WO2009111183A1

WO2009111183A1 - Surface display of recombinant proteins in lower eukaryotes

Info

Publication number: WO2009111183A1
Application number: PCT/US2009/034631
Authority: WO
Inventors: Dongxing Zha; Stefan Wildt
Original assignee: Glycofi, Inc.
Priority date: 2008-03-03
Filing date: 2009-02-20
Publication date: 2009-09-11
Also published as: EP2263089A4; CA2715212A1; US9845464B2; US20100331192A1; JP5731827B2; JP2015130868A; US20150211000A1; EP2263089B1; US8877686B2; JP2011514807A; EP2263089A1

Abstract

Methods for display of recombinant proteins or protein libraries on the surface of lower eukaryotes such as yeast and filamentous fungi are described. The methods are useful for screening libraries of recombinant proteins in lower eukaryotes to identify particular proteins with desired properties from the array of proteins in the libraries. The methods are particularly useful for constructing and screening antibody libraries in lower eukaryotes.

Description

TITLE OF THE INVENTION

SURPACE DISPLAY OF RECOMBINANT PROTEINS IN LOWER EUKARYOTES

BACKGROUND OF THE INVENTION

(1) Field ofthe Invention

The present invention relates to methods for display of recombinant proteins or protein libraries on the surface of lower eukaryotes such as yeast and filamentous fungi. The methods are useful for screening libraries of recombinant proteins in lower eukaryotes to identify particular proteins with desired properties from the array of proteins in the libraries. The methods are particularly useful for constructing and screening antibody libraries in lower eukaryotes.

(2) Description of Related Art The discovery of monoclonal antibodies has evolved from hybridoma technology for producing the antibodies to direct selection of antibodies from human cDNA or synthetic DNA libraries. This has been driven in part by the desire to engineer improvements in binding affinity and specificity of the antibodies to improve efficacy of the antibodies. Thus, combinatorial library screening and selection methods have become a common tool for altering the recognition properties of proteins (Ellman et al., Proc. Natl. Acad. Sci. USA 94: 2779-2782 (1997): Phizicky & Fields, Microbiol. Rev. 59: 94-123 (1995)). The ability to construct and screen antibody libraries in vitro promises improved control over the strength and specificity of antibody-antigen interactions.

The most widespread technique for constructing and screening antibody libraries is phage display, whereby the protein of interest is expressed as a polypeptide fusion to a bacteriophage coat protein and subsequently screened by binding to immobilized or soluble biotinylated ligand. Fusions are made most commonly to a minor coat protein, called the gene III protein (pill), which is present in three to five copies at the tip of the phage. A phage constructed in this way can be considered a compact genetic "unit", possessing both the phenotype (binding activity of the displayed antibody) and genotype (the gene coding for that antibody) in one package. Phage display has been successfully applied to antibodies, DNA binding proteins, protease inhibitors, short peptides, and enzymes (Choo & Klug, Curr. Opin. Biotechnol. 6: 431- 436 (1995); Hoogenboom, Trends Biotechnol. 15: 62-70 (1997); Ladner, Trends Biotechnol. 13: 426-430 (1995); Lowman et al., Biochemistry 30: 10832-10838 (1991); Markland et al., Methods Enzymol. 267: 28-51 (1996); Matthews & Wells, Science 260: 11 13-1117 (1993); Wang et al., Methods Enzymol. 267: 52-68 (1996)). Antibodies possessing desirable binding properties are selected by binding to immobilized antigen in a process called "panning." Phage bearing nonspecific antibodies are removed by washing, and then the bound phage are eluted and amplified by infection of E. coli. This approach has been applied to generate antibodies against many antigens. Nevertheless, phage display possesses several shortcomings. Although panning of antibody phage display libraries is a powerful technology, it possesses several intrinsic difficulties that limit its wide-spread successful application. For example, some eukaryotic secreted proteins and cell surface proteins require post-transiational modifications such as glycosylation or extensive disulfide isomerization, which are unavailable in bacterial cells. Furthermore, the nature of phage display precludes quantitative and direct discrimination of ligand binding parameters. For example, very high affinity antibodies (K d≤= 1 nM) are difficult to isolate by panning, since the elution conditions required to break a very strong antibody- antigen interaction are generally harsh enough (e.g., low pH, high salt) to denature the phage particle sufficiently to render it non-infective. Additionally, the requirement for physical immobilization of an antigen to a solid surface produces many artifactual difficulties. For example, high antigen surface density introduces avidity effects which mask true affinity. Also, physical tethering reduces the translational and rotational entropy of the antigen, resulting in a smaller DS upon antibody binding and a resultant overestimate of binding affinity relative to that for soluble antigen and large effects from variability in mixing and washing procedures lead to difficulties with reproducibility. Furthermore, the presence of only one to a few antibodies per phage particle introduces substantial stochastic variation, and discrimination between antibodies of similar affinity becomes impossible. For example, affinity differences of 6-fold or greater are often required for efficient discrimination (Riechmann & Weill, '93). Finally, populations can be overtaken by more rapidly growing wild-type phage. In particular, since pill is involved directly in the phage life cycle, the presence of some antibodies or bound antigens will prevent or retard amplification of the associated phage.

