More than just letters – the impact of protein sequence variation on expressibility and host cell responses (reference code 229)

Background.

Yeasts such as Saccharomyces cerevisiae or Komagataella phaffii (syn. Pichia pastoris) are efficient hosts for recombinant protein production, both for biotherapeutics and industrial enzymes (Barone et al. 2023; Rettenbacher et al. 2022). Many of them are produced in a secretory manner. Production levels of such proteins range from µg to several g per liter, and it is very hard to predict a priori whether a protein will be well or hardly produced. Even for very closely related proteins (e.g. such obtained by protein engineering or isoforms from different species), production levels can differ significantly.
It is hypothesized that in eukaryotes, secretion yields are related to stability, with less stable proteins being more prone to misfolding and subsequent degradation rather than successful secretion. Indeed, it has already been experimentally verified in the late 1990s that for some proteins, such as bovine trypsin inhibitor (BPTI) expressed in S. cerevisiae, secretion yields were inversely correlated with thermodynamic stability (Kowalski et al. 1998). Similarly, a negative relationship between the stability and secretion levels of 8 disease-related mutants of human lysozyme (HuL) in K. phaffii was established (Kumita et al. 2006). HuL variants with lower stability led to lower protein yields, and showed induction of several aversive cellular pathways such as endoplasmic reticulum (ER) stress response, unfolded protein response (UPR), ER-associated degradation (ERAD) and ER-phagy (Whyteside et al. 2011). Similar findings correlating high production yields and thermodynamic stability were also reported for other host platforms (e.g. Yokota et al. 2017, Kumita et al. 2012). However, all these studies were performed on a limited number of variants that were already pre-selected based on their different stability and aggregation potential (e.g. 8 HuL variants occurring during diseases, 6 BPTI variants with destabilized disulfide bonds), which is far away from a real-life scenario in protein engineering, where the goal is usually to alter binding affinity or substrate specificity of the protein of interest (POI). Moreover, many of the previously investigated instable proteins were prone to aggregation, which is a known factor to trigger ER-quality control mechanisms such as ERAD (Sun & Brodsky, 2018).
While we agree that there is a correlation between stability and protein yields, we have evidence that thermodynamic stability is not the only determinant affecting the expressibility of proteins. For example, zymogens from different species have >89% identity, but vary in their expression levels over 100-fold (BOKU-IMMB, unpublished data). The introduction of stability enhancing mutations led to higher activity, but not higher amounts of secreted recombinant protein.
This also becomes obvious when looking at the published variants, where usually only the most and least stable variants cause the strong correlation. Foldability and the abundance of certain amino acids are among the discussed factors (e.g. Listov et al. 2022; van den Berg. 2012).

Many of these studies were limited by the number of variants that were experimentally screened. Yeast surface display (YSD) is a valuable tool for high-throughput (HTP) screening of protein libraries established in both S. cerevisiae and K. phaffii (van Deventer et al. 2015; Kang et al. 2022; Stadlmayr et al. 2010), and the displayed protein levels usually correlate well with the amounts of secreted soluble protein (Shusta et al. 1999; Laurent et al. 2021). Alternatively, single cell encapsulation techniques in combination with a suitable protein detection assay or plate-based assay that yield easily visible changes upon the production of the target protein (such as colorimetric changes or halos) can be used for HTP screening.

Research questions.

• In which ways do variations in protein sequence (e.g. such as introduced through protein engineering) affect expressibility of such proteins in yeast?
• Which cellular pathways are affected if very closely related proteins (engineered protein variants vs. wild-type proteins) are expressed in yeast?
• Is there a correlation between expressibility and physico-chemical properties that goes beyond thermodynamic stability?

Aims and expected outcomes.

(i) Quantitative determination of the impact of protein sequence variations on expressibility in yeast for three different protein formats
(ii) Correlation of cellular responses of yeast cells that secrete protein variant libraries with their expressibility
(iii) Correlation of expressibility to physico-chemical protein properties

Project plan and methods.

Mutants of model enzymes (selected from B. subtilis serine protease, Crassicarpon hotsonii cellbiose dehydrogenase) and binders (non-antibody binder scaffold, single chain/single domain antibody fragment) will be analysed regarding their secretion yields, stress responses and biophysical properties. Initially, available data and existing mutants will be investigated for their thermodynamic stability, stress response and secretion yields. For one model protein, a binder scaffold that was selected via YSD in S. cerevisiae, it will be validated that expressibility is host species-independent by expressing selected protein sequence variants in K. phaffii as well as in mammalian cells (human T-cells).
To enlarge the design space, additional mutations will be introduced into the model poteins through random mutagenesis or targeted approaches, the latter aiming to generate mutants with similar thermodynamic stability but altered properties. To obtain a high-throughput correlation of cell stress to protein sequence variations, different sequence variants will be expressed as YSD-protein sequence libraries in yeast cells (preferably K. phaffii) that contain fluorescent sensors for folding- and secretion-related cellular responses (e.g. UPR biosensor) and analysed by flow cytometry.
Selected variants will be analyzed in order to differentiate between mutations affecting folding/secretion and activity/binding properties, as well as the more in-depth cellular response to their expression.

