Written and maintained by CASRAI Editorial Board
Published
On this page
Yeast display, also called yeast surface display (YSD), is a protein engineering method that attaches a protein of interest to the outside of a yeast cell, usually Saccharomyces cerevisiae, so that each cell carries both the protein and the gene that encodes it. Cells can then be sorted by what the displayed protein does, most often by binding a fluorescent target in a flow cytometer, and the winning genes read out from the sorted cells.
Short answer: what yeast display does
The technique solves a screening problem: a library may hold millions of protein variants, and the researcher needs the few that bind a target well. Because every yeast cell displays its own variant on its surface and contains the matching DNA, selecting a cell selects a gene (the genotype-phenotype link). The most widely used format, developed by Boder and Wittrup in 1997, fuses the protein to Aga2p, which docks onto the cell-wall anchor Aga1p (Lopez-Morales et al., 2023). Yeast display sits alongside phage display as a standard route in protein engineering, and the two are compared below.
Definition and scope
A yeast-displayed protein is a recombinant protein that is synthesised inside the yeast cell, passed through the secretory pathway and retained at the cell wall instead of being released into the medium. A review of the field describes the construct as needing two parts: a signal sequence that sends the protein into secretion, and an anchor that holds it on the wall; the anchor is typically a glycosylphosphatidylinositol (GPI)-linked cell-wall protein (Teymennet-Ramirez et al., 2021).
The terms yeast display, yeast surface display and the broader surface display are used for the same family of methods. “Surface display” on its own is wider: it also covers display on bacteria and phage. This page covers the yeast version. Hosts other than S. cerevisiae are used too: the same review notes that S. cerevisiae, Pichia pastoris and Yarrowia lipolytica all have GRAS (generally recognised as safe) status with the US FDA, and lists the strain EBY100 among the S. cerevisiae strains used for display. See the site’s guides to Saccharomyces cerevisiae and Pichia pastoris for the organisms themselves.
Why a eukaryotic host matters
Yeast is a eukaryote with a secretory apparatus, so displayed proteins can undergo oxidative folding (disulfide bond formation) and N-linked glycosylation. Gai and Wittrup (2007) list proteins displayed this way, among them epidermal growth factor, interleukin-2, single-chain antibody fragments (scFv), green fluorescent protein, an integrin I domain, fibronectin domains, the West Nile virus envelope protein, the EGFR ectodomain and MHC class II. The same review notes that evaluating affinity and stability on the cell surface avoids having to express and purify each clone as a soluble protein first.
How the Aga1p-Aga2p display system works
In the Boder and Wittrup system, a-agglutinin is the scaffold. According to Lopez-Morales et al. (2023), the a-agglutinin complex has two subunits, Aga1p and Aga2p, which are cell-adhesion glycoproteins that normally help budding yeast cells aggregate during mating. Aga1p carries a GPI motif that anchors it to glycans in the cell wall. Aga2p is bound to Aga1p through a pair of disulfide bonds. The protein of interest is fused to Aga2p, so it is carried out to the surface on the Aga1p anchor.
Induction and expression control
In the standard system the AGA1 and AGA2 genes are under a galactose-inducible promoter, so display is switched on by growing cells in galactose. The same review reports that galactose induction typically takes 24 to 48 hours, and that differences in sequence, size or structure among Aga2p fusions can produce large variation in how much protein reaches the surface. That variation matters because uncontrolled copy numbers make clones hard to compare and can introduce multivalency (avidity) effects, for example when panning against an immobilised or insoluble antigen.
Several groups have tried to regulate display level. Lopez-Morales et al. describe a helper display plasmid with two display levels, plasmids that give a range of expression profiles, and a method that shaves displayed proteins off the surface with the reducing agent dithiothreitol. They also describe the yeast-titratable display (YTD) platform from their own group, in which a tetracycline-repressor feedback circuit controls AGA1 and AGA2-fusion transcription and anhydrotetracycline dose sets the display level. In that report YTD cells displayed roughly tenfold less total protein than galactose-induced cells under saturating conditions and reached maximum display within 5 hours of induction.
The cell wall the protein is anchored to
The anchor ends up in a structured wall. Teymennet-Ramirez et al. (2021) give the outer layer as 100 to 200 nm thick, made up of about 50% mannoproteins, 30 to 45% beta-1,3 glucans, 5 to 10% beta-1,6 glucans and 1.5 to 6% chitin. The data table below lists these components with the display-construct parts for quick reference.
How selection works in practice
The selection step is what makes the method a screening platform rather than just an expression system. Lopez-Morales et al. describe the standard readout: the displayed library is labelled with a target and screened by fluorescence-activated cell sorting (FACS). Gai and Wittrup (2007) list the advantages of this: sorting is quantitative, labelling for expression level can be built into the experiment so that expression differences between clones are not mistaken for binding differences, and affinity and stability can be assessed without purifying each protein.
Selections need not use a soluble, purified target. Gai and Wittrup note that cell-surface selections make it possible to work with insoluble or uncharacterised targets. Lopez-Morales et al. summarise newer variants, including panning against cell lysates and membrane proteins (one example combined a round of magnetic-activated cell sorting with a round of FACS and reported a tenfold enrichment over the starting library), screening against human serum for diagnostic uses, and incorporation of noncanonical amino acids and protein-small molecule hybrids. They also discuss in vivo gene diversification and continuous evolution in yeast, and the use of next-generation sequencing on selected pools, as in deep mutational scanning of the SARS-CoV-2 receptor-binding domain from the Bloom laboratory (Starr et al., 2020).
Yeast display vs phage display
Both methods link a displayed protein to its gene. They differ in host, folding environment and screening readout. Gai and Wittrup (2007) compared them head to head on the same immune libraries, and the table below restates what that review reports.
| Point of comparison | Yeast display | Phage display (as reported in the same review) |
|---|---|---|
| Host and folding | Eukaryotic secretory pathway: oxidative folding and N-linked glycosylation | In the same comparison, only 5 of 12 novel clones from yeast display were functional when displayed on phage |
| Repertoire sampling from immune libraries | Sampled the repertoire “considerably more fully” | Sampled less fully in the same comparison |
| Labour | Described as less labour-intensive | More labour-intensive in the same comparison |
| Downstream expression | 11 of the 12 novel clones expressed solubly in E. coli | Reported only for the yeast-derived clones |
Read the comparison as one review’s account of one set of immune-library experiments, not a general ranking. The site’s phage display guide covers the other method on its own terms.
Where yeast display matters for labs and procurement decisions
For a laboratory, the choice of display platform shapes what equipment and skills a project needs, because the quantitative readout is a flow cytometer or cell sorter and the host is a yeast culture rather than a bacterial one. The uses reported in the sources show where the method is applied.
- Library screening and affinity maturation. Teymennet-Ramirez et al. list library screening first among uses, along with vaccines, biosensors, biocatalysis, bioremediation and ethanol production.
- Antibody and scaffold engineering. The displayed formats in the sources include scFv antibody fragments, fibronectin domains and affibodies.
- Enzyme and cell-surface catalysis. The same review reports that displayed proteins may show improved stability to temperature, pH, solvents and proteases.
- Diagnostics. Lopez-Morales et al. cover serological assays built on yeast-displayed libraries panned against serum.
The comparison with lower-throughput methods is also summarised on the site’s protein engineering guide, which cites yeast display at 107 to 108 variants per day against 104 to 105 for lower-throughput approaches.
Limits of yeast display
- Some proteins fold poorly in the yeast secretory pathway. Gai and Wittrup caution that it is likely inefficient for some cytoplasmic or nuclear proteins with multiple reduced cysteines, such as zinc fingers.
- Expression variation between clones. Lopez-Morales et al. note that Aga2p fusions differ in display level, so clone comparisons need the expression labelling or a titratable system described above.
- Avidity effects. Uncontrolled copy numbers of the displayed protein can introduce multivalency effects, particularly when panning against immobilised or insoluble antigens (Lopez-Morales et al.).
- Epitope mapping needs an independent check. Gai and Wittrup point out that an independent means is needed to verify that the antigen is folded correctly.
- Clone transfer is not automatic. In the yeast-versus-phage comparison, only 5 of the 12 novel yeast-derived clones were functional when displayed on phage, which shows that behaviour can depend on the display platform.
Common misunderstandings
Is yeast display the same as expressing a protein in yeast?
No. Expression in yeast normally means the protein is secreted into the medium or kept inside the cell and harvested. In yeast display the protein is tethered to the wall through an anchor such as the Aga1p-Aga2p pair, so the cell, with its gene, can be sorted.
Is “surface display” always yeast?
No. Surface display is the general idea of presenting a protein on a cell or particle surface; yeast, bacteria and phage are different hosts, and the Aga2p fusion described here is specific to yeast.
Will a clone selected by yeast display always work as a soluble protein?
Not automatically. In the Gai and Wittrup comparison, 11 of 12 novel yeast-selected clones expressed solubly in E. coli, so most did, but that is one dataset and a clone should be tested in the format in which it will be used.
Is display level just a detail?
It determines how reliably clones can be compared, which is why titratable systems such as YTD were developed.
Sources
- Gai SA, Wittrup KD. Yeast surface display for protein engineering and characterization. Curr Opin Struct Biol 2007;17(4):467-473, checked 2026-11-17
- Teymennet-Ramirez KV et al. Yeast Surface Display System: Strategies for Improvement and Biotechnological Applications. Front Bioeng Biotechnol 2021;9:794742, checked 2026-11-17
- Lopez-Morales J et al. Protein Engineering and High-Throughput Screening by Yeast Surface Display: Survey of Current Methods. Small Sci 2023;3(12):2300095, checked 2026-11-17
- Starr TN et al. Deep Mutational Scanning of SARS-CoV-2 Receptor Binding Domain Reveals Constraints on Folding and ACE2 Binding. Cell 2020, checked 2026-11-17
Data: Yeast cell wall components and display-construct parts in yeast surface display
Compiled from three open-access reviews (Teymennet-Ramirez 2021, Lopez-Morales 2023, Gai and Wittrup 2007): the cell-wall components that anchor a displayed protein, with the reported share of each wall component, and the roles of the construct parts used in the Aga1p-Aga2p system.
| Component | Category | Reported detail |
|---|---|---|
| Mannoproteins | Cell-wall component | About 50% of the wall (outer layer 100-200 nm thick) |
| Beta-1,3 glucans | Cell-wall component | 30-45% of the wall |
| Beta-1,6 glucans | Cell-wall component | 5-10% of the wall |
| Chitin | Cell-wall component | 1.5-6% of the wall |
| Aga1p | Construct part | Cell-adhesion glycoprotein with a GPI motif that anchors it to glycans in the cell wall |
| Aga2p | Construct part | Bound to Aga1p by two disulfide bonds; the protein of interest is fused to it |
| Galactose promoter | Construct part | Controls AGA1 and AGA2 so display is induced by galactose |
| Signal sequence | Construct part | Directs the fusion into the secretory pathway |
Sources
- Teymennet-Ramirez et al. Yeast Surface Display System (accessed 2026-11-17)
- Lopez-Morales et al. Protein Engineering and High-Throughput Screening by Yeast Surface Display (accessed 2026-11-17)
Compiled by CASRAI from the sources above. Free to reuse under CC BY 4.0 with attribution.