Several bacterial cell surface display methods have been developed (Francisco, et al., Proc. Natl Acad. Sci. USA 90: 10444-10448 (1993); Georgiou et al., Nat. Biotechnol. 15: 29-34 (1997)). However, use of a prokaryotic expression system occasionally introduces unpredictable expression biases (Knappik & Pluckthun, Prot. Eng. 8: 81-89 (1995); Ulrich et at, Proc. Natl. Acad. Sci. USA 92: 11907-1 1911 (1995); Walker & Gilbert, J. Biol. Chem 269: 28487-28493 (1994)) and bacterial capsular polysaccharide layers present a diffusion barrier that restricts such systems to small molecule ligands (Roberts, Annu. Rev. Microbiol. 50: 285-315 (1996)). E. coli possesses a lipopolysacchari.de layer or capsule that may interfere sterically with macromolecular binding reactions. In fact, a presumed physiological function of the bacterial capsule is restriction of macromolecular diffusion to the cell membrane, in order to shield the cell from the immune system (DiRienzo et al., Ann. Rev. Blochem. 47: 481-532, (1978)). Since the periplasm of E. coli has not evolved as a compartment for the folding and assembly of antibody fragments, expression of antibodies in E. coli has typically been very clone dependent, with some clones expressing well and others not at all. Such variability introduces concerns about equivalent representation of all possible sequences in an antibody library expressed on the surface of E. coli. Moreover, phage display does not allow some important posttranslational modifications such as glycosylation that can affect specificity or affinity of the antibody. About a third of circulating monoclonal antibodies contain one or more N-linked glycans in the variable regions. In some cases it is believed that these N-glycans in the variable region may play a significant role in antibody function. The discovery of novel therapeutics would be facilitated by the development of alternative selection systems that relied upon eukaryotic cells, such as yeast cells. The structural similarities between B-cells displaying antibodies and yeast cells displaying antibodies provide a closer analogy to in vivo affinity maturation than is available with filamentous phage. Moreover, the ease of growth culture and facility of genetic manipulation available with yeast will enable large populations to be mutagenized and screened rapidly. By contrast with conditions in the mammalian body, the physicochemical conditions of binding and selection can be altered for a yeast culture within a broad range of pH, temperature, and ionic strength to provide additional degrees of freedom in antibody engineering experiments. The development of yeast surface display system for screening combinatorial protein libraries has been described. U.S. Patent Nos. 6,300,065 and 6,699,658 describe the development of a yeast surface display system for screening combinatorial antibody libraries and a screen based on antibody- antigen dissociation kinetics. The system relies on transforming yeast with vectors that express an antibody or antibody fragment fused to a yeast cell surface anchoring protein, using mutagenesis to produce a variegated population of mutants of the antibody or antibody fragment and then screening and selecting those cells that produce the antibody or antibody fragment with the desired enhanced phenotypic properties. U.S. Patent No. 7,132,273 discloses various yeast cell wall anchor proteins and a surface expression system that uses them to immobilize foreign enzymes or polypeptides on the cell wall.

