Engineered Yeast Yields a Long, Stable Collagen Fragment for Skin and Tissue Repair

Yeast can be programmed to make almost any human protein, until its own digestive enzymes start chewing up the chain you asked it to build. A new peer-reviewed Wiley study pinned that proteolysis bottleneck for type III collagen and worked around it by retuning the secretion signal, lifting yield to 6.22 grams per liter of a continuous 585-amino-acid fragment with the peptide backbone preserved.

Type III collagen is one of the body's structural workhorses: a fibrous protein that gives skin, blood vessels, and internal organs their tensile strength. It also shows up in wound dressings, injectable skin-repair formulas, and scaffolds for tissue engineering. Today, most of that supply is extracted from animal hides and tendons, with the variability and contamination risk that implies. Making it in microbial hosts, yeast or bacteria grown in steel tanks, would be cleaner, cheaper, and easier to scale, and the broader industry has been moving that way for years.

The hitch is the host. Pichia pastoris, a yeast long used as a protein factory, secretes almost anything you give it a gene for, but it also releases proteases, the enzymes that chop up proteins for recycling. Those proteases are fine for the yeast's housekeeping. They are lethal for a long, fragile recombinant chain. Earlier work on type III in Pichia ended in fragments, with low yields and inconsistent identity.

The new paper, published in Biotechnology and Bioengineering and indexed at PubMed (PMID 42604590), attacks the problem at the secretion step. Every secreted protein carries a short address tag, a signal peptide, that tells the cell where to send it. The team treated that tag as a tunable part rather than a fixed one, screening variants to find the version the host's own proteases were least likely to attack on the way out. The high-producing strains they selected now secrete the type III fragment at 6.22 g/L.

Mass spectrometry confirmed the product's identity. LC-MS/MS mapped a continuous 585-amino-acid stretch (N611 to P1195 of the human sequence), long and unbroken, the structural feature the team was after. A two-step multimodal chromatography cleanup brought the purified fragment to more than 90% purity with under 10 endotoxin units per milligram.

Circular dichroism (CD) spectroscopy, a technique that reads the shape of a protein in solution, found no sign of the canonical triple-helix conformation that gives native type III collagen its springiness. The fragment's peptide backbone is intact; FTIR spectroscopy confirms that much. The fragment is not, however, a one-to-one structural stand-in for the natural protein. That is the trade the authors are proposing: a stable, scalable, animal-free fragment, not a full triple-helix substitute.

What it can do, the cell work suggests, is hold up its end of the cosmetic and tissue-repair brief. In assays on human skin fibroblasts (HSF cells), the fragment supported cell viability, adhesion, and migration at levels the authors describe as favorable for cytocompatibility. Those are early in vitro signals, bench evidence that justifies further work, not the kind that justifies a clinical claim.

The result is one methods paper from one lab, working in one host, on one fragment of one collagen type. It does not establish that recombinant type III is now commercial-ready, and it does not show the fragment will behave like native collagen in a real wound or scaffold. What it does show is that the long-fragment, high-yield version of microbial type III is no longer purely hypothetical: a specific secretion-engineering fix produced a 6.22 g/L titer of a continuous 585-amino-acid fragment, and the resulting material is clean enough and stable enough to justify the next round of testing.

The next question is whether the same trick, swapping signal peptides instead of swapping hosts, translates to the longer, full-length type III chains and to the closely related type I collagen that dominates the cosmetic and surgical biomaterials market. The papers to watch are the ones that either scale this fragment into gram-or-better fermentations or push the same approach toward a triple-helix-compatible product.