Application of modified yeast surface display technologies for non-Antibody protein engineering

Accepted Manuscript Title: Application of Modified Yeast Surface Display Technologies for Non-Antibody Protein Engineering Author: Meng Mei Yu Zhou Wenfang Peng Chan Yu Lixin Ma Guimin Zhang Li Yi PII: DOI: Reference:

S0944-5013(16)30398-6 http://dx.doi.org/doi:10.1016/j.micres.2016.12.002 MICRES 25972

To appear in: Received date: Revised date: Accepted date:

29-6-2016 21-10-2016 9-12-2016

Please cite this article as: Mei Meng, Zhou Yu, Peng Wenfang, Yu Chan, Ma Lixin, Zhang Guimin, Yi Li.Application of Modified Yeast Surface Display Technologies for Non-Antibody Protein Engineering.Microbiological Research http://dx.doi.org/10.1016/j.micres.2016.12.002 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Application of Modified Yeast Surface Display Technologies for NonAntibody Protein Engineering

Meng Mei1, Yu Zhou1, Wenfang Peng1, Chan Yu1, Lixin Ma1, Guimin Zhang1*, Li Yi1* 1, Hubei Collaborative Innovation Center for Green Transformation of Bio-Resources, Hubei Key Laboratory of Industrial Biotechnology, College of Life Sciences, Hubei University, Wuhan, 430062, People's Republic of China *, Correspondent Author

Correspondent Authors Li Yi, Ph.D., Department of Bioengineering, College of Life Science, Hubei University, Wuhan, Hubei, China 430062; Tel.: +86-27-88661237; Fax: +86-27-88663882; Email: [email protected] Guimin Zhang, Ph.D., Department of Bioengineering, College of Life Sciences, Hubei University, Wuhan, Hubei, China 430062; Tel.: +86-27-88661746; Fax: +86-27-88663882; E-mail: [email protected]

1

Abstract Yeast surface display (YSD) system has been widely used in protein engineering since it was established 20 years ago. Combined with fluorescence-activated cell sorting (FACS) technology and directed evolution, YSD has been proven of its extraordinary effectiveness for molecular engineering of various target proteins, especially for antibodies. Recently, a few remarkable efforts were exploited to modify the original Aga1-Aga2 YSD for the non-antibody protein engineering with successful outcomes, expanding its application on oxidase, Class II major histocompatibility complex (MHC II), protease, sortase, lipoic acid ligase etc. Here, the methodologies of these optimized Aga1-Aga2 YSD technologies were introduced, and the recent progress of non-antibody protein engineering using these methods was summarized.

Keywords: Yeast surface display, FACS, Directed evolution, Non-antibody protein engineering

Introduction The background of microbial surface display The cell surface is a functional interface between the inside and outside of the cell. In biotechnology, the cell surface can be exploited by making use of known mechanisms of transporting heterologous proteins to the cell surface. However, lacking of effective platforms to present protein of interest on the cell surface largely hindered the development of surface display technology. In 1990, Scott and Smith discovered that short peptides could be displayed on the virion surface by fusing to the anchor protein of the filamentous phages without affecting their infection ability (Scott and Smith, 1990). Since then, various surface display platforms have been established and rapidly developed (Tanaka et al., 2012; Schuurmann et al., 2014; Domingo-Calap et al., 2016). Surface display system has its major advantage that the target proteins can be displayed on the cell surface to make the manipulation of enzymatic reactions more feasible and easier. Three major microbial cell-surface display systems have been developed so far, including phage display, bacteria surface display, and yeast surface display (YSD) (Lipke and Kurjan, 1992; Little et al., 1993; Griffiths 2

et al., 1994). The principle of phage display is displaying heterologous peptides or proteins through fusing them with coat protein of filamentous phages, which is widely used to isolate ligands, antigens, and antibodies (Hockney, 1994; Delhalle et al., 2012). Because of the phage’s small bulk, the size and variety of the target peptides or proteins are highly restricted. In the bacteria surface display system, e.g. the E.coli OmpA system (Schreuder et al., 1993; Georgiou et al., 1997), heterologous proteins are normally inserted into the loop region of the cellular outer membrane protein, forming a protein complex, which is then co-displayed on the bacterial cell surface. However, inserting the foreign protein into the cellular outer-membrane protein frequently disrupts its functional structure, causing low surface display efficiency. These, together with the fact that many eukaryotic proteins are usually not well folded or misfolded in bacteria and phage, have put a request of developing a surface display system in eukaryotic cells. Under this circumstance, YSD, which uses the protein anchored on the yeast cell wall as a surface carrier, was developed (Kondo and Ueda, 2004). Compared to phage and bacteria display systems, YSD has two extraordinary advantages. First, yeast is a unicellular eukaryote, favoring the expression and folding of the eukaryotic proteins with its post-translational system. This is important, as human antibody engineering, the most profitable protein engineering research field, has encountered massive difficulties in phage and bacteria display systems because of misfolding issues. Second, the heterologous protein in YSD can be alternatively fused at either the N- or C-terminal of the surface anchor protein without disrupting its core structure, thus keeping the structure of the surface anchor protein largely unchanged without undermining its surface displaying efficiency. In addition, other advantages also include high display efficiency and full exposure of target proteins out of the cells in YSD. Compared to phage and bacteria surface display systems, the mostly mentioned disadvantage of YSD is that yeast cells exhibit slower growth rate and lower foreign plasmid transformation efficiency. However, recent developments of yeast technology have already enhanced the yeast transformation efficiency to 108, making it feasible of generating variant libraries with sufficient information in laboratory scale (Kondo and Ueda, 2004; Benatuil et al., 2010).

Brief theory introduction to YSD In the YSD system, heterologous proteins expressed in the yeast are fused to surface anchor protein, forming a protein complex. This protein complex was then displayed on the yeast cell surface, 3

leading to feasible reproduction of the in vitro reaction system for the displayed biocatalysts, and easy detection of the products from catalyst. Nowadays, a matured high-throughput screening approach for protein engineering has been established when YSD is combined with fluorescence-activated cell sorting (FACS) technology and directed evolution (Figure 1). As an easy manipulation platform, the YSD system appears unique advantages for protein engineering comparing to the bacteria and phage display systems, especially for the eukaryotic functional proteins requiring post-translational modifications. In most YSD systems, glycosylphosphatidylinositol (GPI)-anchored proteins were typically used for displaying heterologous proteins, including agglutinin (α-agglutinin and aagglutinin), flocculin Flo1p, Cwp1p, Cwp2p, Sed1p, Tip1p, YCR89w, and Tir1 (Lipke et al., 1989; Lipke and Kurjan, 1992; Kuroda and Ueda, 2014). Through the yeast endoplasmic reticulum (ER)Golgi secretory pathway, the GPI-anchored proteins are transported to the yeast cell surface, forming a β-1, 6-glucan bridge with the mannoprotein layer of the cell wall (Scott and Smith, 1990; Lu et al., 1995; Ueda and Tanaka, 2000; Dudgeon et al., 2012). GPI-anchored proteins have been demonstrated to mediate the display of a range of heterologous proteins upon protein fusion. In 1993, α-galactosidase from Cyamopsis tetragonoloba became the first heterologous protein that was displayed on the yeast cell surface after being fused to the C-terminal of the protein (Schreuder et al., 1993). Since then, many other proteins, including Class II major histocompatibility complex (MHC-II) (Boder et al., 2005), epidermal growth factor receptor (EGFR) fragments (Chao et al., 2004), lipase from Rhizopus oryzae (ROL) (Tanino et al., 2006), single-chain variable fragment (scFv) (Feldhaus et al., 2003), and other enzymes were successfully displayed on the yeast cell surface for further engineering. Moreover, the combination with other newly developed technologies has greatly strengthen the capability of YSD, expanding its application in engineering a number of functional proteins (Wen et al., 2011; Bagriantsev et al., 2014; Maute et al., 2015; Frago et al., 2016), as well as catalytic enzymes (Matsuura et al., 2013; Jin et al., 2014), antibodies (Doerner et al., 2014; Rhiel et al., 2014; Van Deventer and Wittrup, 2014), and combinatorial protein libraries (Ueda, 2009). Recently, the YSD system was also developed to display redox active enzymes for microbial fuel cell applications (Szczupak et al., 2012), actinidain for fast characterization of food allergens (Popovic et al., 2015), xylose reductase along with other enzymes for xylitol production (Chen et al., 2016), and α-amylase and glucoamylase simultaneously for maximum ethanol production (Inokuma et al., 2015). Among the GPI-anchored proteins for surface display, Flo-anchoring protein displays its C4