U.S. Published Application No. 2005/0142562 discloses Compositions, kits and methods are provided for generating highly diverse libraries of proteins such as antibodies via homologous recombination in vivo, and screening these libraries against protein, peptide and nucleic acid targets using a two-hybrid method in yeast. The method for screening a library of tester proteins against a target protein or peptide comprises expressing a library of tester proteins in yeast cells, each tester protein being a fusion protein comprised of a first polypeptide subunit whose sequence varies within the library, a second polypeptide subunit whose sequence varies within the library independently of the first polypeptide, and a linker peptide which links the first and second polypeptide subunits; expressing one or more target fusion proteins in the yeast cells expressing the tester proteins, each of the target fusion proteins comprising a target peptide or protein; and selecting those yeast cells in which a reporter gene is expressed, the expression of the reporter gene being activated by binding of the tester fusion protein to the target fusion protein. Of interest are Tanino et al, Biotechnol. Prog. 22: 989-993 (2006), which discloses construction of a Pichiα pαstoris cell surface display system using Flo Ip anchor system; Ren et al., Molec. Biotechnol, 35:103-108 (2007), which discloses the display of adenoregulin in a Pichiα pαstoris cell surface display system using the Flo Ip anchor system; Mergler et al., Appl. Microbiol. Biotechnol. 63:418-421 (2004), which discloses display of K. lαctis yellow enzyme fused to the C-terminus half of S. cerevisiαe α-agglutinin; Jacobs et al., Abstract T23, Pichia

Protein expression Conference, San Diego, CA (October 8-1 1, 2006), which discloses display of proteins on the surface of Pichiα pαstoris using α-agglutinin; Ryckaert et al., Abstracts BVBMB Meeting, Vrije Universiteit Brussel, Belgium (December 2, 2005), which discloses using a yeast display system to identify proteins that bind particular lectins; U.S. Patent No. 7,166,423, which discloses a method for identifying cells based on the product secreted by the cells by coupling to the cell surface a capture moiety that binds the secreted product, which can then be identified using a detection means; U.S. Published Application No. 2004/0219611, which discloses a biotin-avidin system for attaching protein A or G to the surface of a cell for identifying cells that express particular antibodies; U.S. Patent No. 6,919,183, which discloses a method for identifying cells that express a particular protein by expressing in the cell a surface capture moiety and the protein wherein the capture moiety and the protein form a complex which is displayed on the surface of the cell; U.S. Patent No. 6,114,147, which discloses a method for immobilizing proteins on the surface of a yeast or fungal using a fusion protein consisting of a binding protein fused to a cell surface anchoring protein which is expressed in the cell. The potential applications of engineering antibodies for the diagnosis and treatment of human disease such as cancer therapy, tumor imaging, sepsis are far-reaching. For these applications, antibodies with high affinity (i.e., K d≤ 10 nM) and high specificity are highly desirable. Anecdotal evidence, as well as the a priori considerations discussed previously, suggest that phage display or bacterial display systems are unlikely to consistently produce antibodies of sub-nanomolar affinity. To date, yeast display will fill this gap and as such should be a key technology of tremendous commercial and medical significance.

Development of further protein expression systems for yeasts and filamentous fungi, such as Pichiα pαstoris, based on improved vectors and host cell lines in which effective protein display facilitates development of genetically enhanced yeast strains for recombinant production of proteins, and in particular, for recombinant production of monoclonal antibodies, is a desirable objective. BRIEF SUMMARY OF THE INVENTION

The present invention provides a method for expressing and displaying proteins on the surface of a lower eukaryote in a form that is accessible for detection. Combining this method with øuorescence-activated cell sorting (FACS) provides a means for selecting cells that express proteins with increased or decreased affinity for another molecule, catalytic activity, altered specificity, or conditional binding. The method is particularly useful for constructing and screening antibody libraries in lower eukaryotes such as yeast or filamentous fungi.

In a further aspect, there is provided a method of genetic fusion of a lower eukaryote cell surface anchoring protein to a first binding moiety and genetic fusion of a polypeptide of interest to a second binding moiety that is capable of pairwise binding to the first binding moiety.

Nucleic acids comprising the genetic fusions are transformed into host cells. Expression of the genetic fusions provides a cell surface anchoring protein wherein the first binding moiety binds the second binding moiety and at the cell surface presents the protein of interest. When the protein of interest is an antibody, one effectively can mimic the cell surface display of antibodies by B cells in the immune system.

In further aspects, the first and second binding moieties are adapter peptide are derived from the sequences of homodimeric or heterodimeric proteins that are involved in the formation of stable protein complexes. In particular embodiments, these peptides are coiled coil peptides that are capable of forming a dimeric complex such as the coiled coil peptides comprising the GABAB-R1/GABA-R2 receptors.

In one aspect of the present invention, provided is a method for selecting proteins with desirable binding properties comprising: transforming lower eukaryote host cells with nucleic acids expressing a host cell wall binding protein fused at its N- or C-terminus to a first adapter peptide and a protein to be tested fused at its C-terminus to a second adapter peptide capable of pairwise binding to the first adapter peptide; labeling the host cells with a first label, wherein the first label associates with or binds to host cells expressing the protein to be tested and does not associate with or bind to host cells which do not express the protein to be tested; selecting for the host cells with which the first label is associated; and quantitating said first label, wherein a high occurrence of the first label indicates the protein to be tested has desirable binding properties and wherein a low occurrence of the first label indicates the protein to be tested does not have desirable binding properties.

A further embodiment of the present invention further includes the steps of: labeling the host cells with a second label, wherein the second label associates with or binds to host cells expressing an epitope tag fused to the protein to be tested and encoded by the nucleic acid and does not associate with or bind to host cells which do not express the epitope tag encoded by the nucleic acid; quantitating the second label, wherein an occurrence of the second label indicates a number of expressed copies of the epitope-tagged protein to be tested on the host cell surface: and comparing said quantitation of the first label to the quantitation of the second label to determine the occurrence of the first label normalized for the occurrence of the second label, wherein a high occurrence of the first label relative to the occurrence of the second label indicates the protein to be tested has desirable binding properties. Another further embodiment of the present invention includes the steps of: labeling the host cells with a third label that competes with the first label for binding to the protein to be tested; labeling the yeast cells with the first label; quantitating said first label; labeling the host cells with the second label; quantitating the second label; and comparing the quantitation of the first label to the quantitation of the second label to determine the occurrence of the first label normalized for the occurrence of the second label, wherein a low occurrence of the first label relative to the occurrence of the second label indicates the protein to be tested has desirable binding properties.

In one embodiment of the present invention, the first label is a fluorescent label attached to a ligand and the second label is a fluorescent label attached to an antibody. When the labels are fluorescein the quantitation step is performed by flow cytometry or confocal fluorescence microscopy.

Another aspect of the present invention provides vectors for performing the method of the present invention, a vector comprising a nucleic acid encoding a cell wall binding protein fused to a first adapter peptide and a vector comprising a nucleic acid encoding a protein of interest fused at its C-terminus to a second adapter peptide capable of pairwise binding to the first adapter peptide. Further embodiments of this aspect include means for expressing a polypeptide epitope tag fused to the protein of interest in the host cells. Further still embodiments provide that the cell wall binding protein is GPΪ-anchored cell surface anchoring protein, in particular embodiments, a SEDl protein. Further provided is a method for selecting antibodies and fragments thereof with desirable binding properties, performed as described above using a vector in which a single stop codon is place between the nucleic acid encoding the antibody sequence and the nucleic acid encoding the second adapter peptide. The vector is transformed into lower eukaryote host cells comprising nucleic acids expressing a host cell wall binding protein fused at its N- or C-terminus to a first adapter peptide that is capable of pairwise binding to the second adapter peptide.

Translation of mRN A transcribed from the vector is performed under conditions that increases translational readthrough through the stop codon thereby resulting in the production of antibodies that are fused to the second adapter. Labeling the host cells with a first label, wherein the first label associates with or binds to host cells expressing the desired antibodies and does not associate with or bind to host cells which do not express the desired antibodies enables identification and selection of those host cells that produce the desired antibodies. After the host cells that produce the desired antibodies have been selected and isolated, the host cells are grown under conditions that do result in an increase in translational readthrough through the stop codon. Under the second conditions, the host cells produce antibodies or fragments thereof that are not fused to the second adapter peptide.

