inducible by cellulose-type sugars. The natural hfb2 gene is itself highly expressed on cellulose and lactose media2.
Translation start5′ UTR + Kozakutr-cbh1, kozak-fungalNative cbh1 leader; fungal Kozak-likeCorrect translation initiation in a fungal host.
TargetingCDS (+start −stop)sp-hfb2Residues 1–15 of P79073; designed DNAThe protein’s own leader sends it out of the cell.
Coding sequenceCDS (−start +stop)hfb2-matureResidues 16–86 of P79073; designed DNA, ends TGATAAThe mature moisture-controlling protein.
Transcription stop3′ UTR + polyAt-cbh1T. reesei cbh1 terminatorEnds the transcript cleanly; Kernel expects a 3′ UTR and polyA per promoter.
Vector contextBackbonebb-fungal-hphExisting fungal transformation plasmidReused: origin and antibiotic resistance for E. coli, plus a fungal selection marker.

Designed DNA (one common codon per amino acid, back-translated by hand):

sp-hfb2 (45 bp)
ATGCAGTTCTTCGCCGTCGCCCTCTTCGCCACCTCCGCCCTCGCC

hfb2-mature (219 bp, includes TGATAA stop)
GCCGTCTGCCCCACCGGCCTCTTCTCCAACCCCCTCTGCTGCGCCACCAACGTCCTCGACCTCATCGGCGTCGACTGCAAGACCCCCACCATCGCCGTCGACACCGGCGCCATCTTCCAGGCCCACTGCGCCTCCAAGGGCTCCAAGCCCCTCTGCTGCGTCGCCCCCGTCGCCGACCAGGCCCTCCTCTGCCAGAAGGCCATCGGCACCTTCTGATAA

Part 3Assemble your construct

Promoter5′ UTRKozakSignal peptideMature HFBII3′ UTR / polyA
Construct map of the hfb2 expression cassette

A single fungal expression cassette in a reused backbone. The signal peptide and mature sequence share one reading frame and encode the 86-amino-acid precursor, so the protein is made and then secreted. The coding region is 264 bp (45 + 213 + 6 for the TGATAA stop) and reads MQFFAVALFATSALA straight into AVCPTGLFSN….

Kernel Construct Checker result

Part 4Optimize and quality-check

Codon optimization

Codon usage matters because a host’s tRNA pools favor some codons, so a gene written with rarely used codons can be translated slowly or inaccurately. My draft already uses common, GC-ending codons.

Codon optimization before and after

Sequence search

Checks whether my sequence matches known genes. It encodes a natural protein through a designed DNA sequence, so I compare the protein-level and DNA-level matches separately.

Sequence search result

Safety screen

Screens the construct against databases of sequences of concern.

Safety screen result

Part 5From design to build

How would I build it?

The gene is small (264 bp), but the backbone and regulatory parts are much longer. I would order the coding region as synthetic DNA with overlaps to its neighbors and join it to the regulatory parts and backbone by Gibson assembly5, where overlapping fragments are joined in one isothermal reaction by an exonuclease, a polymerase and a ligase. A chip-synthesized oligo pool pays off when many genes or variants are built together6,8; for one gene it is not worth it.

What does synthesis constrain?

Synthesis has error rates that grow with length and complexity. Error correction such as the mismatch-binding approach of Carr et al.7 removes faulty molecules, and I would sequence-verify the final construct. My design has two trouble spots. My back-translation is about 69% GC (calculated from the designed sequence), because I picked C/G-ending codons, and very high GC makes synthesis and assembly harder. The paired cysteines also create short repeats such as TGCTGC. Shorter fragments of moderate GC joined by overlaps are easier to make and can each be checked before assembly.

Prototype vs. scale

The prototype is one construct built once: a synthetic coding fragment, Gibson assembly, sequencing, then fungal transformation. At product scale DNA is not used up: once a strain carries the cassette, the fungus copies it as it grows, so five 20 × 20 × 5 cm pieces (about 15 L of growth material) come from growing mycelium, not from making more DNA. DNA synthesis scale matters for design exploration. To tune moisture behavior I would want several expression levels (promoter and UTR variants), and there a pooled oligo approach6 becomes attractive.

Part 6What don’t I know yet?

1 · Expression level and moisture behavior

I do not know how much hydrophobin gives the surface I want. There is also a tension: hydrophobin repels water, while my pieces should respond to moisture. To find out: build two or three variants with weaker and stronger promoters or UTRs, measure secreted protein in the culture medium (gel or western blot), and measure the water-droplet contact angle on dried mycelium samples held at different humidities.

2 · Folding and secretion

HFBII has 8 cysteines that must form the right disulfide bonds, and I do not know if the native leader will secrete it well in this host. To find out: check the culture medium for the protein and test whether it is active, for example whether it changes how a water droplet spreads on a treated surface. If secretion is poor, try an alternative leader.

3 · Host mismatch

I designed for T. reesei, but the mycelium in the pieces is probably a different species whose promoters and leader may behave differently, or that may be hard to transform. To find out: check the literature for a transformation protocol in the species I would use, and test the cassette there with a simple reporter before committing to the design.

ReferencesAll references on this page

  1. UniProt entry P79073, Hydrophobin-2, Trichoderma reesei. uniprot.org/uniprotkb/P79073
  2. Nakari-Setälä, T. et al. (1997). Differential expression of the vegetative and spore-bound hydrophobins of Trichoderma reesei: cloning and characterization of the hfb2 gene. Eur. J. Biochem. 248(2), 415–423.
  3. Hakanpää, J. et al. (2004). Atomic resolution structure of the HFBII hydrophobin, a self-assembling amphiphile. J. Biol. Chem. 279, 534–539.
  4. Kottmeier, K., Ostermann, K. & Rödel, G. Secretion system on the basis of a hydrophobin signal sequence from Trichoderma reesei. Patent WO2011073367A1 (2011). Source of the 15-residue HFBII leader (SEQ ID No. 14).
  5. Gibson, D. G. et al. (2009). Enzymatic assembly of DNA molecules up to several hundred kilobases. Nature Methods 6, 343–345. 10.1038/nmeth.1318
  6. Kosuri, S. et al. (2010). Scalable gene synthesis by selective amplification of DNA pools from high-fidelity microchips. Nature Biotechnology 28, 1295–1299. 10.1038/nbt.1716
  7. Carr, P. A. et al. (2004). Protein-mediated error correction for de novo DNA synthesis. Nucleic Acids Research 32, e162. 10.1093/nar/gnh160
  8. Tian, J. et al. (2004). Accurate multiplex gene synthesis from programmable DNA microchips. Nature 432, 1050–1054. 10.1038/nature03151