terminal end associated protein on the yeast cell surface through non-covalent interactions between its flocculation functional domain and the cell-wall α-mannan (Van Mulders et al., 2009). Besides the Flo protein, other mannose proteins, such as Cwp1p, Cwp2p, Sed1p, Tip1p, Tir1p and YCR89W (Table 1), are also developed as anchor as well as carrier proteins for immobilizing the target protein on the cell wall of S. cerevisiae or Pichia pastoris (Tokuhiro et al., 2008; Wasilenko et al., 2010; Liu et al., 2014b), among which Cwp2p was identified to be the only GPI-anchored cell wall mannoprotein in the interior of the cells (Van der Vaart et al., 1997). Different from these cell surface proteins that function as both the surface anchor and heterologous protein carrier, the Aga1-Aga2 YSD system is composed of separated surface anchor and protein carrier. Aga1 and Aga2 proteins both belong to the a-agglutinin proteins. However, unlike the α-agglutinin protein that only consists of a single Agα1 unit, functioning as both anchor and carrier, the Aga1 and Aga2 function as surface anchor and heterologous protein carrier, respectively. In the Aga1-Aga2 YSD system, the heterologous protein of interest is expressed as a fusion to the Aga2 mating agglutinin protein, which is linked to cell wall covalently associated Aga1 through two disulfide bonds (Boder and Wittrup, 1997). It is worth pointing out that the heterologous proteins can be fused at either end of Aga2, providing alternative options to maintain the original biological properties of target proteins. Additionally, an engineered Saccharomyces cerevisiae strain EBY100 is used in the Aga1-Aga2 YSD system, in which a heterologous Aga1 gene under the control of the galactose-inducible GAL promoter is inserted into the yeast chromosome. Since the Aga2 fusion complex gene in the exogenous plasmid is also under the control of the galactose-inducible GAL promoter, the expression of Aga1 and Aga2 fusion complex can be strongly induced simultaneously for the highest surface display efficiency, with an average of more than 3x104 heterologous protein molecules being presented on one yeast cell (Boder and Wittrup, 1997).

Non-antibody protein engineering using Aga1-Aga2 YSD system Saccharomyces cerevisiae has been used for decades as a useful eukaryotic host for protein production. Because of the significant commercial profits of the antibody industry, YSD was mostly developed for antibody engineering. Pioneered by Wittrup and other researchers, various antibodies were displayed on the yeast surface using Aga1-Aga2 YSD system and engineered for high specificity against their antigens, such as fluorogens, cancer antigens, plasma membrane proteins, and viral envelop proteins (Boder and Wittrup, 1997; Holt et al., 2003; Wesolowski et al., 2009). Recently, the 5

Aga1-Aga2 YSD technology was rapidly expanded to non-antibody protein engineering (Cochran et al., 2006; Sharma et al., 2013; Weiskopf et al., 2013; Zaretsky et al., 2013; Zhang et al., 2013a; Zhang et al., 2013b; Kariolis et al., 2014; Heimer et al., 2015; Maute et al., 2015; Harris et al., 2016) (Table 2). Moreover, due to the different properties of various enzymes, the original Aga1-Aga2 YSD technology has also been modified for some specific enzyme engineering, e.g. MHC- II (Jiang and Boder, 2010), LplA (Puthenveetil et al., 2009), glucose oxidase (Wang et al., 2013), HRP (Lipovsek et al., 2007), TEV-P (Yi et al., 2013), SrtA (Chen et al., 2011) (Table 3). In this review, these recently developed new methods based on the Aga1-Aga2 YSD system combined with directed evolution and FACS screening technology in the non-antibody protein engineering will be discussed.

Engineering protein-peptide binding affinity Aga1-Aga2 YSD system has been successfully used to engineer many non-antibody proteins for higher affinity against its interacting proteins or peptides, including T cell receptors (TCRs) (Holler et al., 2000; Weber et al., 2005; Buonpane et al., 2007), interleukin receptor (Zaretsky et al., 2013), Tolllike receptor (Mattis et al., 2015), integrin I domain (Jin et al., 2006), epidermal growth factor (EGF) (Cochran et al., 2006), Kelch-repeat (KR) domain of the Keap1 protein (Zhang et al., 2013a) etc. One recent important progress was that Jiang and Boder established a yeast co-display system based on the Aga1-Aga2 YSD system, which was successfully utilized to engineer a heterodimer, MHC-II, for altered affinity against different target peptides (Jiang and Boder, 2010). MHC-II engineering MHC-II belongs to the heterodimeric trans-membrane protein family, which captures antigenic peptides processed in professional antigen-presenting cells (APCs) and then presents the captured antigens on the surface of the APCs to trigger T cell responses (Sette and Fikes, 2003; Watts, 2004). Therefore, decoding the peptide-binding specificity of MHC-II and exploring the molecular mechanisms of how MHC-II binds to the relative peptides are helpful for vaccine design as well as the understanding of autoimmunity, infection progression, and transplantation rejection involved in a number of diseases (Casares et al., 2002; Jiang and Boder, 2010). In the yeast co-display system that was established by Jiang and Boder, HLA-DR1 (human leukocyte antigen-D Related 1) (Figure 2A), a model MHC-II, and its bound peptide FLU were codisplayed on the yeast cell surface (Jiang and Boder, 2010). HLA-DR1α, HLA-DR1β and FLU were 6

first co-expressed and transported into the yeast ER, in which effective HLA-DR1α and HLA-DR1β could form a functional HLA-DR1 heterodimer, recognizing FLU to form an HLA-DR1-FLU complex. Since the FLU was fused to the Aga2, the HLA-DR1-FLU complex could be transported to the yeast cell surface through Aga1-Aga2 YSD system. Among all the cell surface displayed FLU peptides, the higher extent of HLA-DR1 binding indicates the stronger affinity between HLA-DR1 variants and FLU peptide. Through the V5 and HA epitope tags flanked at the N- and C- terminal sides of the FLU peptide, the total surface displaying level of FLU was quantitated by the anti-V5 antibody and anti-HA antibody, respectively, followed by fluorophore labeling and FACS analysis. At the same time, anti-DR antibody was used to recognize the HLA-DR1-FLU complex on the cell surface, providing quantitative information of how much HLA-DR1-FLU complex were formed. The more HLA-DR1-FLU complex displayed on the cell surface indicates the higher affinity of HLA-DR1 variant against FLU peptide. Using this yeast co-display system, Jiang and Boder firstly analyzed the FLU P1 residue specificity of HLA-DR1 (Jiang and Boder, 2010). Compared to the wild-type FLU with Tyr at P1, the sorting results indicated that wild-type HLA-DR1 exhibited similar affinity to another two aromatic amino acids, Phe and Trp. The results also indicated that HLA-DR1 presented mild affinity to Met, Leu, Ile and Val. Other than that, HLA-DR1 barely recognized other amino acids. To expand the potential application of HLA-DR1 to efficiently recognize FLU variants with Val, Ala, or Glu at P1 position, Jiang and Boder then carried out directed evolution of HLA-DR1 using the yeast co-display system (Jiang and Boder, 2010). Three interesting variants, S4E1.3, S4V1.5, and S4A1.9, were finally obtained, which presented altered affinities against different FLU variants. Among these three HLADR1 variants, S4A1.9 exhibited very low affinity to wild-type FLU with Tyr at P1 position, but extraordinary increased binding affinity to FLU variant with Ala at P1 position. In addition, S4E1.3 variant presented promiscuous affinity against all 20 amino acids at P1 site in the FLU. This yeast codisplay system developed by Jiang and Boder efficiently broadened the Aga1-Aga2 YSD system on displaying heterodimeric/homodimeric proteins.