Therefore, in particular aspects, provided is a method for selecting proteins for displayability on a lower eukaryote host cell surface, comprising: (a)providing a host cell that expresses a capture moiety comprising a cell surface anchoring protein fused to a first binding moiety; (b) transforming the host cell with a nucleic acid encoding a protein fused to a second binding moiety that is capable of specifically interacting with the first binding moiety fused to the cell surface anchoring protein, wherein mutagenesis is used to generate a plurality of host cells encoding a variegated population of mutants of the proteins; (c) contacting the plurality of host cells with a detection means that specifically binds to proteins that are displayed on the surface of the host cell and does not bind to proteins that are not displayed on the surface of the host cell; and (d) isolating the host cells with which the detection means is bound, wherein the presence of the detection means bound to a protein on the surface of the host cells indicates the protein is displayable on the lower eukaryote cell surface.

In a further aspect, provided is a method for selecting a recombinant lower eukaryote host cell that displays a desired protein on the surface of the host cell, comprising: (a) providing host cells that expresses a capture moiety comprising a cell surface anchoring protein fused to a first binding moiety; (b) transforming the host cells with nucleic acids encoding proteins, each fused to a second binding moiety that is capable of specifically interacting with the first binding moiety fused to the cell surface anchoring protein, to produce a plurality of host cells wherein at least one host cell is suspected of displaying the desired protein on the cell surface; (c) contacting the transformed host cells with a detection means that specifically binds to the desired proteins that are displayed on the cell surface; and (d) isolating the host cells with which the detection means is bound to select the host cell that displays the desired protein.

In a further aspect, provided is a method for producing an antibody comprising: (a) providing a host cell that expresses a capture moiety comprising a cell surface anchoring protein fused to a first binding moiety; (b) transforming the host cell with a nucleic acid encoding the heavy and light chains of an antibody wherein the heavy chain (HC) is fused to a second binding moiety that is capable of specifically interacting with the first binding moiety fused to the cell surface anchoring protein, wherein mutagenesis is used to generate a plurality of host cells encoding a variegated population of mutants of the antibodies; (c) contacting the plurality of host cells with a detection means that specifically binds to antibodies that are displayed on the surface of the host cell and does not bind to antibodies that are not displayed on the surface of the host cell; and (d) isolating the host cells with which the detection means is bound, wherein the presence of the detection means bound to an antibody on the surface of the host cell indicates the host cell produces the antibody. In a fUrther aspect, provided is a method for selecting a recombinant lower eiikaryote host cell that displays a desired antibody on the surface of the host cell, comprising: (a) providing host cells that expresses a capture moiety comprising a cell surface anchoring protein fused to a first binding moiety; (b) transforming the host cell with nucleic acids encoding the heavy chain and light chain (LC) of antibodies wherein the heavy chains are fused to a second binding moiety that is capable of specifically interacting with the first binding moiety fused to the cell surface anchoring protein, to produce a plurality of host cells wherein at least one host cell is suspected of displaying the desired antibody on the cell surface; (c) contacting the transformed host cells with a detection means that specifically binds to the desired antibody that is displayed on the cell surface; and (d) isolating the host cell with which the detection means is bound to select the host cell that displays the desired antibody.

Further provided is a method of producing a member of a specific binding pair, wherein the specific binding pair member is an antibody or antibody fragment, comprising an antibody VH domain and an antibody VL domain, and having an antigen binding site with binding specificity for an antigen of interest, the method comprising (a) providing a library of lower eukaryote host cells displaying on their surface a specific binding pair member, which specific binding pair member is an antibody or antibody fragment comprising a synthetic human antibody VH domain and a human antibody VL domain, wherein the library is created by: (i) providing lower eukaryote host cells that express a capture moiety comprising a cell surface anchoring protein fused to a first binding moiety; (ii) providing a library of nucleic acid sequences encoding a genetically diverse population of the specific binding pair member, wherein the VH domains of the genetically diverse population of the specific binding pair member are biased for one or more VH gene families and wherein the specific binding pair member includes a second binding moiety that is capable of specifically interacting with the first binding moiety fused to the cell surface anchoring protein; (iii) expressing the library of nucleic acid sequences in the lower eukaryote host cells, whereby each specific binding pair member is displayed at the surface of a lower eukaryote host cell; (b) selecting one or more specific binding pair members having a binding specificity for the antigen of interest, by binding the one or more specific binding pair members with the antigen of interest, each thus selected specific binding pair member being displayed on the lower eukaryote host cell.

The further aspects, the specific binding pair member comprises a synthetic human antibody VH domain and a synthetic human antibody VL domain and wherein the synthetic human antibody VH domain and the synthetic human antibody VL domain comprise framework regions and hypervariable loops, wherein the framework regions and first two hypervariable loops of both the VH domain and VL domain are essentially human germ line, and wherein the VH domain and VL domain have altered CDR3 loops. In further still aspects in addition to having altered CDR3 loops, the human synthetic antibody VH and VL domains contain mutations in other CDR loops. In further aspects, each human synthetic antibody VH domain CDR loop is of random sequence. In further still aspects, the human synthetic antibody VH domain CDR loops are of known canonical structures and incorporate random sequence elements. The binding pari member can be a full-sized or whole antibody or a fragment such as a single-chain Fv antibody fragment.

In further aspects of any one of the aforementioned methods, the first binding moiety is a first adapter peptide and the second binding moiety is a second adapter peptide wherein the first and second adapter peptides are capable of a specific pairwise interaction. In particular aspects, the first and second adapter peptides are coiled coil peptides that capable of the specific pairwise interaction. In further aspects, the coiled coil peptides are GABAB-Rl and GABAB-R2 subunits that are capable of the specific pairwise interaction.

In further aspects of any one of the aforementioned methods, the cell surface anchoring protein is a GPI protein, for example, a GPI protein selected from the group consisting of α- agglutinin, Cwplp, Cwp2p_? Gaslp, Yap3p, Flolp, Crh2p, Pirlp, Pir4p, Sedlp, Tiplp, Wpip, Hpwplp, Als3p, and Rbt5p. In particular aspects, the cell surface anchoring protein is Sedlp.

In further still aspects, the lower eukaryote is a yeast, including but not limited to, Pichia pastoris .

In further aspects of any one of the aforementioned methods, expression of the nucleic acids encoding the capture moiety and proteins is constitutive or the expression of the nucleic acids encoding the capture moiety and the proteins is induced simultaneously, or the expression of the nucleic acids encoding the capture moiety and the mutants of the proteins are induced sequentially.

In further aspects of any one of the aforementioned methods, Oglycosylation of glycoproteins in the host cell is controlled. That is, O-glycan occupancy and mannose chain length are reduced. In lower eukaryote host cells such as yeast, O-glycosylation can be controlled by deleting the genes encoding one or more protein O-mannosyltransferases (DoI-P- Man:Protein (Ser/Thr) Mannosyl Transferase genes) (PMTs) or by growing the host in a medium containing one or more Pmtp inhibitors. In further aspects, the host cell includes a deletion of one or more of the genes encoding PMTs and the host cell is cultivated in a medium that includes one or more Pmtp inhibitors. Pmtp inhibitors include but are not limited to a benzylidene thiazolidinedione. Examples of benzylidene thiazolidinediones that can be used are 5-[[3,4- bis(phenylmethoxy) phenyl] methylene] -4-oxo-2-thioxo-3-thiazolidineacetic Acid; 5-[[3-(l- Phenylethoxy)-4-(2-phenylethoxy)]phenyl]methylene]-4-oxo-2~thioxo-3-thiazoIidineacetic Acid; and 5-[[3-(l-Phenyl-2-hydroxy)ethoxy)-4-(2-phenylethoxy)]phenyl]methylene]-4-oxo-2-thioxo- 3-thiazolidineacetic Acid. In further still aspects, the host cell further includes a nucleic acid that encodes an alpha- 1 ,2-mannosidase that has a signal peptide that directs it for secretion. In another aspect, genes encoding one or emore endogenous mannosyltransferase enzymes are deleted. This deletion(s) can be in combination with providing the secreted alpha- 1,2-mannosidase and/or PMT inhibitors or can be in lieu of providing the secreted alpha- 1,2-mannosidase and/or PMT inhibitors.