- Expression of model antibodies/binders and enzymes in YSD format
- Construction of libraries (gene sequences subjected to random mutagenesis)
- Correlation of cellular stress response (via fluorescent genetically encoded biosensors for ER stress, UPR and ROS) and YSD yields (as proxi for secreted protein titers) via flow cytometry and fluorescence activated cell sorting
- Isolation of mutants with increased OR decreased secretion; identification of encoded protein sequence; verification as soluble secreted protein
- High throughput assays will be used to evaluate secretion and/activity e.g. B. subtilis serine protease producing cells will be screened for clearing halos on casein plates; cellobiose dehydrogenase gene will be screened for oxidoreductase activity in an ultra-high-throughput-assay; binders will be detected immunofluorescent staining.
- Variants with significantly increased as well as decreased clearing signals will be selected and quantitatively analyzed for the amount of secreted soluble recombinant protein and the specific activity of the produced protein.
- Analysis of variants for functionality (catalytic activity, binding properties) and physico-chemical properties (thermodynamic stability)
- Analysis of variants for cellular response of yeast expression host: Correlate intracellular and secreted protein amounts for protein sequence variants; optionally analyse differential gene expression levels of selected strains using transcriptomics; optionally determine intracellular localization of protein sequence variants as well as potential ER-phagy by (immunofluorescence) microscopy

REFERENCES
Barone GD, Emmerstorfer-Augustin A, Biundo A, Pisano I, Coccetti P, Mapelli V, Camattari A. 2023. Industrial Production of Proteins with Pichia pastoris-Komagataella phaffii. Biomolecules. 13(3):441. doi: 10.3390/biom13030441.
Kowalski JM, Parekh RN, Mao J, Wittrup KD. 1998. Protein folding stability can determine the efficiency of escape from endoplasmic reticulum quality control. J Biol Chem. 273(31):19453-8. doi: 10.1074/jbc.273.31.19453.
Kowalski JM, Parekh RN, Wittrup KD. 1998. Secretion efficiency in Saccharomyces cerevisiae of bovine pancreatic trypsin inhibitor mutants lacking disulfide bonds is correlated with thermodynamic stability. Biochemistry. 37(5):1264-73. doi: 10.1021/bi9722397.
Kumita JR, Johnson RJ, Alcocer MJ, Dumoulin M, Holmqvist F, McCammon MG, Robinson CV, Archer DB, Dobson CM. 2006. Impact of the native-state stability of human lysozyme variants on protein secretion by Pichia pastoris. FEBS J. 273(4):711-20. doi: 10.1111/j.1742-4658.2005.05099.x.
Kumita JR, Helmfors L, Williams J, Luheshi LM, Menzer L, Dumoulin M, Lomas DA, Crowther DC, Dobson CM, Brorsson AC. 2012. Disease-related amyloidogenic variants of human lysozyme trigger the unfolded protein response and disturb eye development in Drosophila melanogaster. FASEB J. 26(1):192-202. doi: 10.1096/fj.11-185983.
Listov D, Lipsh-Sokolik R, Rosset S, Yang C, Correia BE, Fleishman SJ. 2022. Assessing and enhancing foldability in designed proteins. Protein Sci. 31(9):e4400. doi: 10.1002/pro.4400.
Kang BH, Lax BM, Wittrup KD. 2022. Yeast Surface Display for Protein Engineering: Library Generation, Screening, and Affinity Maturation. Methods Mol Biol. 2491:C1. doi: 10.1007/978-1-0716-2285-8_27.
Rettenbacher L, Arauzo-Aguilera K, Buscajoni L, Castillo-Corujo A, Ferrero-Bordera B, Kostopoulou A, Moran-Torres R, Núñez-Nepomuceno D, Öktem A, Palma A, Pisent B, Puricelli M, Schilling T, A Tungekar AA, Walgraeve J, Humphreys D, von der Haar T, Gasser B, Mattanovich D, Ruddock L, Jan Maarten van Dijl JM. 2022.Microbial protein cell factories fight back? Trends Biotechnol. 40(5):576-590. doi: 10.1016/j.tibtech.2021.10.003.
Shusta EV, Kieke MC, Parke E, Kranz DM, Wittrup KD. 1999. Yeast polypeptide fusion surface display levels predict thermal stability and soluble secretion efficiency. J Mol Biol. 292(5):949-56. doi: 10.1006/jmbi.1999.3130.
van den Berg BA, Reinders MJ, Hulsman M, Wu L, Pel HJ, Roubos JA, de Ridder D. 2012. Exploring sequence characteristics related to high-level production of secreted proteins in Aspergillus niger. PLoS One. 7(10):e45869. doi: 10.1371/journal.pone.0045869.
Yokota JI, Shiro D, Tanaka M, Onozaki Y, Mizutani O, Kakizono D, Ichinose S, Shintani T, Gomi K, Shintani T. 2017. Cellular responses to the expression of unstable secretory proteins in the filamentous fungus Aspergillus oryzae. Appl Microbiol Biotechnol. 101(6):2437-2446. doi: 10.1007/s00253-016-8086-3.