Engineering enzyme-substrate specificity Genome research has revealed that abundant gene resources existed in different creatures, among which the number of catalytic enzymes is in a small portion. In fact, our understanding of the different substrates that can be recognized by a specific catalytic enzyme is very limited. Therefore, identifying 7

un-reported substrates that can be recognized by a known catalytic enzyme is a critical topic for its practical application. As a powerful protein engineering approach, the Aga1-Aga2 YSD system was also used to explore the different substrates that can be recognized by a specific enzyme. For instance, the substrate diversities of lipoate-protein ligase A (LplA) was explored using the YSD due to its broad application. LplA peptide substrates engineering LplA catalyzes a two-step reaction, in which a lipoyl-AMP intermediate is firstly generated through ATP-dependent activation of lipoic acid. Then the lipoyl moiety was subsequently transferred from the lipoyl-AMP to a specific lysine residue of an apo-protein. In addition, LplA can also specifically ligate small molecules, such as photocrosslinkers and alkyl azides, to small tags of 17-22 amino acid peptides named as LAP1 (LplA Acceptor Peptide 1) (Fernandez-Suarez et al., 2007; Baruah et al., 2008). As a useful bio-catalytic tool for specific protein tagging, E.coli LplA has been biochemically and structurally characterized, and efficiently used for fluorescent protein labeling applications (Fujiwara et al., 2001; Fujiwara et al., 2005; Fujiwara et al., 2010). To expand the utilization of LplA for protein tagging, Ting’s research group carried out extensive research to employ the YSD technology (Figure 2B) to isolate new LplA peptide substrates without βhairpin structures (Puthenveetil et al., 2009). Through library sorting, a new 13-amino acid polypeptide, LAP2 (GFEIDKVWYDLDA), which was recognized by E.coli LplA more efficiently than the enzyme’s previously designed substrate, LAP1 (DEVLVEIETDKAVLEVP), was finally obtained (Puthenveetil et al., 2009). LAP2 polypeptide was conjugated to lipoic acid by LplA with a kcat/KM of 0.99 μM-1min-1, which is 8-fold lower than that of LplA’s natural protein substrate H-protein, but >70fold higher than that of LAP1 polypeptide substrate. In the study, a library containing 107 LAP polypeptide variants was displayed on the yeast surface through fusion to the Aga2. A c-Myc epitope tag was added to the C-terminus of the LAP, allowing the quantification of the LAP variant expression levels through cell surface labeling with an Alexa Fluor 488 conjugated anti-c-Myc antibody. After being surface displayed, the solution added lipoic acid could be ligated to effective LAP variants under the catalysis of LplA, in which the extent of the ligated LAP variants were detected with an anti-lipoic acid antibody conjugated with PE fluorophore. The ligation efficiencies of different LAP variants for lipoic acid were then monitored by the fluorescence intensity ratio of PE/Alexa Fluor 488 using FACS. One interesting strategy utilized in this study is that the enzyme concentration of LplA was gradually 8

decreased from 5 M to 200 nM during the total four rounds of cell sorting, providing extra pressure to differentiate the most suitable substrate for LplA. The evolved LAP2/LplA was later expanded for protein labeling in living mammalian cells. The HEK cells expressing the LAP2-LDL receptor, in which LAP2 polypeptide fused to the low density lipoprotein (LDL) receptor, was specifically and efficiently labeled with 11-Br probe under the catalysis of LplA (Puthenveetil et al., 2009). Additional examples using the LAP2/LplA set also include cytoskeletal proteins labeled with green and red fluorophores (Liu et al., 2012b), and neurexin labeled with HaloTag-QD605 in living cells (Liu et al., 2012a).

Engineering enzyme for increased catalytic activity An enzyme with high activity is always the goal in enzyme engineering, which is also a big obstacle to hinder the enzyme’s practical application. As a powerful method, YSD system has also been used for this purpose. In general strategy, target enzyme library was displayed on the yeast cell surface, followed by isolation of characteristic variants, including Rhizopus oryzae lipase engineering (Shiraga et al., 2005), DhbE adenylation domain engineering (Zhang et al., 2013b), glucose oxidase (GOx) engineering (Ostafe et al., 2014) etc. Among these examples, a smart modification was applied to the Aga1-Aga2 YSD system to facilitate the GOx engineering.

GOx engineering In the glucose oxidase (GOx) engineering, the strategy of ligating fluorescent substrate to the tyrosine residues of the yeast cell surface protein by HRP was used to establish a new high-throughput screening method based on the Aga1-Aga2 YSD system (Ostafe et al., 2014). GOx is an oxidoreductase, catalyzing the oxidation of glucose in the presence of oxygen to generate gluconic acid and hydrogen peroxide. In industrial production, GOx has been utilized to produce miniature biofuel cells that could use glucose from human blood as a source of energy (Wong et al., 2008). Wild-type GOx has an optimal pH at ~pH 5.0 and KM against glucose of ~30 mM. However, under the physiological conditions in human blood, the pH is ~ 7.4 and the concentration of glucose is ~ 4 mM. To expand its medical applications, an engineered GOx variant adapting to the physiological conditions is needed. Emulsion technology combined with FACS was used for GOx engineering. However, one 9

significant challenge in GOx engineering is how to coordinate to start the reaction at the same time through precise addition of glucose substrate, as the glucose substrate can also be used by the host cells if it is added to the reaction mixture at the very beginning. Recently, Ostafe et al. developed a modified Aga1-Aga2 YSD system combined with directed evolution, emulsion technology, and FACS to engineer GOx for higher activity under physiological conditions (Ostafe et al., 2014). In this modified Aga1-Aga2 YSD system, GOx variants were firstly expressed and presented on the yeast cell surface. Then, the cells were encapsulated in water-in oil single emulsions with components including glucosidase, HRP, and tyramide-fluorescein. Instead of adding glucose substrate, β-octylglucoside was added to stay at the oil-water interface of the emulsion compartment, where β-octylglucoside was cleaved by the β-glucosidase to release the glucose into the water phase. Released glucose was then oxidized by GOx to generate H2O2, which facilitated the ligation of tyramide-fluorescein to the tyrosine residue of the yeast cell surface proteins under the catalysis of HRP, staining the cell with fluorescein (Figure 2C). The in vitro compartmentalization established by Ostafe and his colleagues has the advantage to establish an independent reaction system, in which the substrate and reaction condition can be easily manipulated. After a library containing about 107 GOx mutants was analyzed, several variants were obtained to possess slightly increased kcat/KM (1.2 - 1.8 fold). More interestingly, the Aga1-Aga2 YSD system was also used to express and purify GOx and its variants by fusing the enzymes to the Cterminus of Aga2 (Blazic et al., 2013). This progress indicates that Aga1-Aga2 YSD is suitable for the development of screening platforms for isolating engineered GOx variants. Besides, it could also be used as a whole-cell catalyst, either in suspension or immobilized to a solid phase, for the efficient production of gluconic acid. In addition to GOx engineering, the Aga1-Aga2 YSD system has also been recently used for engineering the DhbE adenylation domain of nonribosomal peptide synthetase (NRPS) against different substrates (Zhang et al., 2013b). Through displaying the adenylation domain library of DhbE on the cell surface, mutants with 11- and 6-fold increase of kcat/KM for nonnative substrates 3-hydroxybenzoic acid and 2-aminobenzoic acid, respectively, were isolated.

Engineering enzymes for altered substrate specificity and increased catalytic activity In addition to obtaining a higher catalytic activity, engineering enzyme to recognize the designed 10

substrate other than its native ones is also crucial for its practical application. An engineered enzyme with new substrate specificity, especially with deliberately designed substrate specificity could target the enzyme for specific aims, which is crucial for its application in academia and medicine. Similar to GOx, special modifications have to be utilized to fit different requirements of various catalytic enzymes. Inhere, three examples including HRP, TEV-P and SrtA engineering are briefly described of obtaining both the increased enzyme activity and altered substrate specificity based on the modified Aga1-Aga2 YSD technologies. Horseradish peroxidase engineering Enzymes are attractive catalysts for applications in organic chemistry. HRP, a widely used model enzyme that can catalyze the oxidation of a wide variety of electron donor substrates, contains four disulfide bonds as well as a heme prosthetic group. Since horseradish peroxidase (HRP) is difficult to be expressed as a functional form in bacteria, directed evolution against HRP did not succeed in bacterial or phage platforms. In 2007, Lipovsek and coworkers achieved the engineering of HRP into a different functional version by using a modified Aga1-Aga2 YSD system (Lipovsek et al., 2007). In this newly developed method, the HRP fused with a c-myc epitope tag was placed at the C-terminus of Aga2, followed by display on the yeast cell surface. The yeast cells were then incubated with its Alexa 488 labeled substrates (L-tyrosinol or D-tyrosinol), which could be attached to the tyrosine residues of the yeast endogenous surface proteins under the catalysis of the surface-displayed HRP (Figure 2D). Based on this strategy, two different HRP-derived libraries were constructed to isolate HRP variants with higher substrate specificity against either L-tyrosinol over D-tyrosinol or D-tyrosinol over L-tyrosinol. One variant library was constructed via an error-prone PCR approach and the other as a complete combinatorial library against five positions within the vicinity of the active site. Finally, several HRP variants were obtained with the altered enantioselectivities for L-tyrosinol or D-tyrosinol, among which one variant presented preferred substrate specificity switched from L-tyrosinol to Dtyrosinol with up to 8-fold (Lipovsek et al., 2007). It is worth noting that the variants from the errorprone PCR library exhibited no significantly improved enantioselectivity compared to the complete combinatorial library, which was probably due to the small library size (1.6×66) analyzed in the experiment. The successful work of HRP engineering brings a feasible strategy for engineering other similar enzymes, even if the substrates are not natural components on the cell wall. Using the similar strategy, the first synthetic substrate of the target enzyme is tethered to the surface of yeast, and then 11