In further aspects of any one of the aforementioned methods, host cells further include lower eukaryote cells (e.g., yeast such as Pichia pastoris) that are genetically engineered to eliminate glycoproteins having α-mannosidase-resistant iV-glycans by deleting or disrupting one or more of the β-mannosyltransferase genes (e.g., BMTl, BMT2, BMT3, and BMT4)(See, U.S. Published Patent Application No. 2006/0211085) or abrogating translation of RNAs encoding one or more of the β-mannosyltransferases using interfering RNA₃ antisense RNA, or the like. In further aspects of any one of the methods herein, the host cells can further include lower eukaryote cells (e.g., yeast such as Pichia pastoris) that are genetically engineered to eliminate glycoproteins having phosphomannose residues by deleting or disrupting one or both of the phosphomannosyl transferase genes PNOl and MNN4B {See for example, U.S. Patent Nos. 7,198,921 and 7,259,007), which in further aspects can also include deleting or disrupting the MNN 4 A gene or abrogating translation of RNAs encoding one or more of the phosphomannosyltransferases using interfering RNA₅ antisense RNA, or the like.

In further still aspects, the host cell has been genetically modified to produce glycoproteins that have predominantly an iV-glycan selected from the group consisting of complex JV-glycans, hybrid JV-glycans, and high mannose iV-glycans wherein complex iV-glycans are selected from the group consisting of Man3GlcNAc2, GIcNAC(I _4)Man3GlcNAc2, GaI(I . 4)GlcNAc(i-4)Man3GIcNAc2, and NANA(]_4)Gal(i-4)Mari3GlcNAc2; hybrid JV-glycans are selected from the group consisting of Mans GIcN Ac2, GlcNAcMan5GlcNAc2, GalGlcNAcMansGlcNAc2_} and NANAGalGlcNAcMan5GlcNAc2; and high Mannose JV-glycans are selected from the group consisting of MangGlcNAc2, ManyGlcNAc2, MangGlcNAc2, and Manα.GlcNAc2-

DEFINITIONS

Unless otherwise defined herein, scientific and technical terms and phrases used in connection with the present invention shall have the meanings that are commonly understood by those of ordinary skill in the art. Further, unless otherwise required by context, singular terms shall include the plural and plural terms shall include the singular. Generally, nomenclatures used in connection with, and techniques of biochemistry, enzymology, molecular and cellular biology, microbiology, genetics and protein and nucleic acid chemistry and hybridization described herein are those well known and commonly used in the art. The methods and techniques of the present invention are generally performed according to conventional methods well known in the art and as described in various general and more specific references that are cited and discussed throughout the present specification unless otherwise indicated. See, e.g., Sambrook et al. Molecular Cloning: A Laboratory Manual, 2d ed., Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N. Y. (1989); Ausubel et al, Current Protocols in Molecular Biology, Greene Publishing Associates (1992, and Supplements to 2002); Harlow and Lane, Antibodies: A Laboratory Manual, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N. Y. (1990); Taylor and Drickamer, Introduction to Glycobiology, Oxford Univ. Press (2003); Worthington Enzyme Manual, Worthington Biochemical Corp., Freehold, NJ; Handbook of Biochemistry: Section A Proteins, VoI 1, CRC Press (1976); Handbook of Biochemistry: Section A Proteins, VoI II, CRC Press (1976); Essentials of Glycobiology, Cold Spring Harbor Laboratory Press (1999).

All publications, patents and other references mentioned herein are hereby incorporated by reference in their entireties.

The following terms, unless otherwise indicated, shall be understood to have the following meanings:

As used herein, the terms "N-glycan"and "glycoform" are used interchangeably and refer to an N-linked oligosaccharide, e.g., one that is attached by an asparagine-N- acetylglucosamine linkage to an asparagine residue of a polypeptide. N-linked glycoproteins contain an N-acetylglucosamine residue linked to the amide nitrogen of an asparagine residue in the protein. The predominant sugars found on glycoproteins are glucose, galactose, mannose, fucose, N-acetylgalactosamine (GalΝAc), N-acetylglucosamine (GlcΝAc) and sialic acid (e.g., N- acetyl-neuraminic acid (NANA)). The processing of the sugar groups occurs cotranslationally in the lumen of the ER and continues in the Golgi apparatus for N-linked glycoproteins.

N-glycans have a common pentasaccharide core of Man3GlcΝAc2 ("Man" refers to mannose; "GIc" refers to glucose; and "NAc" refers to N-acetyl; GIcNAc refers to N- acetylglucosamine). N-glycans differ with respect to the number of branches (antennae) comprising peripheral sugars (e.g., GlcΝAc, galactose, fucose and sialic acid) that are added to the Man3GlcΝAc2 ("Man3") core structure which is also referred to as the "triammnose core", the "pentasaccharide core" or the "paucimannose core". N-glycans are classified according to their branched constituents (e.g., high mannose, complex or hybrid). A "high mannose" type N- glycan has five or more mannose residues. A "complex" type Ν-glycan typically has at least one GlcΝAc attached to the 1,3 mannose arm and at least one GlcΝAc attached to the 1,6 mannose ami of a "trimannose" core. Complex N-glycans may also have galactose ("Gal") or N- acetylgalactosamine ("GalΝAc") residues that are optionally modified with sialic acid or derivatives (e.g., "NANA" or "NeuAc", where "Neu" refers to neuraminic acid and "Ac" refers to acetyl). Complex N-glycans may also have intrachain substitutions comprising "bisecting" GlcΝAc and core fucose ("Fuc"). Complex N-glycans may also have multiple antennae on the "trimannose core," often referred to as "multiple antennary glycans." A "hybrid" N-glycan has at least one GlcΝAc on the terminal of the 1,3 mannose arm of the trimannose core and zero or more mamoses on the 1,6 mannose arm of the trimannose core. The various iV-glycans are also referred to as "glycoforms."

Abbreviations used herein are of common usage in the art, see, e.g., abbreviations of sugars, above. Other common abbreviations include "PNGase", or "glycanase" or "glucosidase" which all refer to peptide N-glycosidase F (EC 3.2.2.18).

The term "operably linked" expression control sequences refers to a linkage in which the expression control sequence is contiguous with the gene of interest to control the gene of interest, as well as expression control sequences that act in trans or at a distance to control the gene of interest. The term "expression control sequence" or "regulatory sequences" are used interchangeably and as used herein refer to polynucleotide sequences which are necessary to affect the expression of coding sequences to which they are operably linked. Expression control sequences are sequences which control the transcription, post-transcriptional events and translation of nucleic acid sequences. Expression control sequences include appropriate transcription initiation, termination, promoter and enhancer sequences; efficient RNA processing signals such as splicing and polyadenylation signals; sequences that stabilize cytoplasmic mRNA; sequences that enhance translation efficiency (e.g., ribosome binding sites); sequences that enhance protein stability; and when desired, sequences that enhance protein secretion. The nature of such control sequences differs depending upon the host organism; in prokaryotes, such control sequences generally include promoter, ribosomal binding site, and transcription termination sequence. The term "control sequences" is intended to include, at a minimum, all components whose presence is essential for expression, and can also include additional components whose presence is advantageous, for example, leader sequences and fusion partner sequences. The term "recombinant host cell" ("expression host cell", "expression host system",

"expression system" or simply "host cell"), as used herein, is intended to refer to a cell into which a recombinant vector has been introduced. It should be understood that such terms are intended to refer not only to the particular subject cell but to the progeny of such a cell. Because certain modifications may occur in succeeding generations due to either mutation or environmental influences, such progeny may not, in fact, be identical to the parent cell, but are still included within the scope of the term "host cell" as used herein. A recombinant host cell may be an isolated cell or cell line grown in culture or may be a cell which resides in a living tissue or organism.

The term "eukaryotic" refers to a nucleated cell or organism, and includes insect cells, plant cells, mammalian cells, animal cells and lower eukaryotic cells.

The term "lower eukaryotic cells" includes yeast and filamentous fungi. