Early on in my career while toying with mRNA display and phage display, I dreamed that I could use FACS and flow cytometry to sort populations of peptide binders. I then moved over to antibody discovery and drove a yeast display platform to discover scFv’s and VHH’s from immune repertoires. Here, I realized how “easy” discovery could be: A common workflow for yeast surface display ebbs and flows with yeast growth, allowing about one round of selection per week, equipped only with a proper sorter, flow cytometer, swing-bucket centrifuge, and a shaking programmable incubator (which can enable two rounds per week). This can all be done by a single researcher, and other than growth media, induction media, and a few cell passages, the cells do (almost!) everything that an entire phage display or mRNA display group would do between rounds.
And speaking of yeast doing the heavy lifting — don’t panic, but the Boulder Peptide Symposium is only two weeks away! If your abstract still looks like a yeast culture on day one, now’s the time to induce some expression.
Back to science, the wet-lab schedule of yeast surface display is similar to working in a bakery; yeast grow, they’re induced to express the surface-displayed library, they’re mixed with target antigen, stained with fluorescently-labeled antibodies to ensure expression of the library on the cell surface and another antibody stain that binds to the antigen of interest, which, if bound by the yeast results in a quantifiable double-stained yeast cell. This enables the physical sorting of binders from non-binders, high affinity gating within the population of binders, and collection of single cells in single wells to quickly characterize their affinity even before knowing the sequence.
The physical advantages of yeast surface display far outweigh the avidity-shaming nay-sayers commentary (I also initially thought that it would be silly to attach 1000’s of copies of a construct to a single yeast cell and expect high-affinity binders to win). Beyond the avidity concern, yeast transformation rates are orders of magnitude below other platforms’ diversity, further agonizing my initial bias against a yeast platform because I was accustomed to high diversity and full coverage libraries, which makes identification of the best hits from selections much easier and provides significant SAR within the NGS dataset. The enabling technology demonstrated in this paper satisfies many of these concerns.
OK, to the paper at hand:
Sara Linciano and Ylenia Mazzocato merge the essence of library-based peptide discovery and the advantages of yeast surface display to demonstrate the discovery of peptide hits targeting an array of different proteins. This group takes the best part of peptide display platforms (e.g., using constrained amino acid architectures that enable cyclization, controlled lengths that attempt library diversity coverage, deep NGS analysis of resulting selection output to identify families of binders) and crosses it with the ability to express a diverse library of DNA-encoded constructs on a yeast cell surface, which together enable the selection of peptide binders via fluorescence-activated cell sorting (FACS) all while enjoying the simplicity of propagating yeast culture (again, compared to the laborious effort between rounds of mRNA display to regenerate the library).
Specifically, this group expresses a library of peptides constructed with specific lengths and cyclization architectures (utilizing cysteines) to define the peptide architecture of potential resulting hits. The range of peptide lengths and cycle sizes (with NNK codon utilization in the random regions) results in libraries with almost full coverage in yeast, whose library diversities are limited by transformation efficiencies that hover in the 10^8 to 10^9 range. Common yeast surface display platforms utilize AGA1 and AGA2 binding interactions to present antibody fragments on the surface, however this group has elegantly utilized a different surface protein, GPI, which does not contain free cysteines, to present the disulfide-cyclized peptide libraries as N-terminal peptide-GPI fusion proteins.
Additionally, yeast display enables the huge advantage of being able to perform affinity determinations via flow cytometry while still working within the yeast platform, which avoids SPPS and the associated expenses and enables the screening of many more clones without significant cost (one only needs more yeast growth media and more antibody stains). This paper demonstrates the capability to rank-order different peptide hits against various targets in the nM range of steady-state affinities, and then further hammers it home with step-wise SPR to confirm the residence times and structure determination to further demonstrate target engagement. A proper tour; nice work.
And while they haven’t done this yet, the future ability to perform a TITESEQ experiment, whereby a converged selection’s library is panned against a titration of target concentrations and are all sequenced, revealing the binding rank-order from the abundance of clones in the NGS data at a given concentration. This technique is very rewarding in an antibody discovery and engineering campaigns when a binding population is down to ~1000 members (depending on NGS depth) and is hopefully applicable to the discovery of peptides on yeast which this group has demonstrated here. To push the envelope further, I wonder if they can they perform post-translational modifications or cyclization chemistries that enable further functionalization of their libraries…?
Open Access article: https://doi.org/10.1038/s41467-025-60907-x
Linciano, S., Mazzocato, Y., Romanyuk, Z. et al. Screening macrocyclic peptide libraries by yeast display allows control of selection process and affinity ranking. Nat Commun 16, 5367 (2025).
Zeke Nims, Ph.D.
https://www.linkedin.com/in/zeke-nims/