the second substrate is added in solution mix. TEV protease engineering Tobacco Etch Virus protease (TEV-P), as a cysteine endopeptidase, plays an important role in research and industry for its stringent substrate specificity. TEV-P has been widely used in affinity tag removal during protein and peptide purification (Kapust et al., 2002; Zhang et al., 2015; Julien et al., 2016), modified proenzyme activation (Gray et al., 2010; Manuvera et al., 2015) and so on. As a representative protease, TEV-P has high specificity against its peptide substrate of ENLYFQS, with the cleavage site between Gln and Ser. The broad applications and stringent substrate specificity make the effectiveness of TEV-P engineering being widely used as an evaluation criteria for any newly developed method for protease engineering. Recently, Yi et al. optimized the Aga1-Aga2 YSD system to establish the Yeast ER Sequestration System (YESS) approach, enabling the directed evolution of protease for high hydrolytic activity and selectivity in yeast (Yi et al., 2013). Using TEV-P as an example (Figure 2E), TEV-P and its substrates were co-expressed and guided to the yeast ER, facilitating the proteolytic reactions. To evolve variants with high specificity, both the counter-selection and selection substrates were fused to the C-terminus of Aga2, with two different epitope tags, FLAG and 6His, anchored right before and after the selection substrate, respectively. In addition, the ER retention sequence, FEHDEL, which could significantly retain the protease and its substrate in the yeast ER (Lewis et al., 1990; Semenza et al., 1990), was fused at the C-terminus of both the 6xHis and the TEV-P sequences, eventually forming an FEHDEL-(TEV-P)-GAL1GAL10-Aga2-(counterselection substrate)-FLAG-(selection substrate)-6His-FEHDEL cassette. The use of ER sequestration effect makes the ER as the reaction container, favoring the enzyme folding and strengthening the proteolytic reactions. After the proteolytic reaction, the products fused to Aga2 were transported to the cell surface. Cells were then labeled with anti-FLAG and anti-6His antibodies that were conjugated with phycoerythrin (PE) and fluorescein (FITC) fluorophores, respectively. Cleavage of the selection substrate by different TEV-P variants was evaluated via FACS analysis by monitoring the ratio of FITC to PE fluorescence intensity. Cells exhibiting relatively high PE, but little or no FITC fluorescence, were assumed to indicate specific cleavage at only the selection substrate (Yi et al., 2013). Using the YESS method, the TEV protease was successfully engineered to recognize Glu or His, rather than the wild-type required Gln at the P1 position, of its canonical substrate. The kcat/KM values of the engineered variants against ENLYFES and ENLYFHS were 2.06 ± 0.46 mM-1s-1 and 0.15 ± 0.02 12

mM-1s-1, respectively, exhibiting the overall changes in selectivity for the Glu and His specific TEV variants of up to 5,000-fold and 1,100-fold, respectively. More importantly, the engineered variant TEV-PE10 showed a higher catalytic turnover activity than the wild-type TEV-P, and a variant TEVFast presented more than 4-fold higher activity than the wild-type TEV-P. It is worthy pointing out that the work presented using YESS is the first successful case of the protease engineering using highthroughput screening method, and Yi with his colleagues indicated that it could be potentially used for human protease granzyme K (GrK) and human Abelson tysosine kinase (AblTK) engineering (Yi et al., 2013). Recently, based on similar concept of using yeast ER as the reaction container, Guerrero et al. successfully engineered the human kallikrein 7 to recognize an amyloid beta peptide containing the central hydrophobic core by a FRET based method, protease evolution via cleavage of an intracellular substrate (PrECISE) (Guerrero et al., 2016). Sortase A engineering Staphylococcus aureus sortase A (SrtA), a bacterial transpeptidase possesses a cysteine proteaselike activity (Tsukiji and Nagamune, 2009), which has been widely used in industry and academia for conjugating proteins with various molecules. It binds the specific peptide sequence of LPXTG (where X = any amino acid) and cleaves the Thr-Gly peptide bond. This enzymatic reaction is followed by covalent conjugation of the LPXT to the target protein containing a penta-glycine sequence, finally forming a LPXT-(G)n-protein product. Wild-type SrtA exhibits low catalytic activity and strict specificity against its native substrates, which significantly limits its application. To expand the application of SrtA, Liu’s research group recently developed a modified bondforming enzyme screening system to enable the laboratory evolution of SrtA variants with dramatically altered substrate specificity (Chen et al., 2011; Dorr et al., 2014). In this modified YSD (Figure 2F), a modified Aga1 gene with a DNA fragment encoding the S6 peptide (GDSLSWLLRLLLN) at its 5’ end was firstly integrated into the yeast genome, whose expression was under control of 3-glyceraldehydephosphate dehydrogenase (GPD) promoter. The substrate A (CoA-LPETGG) was then conjugated to the serine residue within the S6 peptide using Bacillus subtilis Sfp phosphopantetheinyl transferase after the S6-Aga1 fusion complex was displayed on the cell surface. Meanwhile, the Aga2 fused HASrtA protein complex was displayed on the cell surface through the covalently attachment between Aga1 and Aga2. When the cells were incubated with the substrate B (GGGYK-biotin), the active SrtA could catalyze the formation of CoA-LPETGG, forming the final product of LPETGGGYK-biotin 13

conjugated on the surface displayed Aga1. Then, the cells were stained with anti-HA-Alexa Fluor 488 and streptavidin-PE antibodies, evaluating the levels of surface displayed SrtA and generated LPETGGGYK-biotin product by FACS technology. The high fluorescence signals ratio of Alexa Fluor 488/PE indicated the highly active SrtA variants. Through this modified Aga1-Aga2 YSD method, a SrtA variant with up to a 140-fold increase in LPETG-coupling activity was isolated (Chen et al., 2011). Continuing this work, Liu’s research group further modified the method to evolve SrtA variants that specifically recognized LAXTG and LPXSG, respectively, by introducing the negative selection (Dorr et al., 2014). For example, the nonbiotinylated LPETG was used as a competitor for biotin– LAETG to facilitate the isolation of SrtA variants possessing higher specificity against LAETG over LPETG, especially when the competitor’s concentration was much higher than that of biotin–LAETG. Using this strategy, a SrtA variant, eSrtA(2A-9), was isolated presenting 510-fold preference for LAETG over LPETG compared to the starting SrtA, which had a 103-fold preference for LPETG over LAETG. These numbers were added together for an extraordinary 51,000-fold specificity alteration. It is also noteworthy that the eSrtA variants were later applied to label the Fetuin A in human plasma (Dorr et al., 2014), conjugate thrombomodulin (TM) with alkylamine derivatives (Qu et al., 2014), prepare the scFv and TM conjugated multifunctional protein micelles (Kim et al., 2015), and in situ strip and recharge LPETG conjugated TM to the pentaglycine-derivated polyurethane catheters in the jugular vein within a rat model (Ham et al., 2016).

Conclusion and prospect Since Boder et al. had established the Aga1-Aga2 YSD system in 1997 (Boder and Wittrup, 1997), the applications of YSD in protein engineering have been quickly developed. After approximately 20 years’ development, the Aga1-Aga2 YSD system has been approved to be the most abundantly used YSD system with many successful applications, expanding from antibody engineering to non-antibody protein engineering. More interestingly, the Aga1-Aga2 YSD system has recently also been implanted to the Pichia pastoris cells (Jacobs et al., 2008), further expanding the potential application of Aga1Aga2 YSD system. As a matured methodology platform, the association of FACS with YSD has proven its power in 14

protein engineering, with its extraordinary advantages of the fine-tuning of protein affinity, specificity, and activity. With the combination of computational structural biology, the variant library can be constructed wisely, enabling more effective variants existing in the same size of variant library (Cherf and Cochran, 2015). For example, Ting’s research group recently further engineered the LplA using Rosetta-based computational design to specifically ligate the red dye resorufin to LAP (Liu et al., 2014a). The deep sequencing technology is another powerful technology that has been recently combined with YSD, providing an alternative way to precisely and efficiently map the enriched variants among a big variant library in parallel (D'Angelo et al., 2014; Van Blarcom et al., 2015). Recently, scientists from Pfizer Inc. combined the rational variant library design, YSD, and deep sequencing technologies to develop a strategy that was used to quickly map the epitopes of a panel of antibodies against α-toxin from Staphylococcus aureus in several weeks (D'Angelo et al., 2014; Van Blarcom et al., 2015). In addition, the display of human cDNA library with YSD also significantly expanded its application in defining potential cross-reactive proteins for small molecules or drugs (Wadle et al., 2005; Bidlingmaier and Liu, 2006; Bidlingmaier et al., 2016). It can be anticipated that the smart library construction, next-generation DNA sequencing, and bioinformatic analysis will facilitate the development of YSD to a powerful screening method with efficiency and simplicity, which will be the future direction of protein engineering.

Compliance with ethical standards Ethical standards This article does not contain any studies with human participants or animals performed by any of the authors.

Conflict of interest The authors declare that they have no competing interests. Acknowledgements This work was supported by National Science Foundation of China (No. 31540068 to L.Y. No. 31670069 to G.Z.), the Natural Science Foundation of Hubei Province of China (No. 2015CFA088 to L.Y.), the Ministry of Science and Technology of China (863 program 2014AA022203C to G.Z.)

15

References Bagriantsev SN, Chatelain FC, Clark KA, Alagem N, Reuveny E, Minor DL, Jr. Tethered protein display identifies a novel Kir3.2 (GIRK2) regulator from protein scaffold libraries. ACS Chem Neurosci 2014;5(9):812-22. Baruah H, Puthenveetil S, Choi YA, Shah S, Ting AY. An engineered aryl azide ligase for site-specific mapping of protein-protein interactions through photo-cross-linking. Angew Chem Int Ed Engl 2008;47(37):7018-21. Benatuil L, Perez JM, Belk J, Hsieh CM. An improved yeast transformation method for the generation of very large human antibody libraries. Protein engineering, design & selection : PEDS 2010;23(4):155-9. Bidlingmaier S, Liu B. Construction and application of a yeast surface-displayed human cDNA library to identify post-translational modification-dependent protein-protein interactions. Mol Cell Proteomics 2006;5(3):533-40. Bidlingmaier S, Ha K, Lee NK, Su Y, Liu B. Proteome-wide Identification of Novel Ceramide-binding Proteins by Yeast Surface cDNA Display and Deep Sequencing. Molecular & cellular proteomics : MCP 2016;15(4):1232-45. Blazic M, Kovacevic G, Prodanovic O, Ostafe R, Gavrovic-Jankulovic M, Fischer R, Prodanovic R. Yeast surface display for the expression, purification and characterization of wild-type and B11 mutant glucose oxidases. Protein expression and purification 2013;89(2):175-80. Boder ET, Wittrup KD. Yeast surface display for screening combinatorial polypeptide libraries. Nature biotechnology 1997;15(6):553-7. Boder ET, Bill JR, Nields AW, Marrack PC, Kappler JW. Yeast surface display of a noncovalent MHC class II heterodimer complexed with antigenic peptide. Biotechnology and bioengineering 2005;92(4):485-91. Buonpane RA, Churchill HR, Moza B, Sundberg EJ, Peterson ML, Schlievert PM, Kranz DM. Neutralization of staphylococcal enterotoxin B by soluble, high-affinity receptor antagonists. Nature medicine 2007;13(6):725-9. Casares S, Hurtado A, McEvoy RC, Sarukhan A, von Boehmer H, Brumeanu TD. Down-regulation of diabetogenic CD4+ T cells by a soluble dimeric peptide-MHC class II chimera. Nat Immunol 2002;3(4):383-91. Chao G, Cochran JR, Wittrup KD. Fine epitope mapping of anti-epidermal growth factor receptor antibodies through random mutagenesis and yeast surface display. Journal of molecular biology 2004;342(2):539-50. Chen I, Dorr BM, Liu DR. A general strategy for the evolution of bond-forming enzymes using yeast display. Proc Natl Acad Sci U S A 2011;108(28):11399-404. Chen X, Xiao Y, Shen W, Govender A, Zhang L, Fan Y, Wang Z. Display of phytase on the cell surface of Saccharomyces cerevisiae to degrade phytate phosphorus and improve bioethanol production. Applied microbiology and biotechnology 2016;100(5):2449-58. Cherf GM, Cochran JR. Applications of Yeast Surface Display for Protein Engineering. Methods in molecular biology (Clifton, NJ) 2015;1319:155-75. Cochran JR, Kim YS, Lippow SM, Rao B, Wittrup KD. Improved mutants from directed evolution are biased to orthologous substitutions. Protein engineering, design & selection : PEDS 2006;19(6):245-53. D'Angelo S, Kumar S, Naranjo L, Ferrara F, Kiss C, Bradbury AR. From deep sequencing to actual 16

clones. Protein engineering, design & selection : PEDS 2014;27(10):301-7. Delhalle S, Schmit JC, Chevigne A. Phages and HIV-1: from display to interplay. International journal of molecular sciences 2012;13(4):4727-94. Doerner A, Rhiel L, Zielonka S, Kolmar H. Therapeutic antibody engineering by high efficiency cell screening. FEBS letters 2014;588(2):278-87. Domingo-Calap P, Georgel P, Bahram S. Back to the future: bacteriophages as promising therapeutic tools. Hla 2016;87(3):133-40. Dorr BM, Ham HO, An C, Chaikof EL, Liu DR. Reprogramming the specificity of sortase enzymes. Proc Natl Acad Sci U S A 2014;111(37):13343-8. Dudgeon K, Rouet R, Famm K, Christ D. Selection of human VH single domains with improved biophysical properties by phage display. Methods in molecular biology (Clifton, NJ) 2012;911:383-97. Feldhaus MJ, Siegel RW, Opresko LK, Coleman JR, Feldhaus JM, Yeung YA, Cochran JR, Heinzelman P, Colby D, Swers J, Graff C, Wiley HS, Wittrup KD. Flow-cytometric isolation of human antibodies from a nonimmune Saccharomyces cerevisiae surface display library. Nature biotechnology 2003;21(2):163-70. Fernandez-Suarez M, Baruah H, Martinez-Hernandez L, Xie KT, Baskin JM, Bertozzi CR, Ting AY. Redirecting lipoic acid ligase for cell surface protein labeling with small-molecule probes. Nature biotechnology 2007;25(12):1483-7. Frago S, Nicholls RD, Strickland M, Hughes J, Williams C, Garner L, Surakhy M. Functional evolution of IGF2:IGF2R domain 11 binding generates novel structural interactions and a specific IGF2 antagonist. 2016. Fujiwara K, Takeuchi S, Okamura-Ikeda K, Motokawa Y. Purification, characterization, and cDNA cloning of lipoate-activating enzyme from bovine liver. J Biol Chem 2001;276(31):28819-23. Fujiwara K, Toma S, Okamura-Ikeda K, Motokawa Y, Nakagawa A, Taniguchi H. Crystal structure of lipoate-protein ligase A from Escherichia coli. Determination of the lipoic acid-binding site. The Journal of biological chemistry 2005;280(39):33645-51. Fujiwara K, Maita N, Hosaka H, Okamura-Ikeda K, Nakagawa A, Taniguchi H. Global conformational change associated with the two-step reaction catalyzed by Escherichia coli lipoate-protein ligase A. The Journal of biological chemistry 2010;285(13):9971-80. Georgiou G, Stathopoulos C, Daugherty PS, Nayak AR, Iverson BL, Curtiss R, 3rd. Display of heterologous proteins on the surface of microorganisms: from the screening of combinatorial libraries to live recombinant vaccines. Nature biotechnology 1997;15(1):29-34. Griffiths AD, Williams SC, Hartley O, Tomlinson IM, Waterhouse P, Crosby WL, Kontermann RE, Jones PT, Low NM, Allison TJ, et al. Isolation of high affinity human antibodies directly from large synthetic repertoires. The EMBO journal 1994;13(14):3245-60. Guerrero JL, O'Malley MA, Daugherty PS. Intracellular FRET-based Screen for Redesigning the Specificity of Secreted Proteases. ACS chemical biology 2016;11(4):961-70. Ham HO, Qu Z, Haller CA, Dorr BM, Dai E, Kim W, Liu DR, Chaikof EL. In situ regeneration of bioactive coatings enabled by an evolved Staphylococcus aureus sortase A. Nature communications 2016;7:11140. Harris DT, Wang N, Riley TP, Anderson SD, Singh NK, Procko E, Baker BM, Kranz DM. Deep mutational scans as a guide to engineering high-affinity T cell receptor interactions with peptidebound MHC. The Journal of biological chemistry 2016. 17

Heimer BW, Tam BE, Sikes HD. Characterization and directed evolution of a methyl-binding domain protein for high-sensitivity DNA methylation analysis. Protein engineering, design & selection : PEDS 2015;28(12):543-51. Hockney RC. Recent developments in heterologous protein production in Escherichia coli. Trends in biotechnology 1994;12(11):456-63. Holler PD, Holman PO, Shusta EV, O'Herrin S, Wittrup KD, Kranz DM. In vitro evolution of a T cell receptor with high affinity for peptide/MHC. Proceedings of the National Academy of Sciences of the United States of America 2000;97(10):5387-92. Holt LJ, Herring C, Jespers LS, Woolven BP, Tomlinson IM. Domain antibodies: proteins for therapy. Trends Biotechnol 2003;21(11):484-90. Inokuma K, Yoshida T, Ishii J, Hasunuma T, Kondo A. Efficient co-displaying and artificial ratio control of alpha-amylase and glucoamylase on the yeast cell surface by using combinations of different anchoring domains. Applied microbiology and biotechnology 2015;99(4):1655-63. Jacobs PP, Ryckaert S, Geysens S, De Vusser K, Callewaert N, Contreras R. Pichia surface display: display of proteins on the surface of glycoengineered Pichia pastoris strains. Biotechnology letters 2008;30(12):2173-81. Jiang W, Boder ET. High-throughput engineering and analysis of peptide binding to class II MHC. Proc Natl Acad Sci U S A 2010;107(30):13258-63. Jin M, Song G, Carman CV, Kim YS, Astrof NS, Shimaoka M, Wittrup DK, Springer TA. Directed evolution to probe protein allostery and integrin I domains of 200,000-fold higher affinity. Proceedings of the National Academy of Sciences of the United States of America 2006;103(15):5758-63. Jin Z, Zhang K, Zhang L, Zheng SP, Han SY, Lin Y. Quantification analysis of yeast-displayed lipase. Anal Biochem 2014;450:46-8. Kariolis MS, Miao YR, Jones DS, 2nd, Kapur S, Mathews, II, Giaccia AJ, Cochran JR. An engineered Axl 'decoy receptor' effectively silences the Gas6-Axl signaling axis. Nature chemical biology 2014;10(11):977-83. Kim W, Haller C, Dai E, Wang X, Hagemeyer CE, Liu DR, Peter K, Chaikof EL. Targeted antithrombotic protein micelles. Angewandte Chemie (International ed in English) 2015;54(5):1461-5. Kondo A, Ueda M. Yeast cell-surface display--applications of molecular display. Appl Microbiol Biotechnol 2004;64(1):28-40. Kuroda K, Ueda M. Generation of arming yeasts with active proteins and peptides via cell surface display system: cell surface engineering, bio-arming technology. Methods in molecular biology (Clifton, NJ) 2014;1152:137-55. Lewis MJ, Sweet DJ, Pelham HR. The ERD2 gene determines the specificity of the luminal ER protein retention system. Cell 1990;61(7):1359-63. Lipke PN, Kurjan J. Sexual agglutination in budding yeasts: structure, function, and regulation of adhesion glycoproteins. Microbiol Rev 1992;56(1):180-94. Lipke PN, Wojciechowicz D, Kurjan J. AG alpha 1 is the structural gene for the Saccharomyces cerevisiae alpha-agglutinin, a cell surface glycoprotein involved in cell-cell interactions during mating. Molecular and cellular biology 1989;9(8):3155-65. Lipovsek D, Antipov E, Armstrong KA, Olsen MJ, Klibanov AM, Tidor B, Wittrup KD. Selection of horseradish peroxidase variants with enhanced enantioselectivity by yeast surface display. Chem 18

Biol 2007;14(10):1176-85. Little M, Fuchs P, Breitling F, Dubel S. Bacterial surface presentation of proteins and peptides: an alternative to phage technology? Trends in biotechnology 1993;11(1):3-5. Liu DS, Phipps WS, Loh KH, Howarth M, Ting AY. Quantum dot targeting with lipoic acid ligase and HaloTag for single-molecule imaging on living cells. ACS nano 2012a;6(12):11080-7. Liu DS, Tangpeerachaikul A, Selvaraj R, Taylor MT, Fox JM, Ting AY. Diels-Alder cycloaddition for fluorophore targeting to specific proteins inside living cells. Journal of the American Chemical Society 2012b;134(2):792-5. Liu DS, Nivon LG, Richter F, Goldman PJ, Deerinck TJ, Yao JZ, Richardson D, Phipps WS, Ye AZ, Ellisman MH, Drennan CL, Baker D, Ting AY. Computational design of a red fluorophore ligase for site-specific protein labeling in living cells. Proceedings of the National Academy of Sciences of the United States of America 2014a;111(43):E4551-9. Liu Y, Zhang T, Qiao J, Liu X, Bo J, Wang J, Lu F. High-yield phosphatidylserine production via yeast surface display of phospholipase D from Streptomyces chromofuscus on Pichia pastoris. Journal of agricultural and food chemistry 2014b;62(23):5354-60. Lu CF, Montijn RC, Brown JL, Klis F, Kurjan J, Bussey H, Lipke PN. Glycosyl phosphatidylinositoldependent cross-linking of alpha-agglutinin and beta 1,6-glucan in the Saccharomyces cerevisiae cell wall. The Journal of cell biology 1995;128(3):333-40. Matsuura H, Yamamoto Y, Muraoka M, Akaishi K, Hori Y, Uemura K, Tsuji N, Harada K, Hirata K, Bamba T, Miyasaka H, Kuroda K, Ueda M. Development of surface-engineered yeast cells displaying phytochelatin synthase and their application to cadmium biosensors by the combined use of pyrene-excimer fluorescence. Biotechnol Prog 2013;29(5):1197-202. Mattis DM, Chervin AS, Ranoa DR, Kelley SL, Tapping RI, Kranz DM. Studies of the TLR4associated protein MD-2 using yeast-display and mutational analyses. Molecular immunology 2015;68(2 Pt A):203-12. Maute RL, Gordon SR, Mayer AT, McCracken MN, Natarajan A, Ring NG, Kimura R, Tsai JM, Manglik A, Kruse AC, Gambhir SS, Weissman IL, Ring AM. Engineering high-affinity PD-1 variants for optimized immunotherapy and immuno-PET imaging. Proceedings of the National Academy of Sciences of the United States of America 2015;112(47):E6506-14. Ostafe R, Prodanovic R, Nazor J, Fischer R. Ultra-high-throughput screening method for the directed evolution of glucose oxidase. Chemistry & biology 2014;21(3):414-21. Popovic M, Prodanovic R, Ostafe R, Schillberg S, Fischer R, Gavrovic-Jankulovic M. Yeast surface display is a novel tool for the rapid immunological characterization of plant-derived food allergens. Immunologic research 2015;61(3):230-9. Puthenveetil S, Liu DS, White KA, Thompson S, Ting AY. Yeast display evolution of a kinetically efficient 13-amino acid substrate for lipoic acid ligase. Journal of the American Chemical Society 2009;131(45):16430-8. Qu Z, Krishnamurthy V, Haller CA, Dorr BM, Marzec UM, Hurst S, Hinds MT, Hanson SR, Liu DR, Chaikof EL. Immobilization of actively thromboresistant assemblies on sterile blood-contacting surfaces. Advanced healthcare materials 2014;3(1):30-5. Rhiel L, Krah S, Gunther R, Becker S, Kolmar H, Hock B. REAL-Select: full-length antibody display and library screening by surface capture on yeast cells. PloS one 2014;9(12):e114887. Schreuder MP, Brekelmans S, van den Ende H, Klis FM. Targeting of a heterologous protein to the cell wall of Saccharomyces cerevisiae. Yeast (Chichester, England) 1993;9(4):399-409. 19

Schuurmann J, Quehl P, Festel G, Jose J. Bacterial whole-cell biocatalysts by surface display of enzymes: toward industrial application. Applied microbiology and biotechnology 2014;98(19):8031-46. Scott JK, Smith GP. Searching for peptide ligands with an epitope library. Science (New York, NY) 1990;249(4967):386-90. Semenza JC, Hardwick KG, Dean N, Pelham HR. ERD2, a yeast gene required for the receptormediated retrieval of luminal ER proteins from the secretory pathway. Cell 1990;61(7):1349-57. Sette A, Fikes J. Epitope-based vaccines: an update on epitope identification, vaccine design and delivery. Curr Opin Immunol 2003;15(4):461-70. Sharma P, Postel S, Sundberg EJ, Kranz DM. Characterization of the Staphylococcal enterotoxin A: Vbeta receptor interaction using human receptor fragments engineered for high affinity. Protein engineering, design & selection : PEDS 2013;26(12):781-9. Shiraga S, Kawakami M, Ishiguro M, Ueda M. Enhanced reactivity of Rhizopus oryzae lipase displayed on yeast cell surfaces in organic solvents: potential as a whole-cell biocatalyst in organic solvents. Applied and environmental microbiology 2005;71(8):4335-8. Szczupak A, Kol-Kalman D, Alfonta L. A hybrid biocathode: surface display of O2-reducing enzymes for microbial fuel cell applications. Chemical communications (Cambridge, England) 2012;48(1):49-51. Tanaka T, Yamada R, Ogino C, Kondo A. Recent developments in yeast cell surface display toward extended applications in biotechnology. Appl Microbiol Biotechnol 2012;95(3):577-91. Tanino T, Fukuda H, Kondo A. Construction of a Pichia pastoris cell-surface display system using Flo1p anchor system. Biotechnology progress 2006;22(4):989-93. Tokuhiro K, Ishida N, Kondo A, Takahashi H. Lactic fermentation of cellobiose by a yeast strain displaying beta-glucosidase on the cell surface. Applied microbiology and biotechnology 2008;79(3):481-8. Tsukiji S, Nagamune T. Sortase-mediated ligation: a gift from Gram-positive bacteria to protein engineering. Chembiochem 2009;10(5):787-98. Ueda M. [Novel high-throughput system for production of new medicines-integration and combination with molecular display and combinatorial bioengineering]. Yakugaku Zasshi 2009;129(11):127784. Ueda M, Tanaka A. Cell surface engineering of yeast: construction of arming yeast with biocatalyst. Journal of bioscience and bioengineering 2000;90(2):125-36. Van Blarcom T, Rossi A, Foletti D, Sundar P, Pitts S, Bee C, Melton Witt J, Melton Z, Hasa-Moreno A, Shaughnessy L, Telman D, Zhao L, Cheung WL, Berka J, Zhai W, Strop P, Chaparro-Riggers J, Shelton DL, Pons J, Rajpal A. Precise and efficient antibody epitope determination through library design, yeast display and next-generation sequencing. J Mol Biol 2015;427(6 Pt B):1513-34. Van der Vaart JM, te Biesebeke R, Chapman JW, Toschka HY, Klis FM, Verrips CT. Comparison of cell wall proteins of Saccharomyces cerevisiae as anchors for cell surface expression of heterologous proteins. Applied and environmental microbiology 1997;63(2):615-20. Van Deventer JA, Wittrup KD. Yeast surface display for antibody isolation: library construction, library screening, and affinity maturation. Methods in molecular biology (Clifton, NJ) 2014;1131:151-81. Van Mulders SE, Christianen E, Saerens SM, Daenen L, Verbelen PJ, Willaert R, Verstrepen KJ, Delvaux FR. Phenotypic diversity of Flo protein family-mediated adhesion in Saccharomyces cerevisiae. FEMS yeast research 2009;9(2):178-90. 20

Wadle A, Mischo A, Imig J, Wullner B, Hensel D, Watzig K, Neumann F, Kubuschok B, Schmidt W, Old LJ, Pfreundschuh M, Renner C. Serological identification of breast cancer-related antigens from a Saccharomyces cerevisiae surface display library. Int J Cancer 2005;117(1):104-13. Wang H, Lang Q, Li L, Liang B, Tang X, Kong L, Mascini M, Liu A. Yeast surface displaying glucose oxidase as whole-cell biocatalyst: construction, characterization, and its electrochemical glucose sensing application. Analytical chemistry 2013;85(12):6107-12. Wasilenko JL, Sarmento L, Spatz S, Pantin-Jackwood M. Cell surface display of highly pathogenic avian influenza virus hemagglutinin on the surface of Pichia pastoris cells using alpha-agglutinin for production of oral vaccines. Biotechnology progress 2010;26(2):542-7. Watts C. The exogenous pathway for antigen presentation on major histocompatibility complex class II and CD1 molecules. Nat Immunol 2004;5(7):685-92. Weber KS, Donermeyer DL, Allen PM, Kranz DM. Class II-restricted T cell receptor engineered in vitro for higher affinity retains peptide specificity and function. Proceedings of the National Academy of Sciences of the United States of America 2005;102(52):19033-8. Weiskopf K, Ring AM, Ho CC, Volkmer JP, Levin AM, Volkmer AK, Ozkan E, Fernhoff NB, van de Rijn M, Weissman IL, Garcia KC. Engineered SIRPalpha variants as immunotherapeutic adjuvants to anticancer antibodies. Science (New York, NY) 2013;341(6141):88-91. Wen F, Sethi DK, Wucherpfennig KW, Zhao H. Cell surface display of functional human MHC class II proteins: yeast display versus insect cell display. Protein Eng Des Sel 2011;24(9):701-9. Wesolowski J, Alzogaray V, Reyelt J, Unger M, Juarez K, Urrutia M, Cauerhff A, Danquah W, Rissiek B, Scheuplein F, Schwarz N, Adriouch S, Boyer O, Seman M, Licea A, Serreze DV, Goldbaum FA, Haag F, Koch-Nolte F. Single domain antibodies: promising experimental and therapeutic tools in infection and immunity. Medical microbiology and immunology 2009;198(3):157-74. Wong CM, Wong KH, Chen XD. Glucose oxidase: natural occurrence, function, properties and industrial applications. Appl Microbiol Biotechnol 2008;78(6):927-38. Yi L, Gebhard MC, Li Q, Taft JM, Georgiou G, Iverson BL. Engineering of TEV protease variants by yeast ER sequestration screening (YESS) of combinatorial libraries. Proceedings of the National Academy of Sciences of the United States of America 2013;110(18):7229-34. Zaretsky M, Etzyoni R, Kaye J, Sklair-Tavron L, Aharoni A. Directed evolution of a soluble human IL17A receptor for the inhibition of psoriasis plaque formation in a mouse model. Chemistry & biology 2013;20(2):202-11. Zhang K, Li H, Bhuripanyo K, Zhao B, Chen TF, Zheng N, Yin J. Engineering new protein-protein interactions on the beta-propeller fold by yeast cell surface display. Chembiochem : a European journal of chemical biology 2013a;14(4):426-30. Zhang K, Nelson KM, Bhuripanyo K, Grimes KD, Zhao B, Aldrich CC, Yin J. Engineering the substrate specificity of the DhbE adenylation domain by yeast cell surface display. Chemistry & biology 2013b;20(1):92-101.

21

Figures Figure 1 Schematic summary of YSD platform for protein engineering

22

Figure 2 Overview of the modified Aga1-Aga2 YSD systems on non-antibody protein engineering

(A) Schematic presentation of the yeast surface co-display system for HLA-DR1 engineering (Jiang and Boder, 2010) (B) Schematic presentation of LAP evolution by YSD (Puthenveetil et al., 2009). K: Lysine. (C) Schematic presentation of GOx engineering by YSD combined with emulsion technology (Blazic et al., 2013). TF: Tyramide Fluorescein; GDL: DGluconolactone. (D) Schematic presentation of HRP engineering by YSD using the yeast surface protein as substrates (Lipovsek et al., 2007). (E) Schematic presentation of the YESS for TEV protease engineering (Yi et al., 2013). The cells exhibiting single fluorescence (circled by red dash lines) was isolated for variants having high specificity against targeted TEV-P substrate. ERS: ER retention sequence; CS: TEV-P counter-selection substrate; PS: TEV-P substrate. (F) Schematic presentation of sortase A engineering with or without competitive selections (Chen et al., 2011; Qu et al., 2014).

23

Tables Table 1 Cell wall proteins used in yeast surface display system Anchors

Carriers

Target protein fusion site

Hosts

Applications

References

MHC, LplA, GOx, HRP, (Jacobs et al., 2008;

Saccharomyces a-agglutinin

a-agglutinin

(Aga1p)

(Aga2p)

TEV-P, SrtA; hGal-1, N or C terminal

Cherf and Cochran,

cerevisiae, hEPO, mIFN-β, mIFN-γ,

2015), etc

Pichia pastoris etc.

α-agglutinin

α-agglutinin

β-glucosidase,

Saccharomyces

hemagglutinin;

cerevisiae

α-galactosidase

(Van der Vaart et al., 1997; Tokuhiro et al.,

C terminal (Agα1p)

Pichia pastoris,

(Agα1p)

2008; Wasilenko et al., 2010)

Flo1p

Flo1p

N or C terminal

Pichia pastoris,

phospholipase D;

(Van der Vaart et al.,

Saccharomyces

cellulolytic enzymes, α-

1997; Liu et al., 2014;

cerevisiae

galactosidase

Mo et al., 2014)

Saccharomyces Cwp1p

Cwp1p

C terminal

α-galactosidase

1997)

cerevisiae Saccharomyces Cwp2p

Cwp2p

C terminal

α-galactosidase

Saccharomyces Sed1p

C terminal

α-galactosidase

Saccharomyces Tip1p

C terminal

α-galactosidase

Saccharomyces Tir1p

C terminal

α-galactosidase

Saccharomyces YCR89w

C terminal cerevisiae

24

(Van der Vaart et al., 1997)

cerevisiae YCR89w

(Van der Vaart et al., 1997)

cerevisiae Tir1p

(Van der Vaart et al., 1997)

cerevisiae Tip1p

(Van der Vaart et al., 1997)

cerevisiae Sed1p

(Van der Vaart et al.,

α-galactosidase

(Van der Vaart et al., 1997)

Table 2 Proteins or substrates engineering using non-modified Aga1-Aga2 YSD technology

Protein/enzyme

Engineered Variants

Fold switch

KD, ªIC50, kcat/KM

NR

IC50 = 4.2±0.9 nM

30-fold

IC50 = 0.16±0.02 nM

EGF clone 28

15-fold

IC50 = 0.29±0.11 nM

EGF clone 30

4-fold

IC50 = 1.1±0.4 nM

WT IL-17RA

NR

KD=2.62±0.04 nM

6-fold

KD=0.45±0.2 nM

4.2-fold

KD=0.63±0.15 nM

Neh[ETGE]

NR

kcat/KM =126±5 mM-1.s-1

Neh2[ETGE]-E79K

NR

NR

Neh[ETGE]

NR

kcat/KM = 3.9±0.5 mM-1.s-1

Neh2[ETGE]-E79K

330

kcat/KM = 4.1±0.7 mM-1.s-1

Neh[ETGE]

NR

kcat/KM = 3.9±0.4 mM-1.s-1

Neh2[ETGE]-E79K

83

kcat/KM = 2.5±0.5 mM-1.s-1

DHB

NR

SA

NR

kcat/KM = 140 mM-1.min-1

3-HBA

NR

kcat/KM = 22 mM-1.min-1

2-ABA

NR

kcat/KM = 5.5 mM-1.min-1

DHB

NR

kcat/KM = 410 mM-1.min-1

3-HBA

200

kcat/KM = 240 mM-1.min-1

KZ12

SA

NR

kcat/KM = 6.6 mM-1.min-1

（W234H）

2-ABA

200

kcat/KM = 34 mM-1.min-1

WT Vβ22

Staphylococcal

NR

KD=100 μM

FL

enterotoxin A

25,000-fold

KD=4 nM

NR

KD=0.3~0.5 μM

12,000-fold

KD=41.3 pM

45,000-fold

KD=11.1 pM

Substrates

WT EGF Epidermal growth

EGF clone 114 hEGFR

factor (EGF)

IL-17A receptor

V3

IL-17A

V10

References

(Cochran et al., 2006)

(Zaretsky et al., 2013)

WT KR

Keap1 protein

KR1 (Zhang et al., 2013a)

KR2

kcat/KM = 1,100 mM-1.min-1

WT DhbE

(Zhang et al., 2013b)

DhbE KZ4（W234H）

Vβ22

(Sharma et al., 2013) WT SIRPα

SIRPα

FD6

CD47

FA4 WT Axl Ig1

ligand growth

NR

KD=33 pM

MYD1 Ig1

arrest-specific 6

12-fold

KD=2.7 pM

NR

KD=5.9±1.3 nM

1.4-fold

KD=4.4±0.4 nM

Var 2/5

2-fold

KD=3.1±1.0 nM

Programmed cell death

WT PD-1

NR

KD=8.2 μM

protein-1

HAC-PD-1-I

76,000-fold

KD=107 pM

(PD-1)

HAC-PD-1-V

74.000-fold

KD=110 pM

RD1- MART1

NR

KD=137 nM

2-fold

KD=76 nM

5-fold

KD=30 nM

(Kariolis et al., 2014)

Axl receptor

WT hMBD2 hMBD2

TCR

(Weiskopf et al., 2013)

Var 1/4

Y50Wα A99Yβ

methylated DNA

hPD-L1

MART1/HLA-A2

(Heimer et al., 2015)

(Maute et al., 2015)

(Harris et al., 2016)

ªIC50: The half-maximal value of binding, which is determined by competition binding of wild-type EGF and mutant proteins with EGFR (Cochran et al., 2006).

25

Table 3 Proteins or substrates engineering using modified Aga1-Aga2 YSD technology Protein/enzyme

Engineered Variants

Substrates

Fold switch

kcat/KM, KH

HLA-DR1

PKYVKQNTLKLAT

NR

NR

S4V1.5

PKVVKQNTLKLAT

3~14-fold

NR

(Jiang and Boder,

S4A1.9

PKAVKQNTLKLAT

NR

NR

2010)

S4E1.3

PKEVKQNTLKLAT

NR

NR

LAP1#

DEVLVEIETDKAVLEVPG

NR

kcat/KM < 0.0135 μM-1 min-1

(Puthenveetil et al.,

GFEIDKVWYDLDA

>70-fold

kcat/KM =0.99 mM-1.min-1

2009)

References

HLA-DR1

LplA LAP2

#

KM=23.19 ± 0.57mM WT GOx

NR

kcat=130.16 ± 3.97 s-1 Glucose

GOx

(Ostafe et al., 2014)

KM= 13.08 ± 0.49mM A2

5.8-fold

kcat =432.17 ± 16.38 s-1 WT HRP

L-tyrosinol

NR

NR

CD8.02

D-tyrosinol

3.8-fold

NR

CL8.01

L-tyrosinol

7-fold

NR

(Lipovsek et al., HRP

2007)

TEV protease

WT TEV-P

ENLYFQS

NR

kcat/KM=1.20±0.09 mM-1.s-1

TEV-PE10

ENLYFES

5000-fold

kcat/KM=2.06±0.46 mM-1.s-1

TEV-PH21

ENLYFHS

1100-fold

kcat/KM=0.15±0.02 mM .s

eSrtA(2A-9)

LAETG

5100-fold

KH=18.7±3.3 mM-1.s-1*

-1

(Yi et al., 2013)

-1

(Chen et al., 2011;

Sortase eSrtA(4S-9)

LPESG

120-fold

-1 -1

KH=295±18 mM .s *

Qu et al., 2014)

* KH, represents the concentration of GGG at which the rates of acyl–enzyme hydrolysis and transpeptidation are equal (Dorr et al., 2014). # LAP: LplA acceptor peptide; NR: not reported.

26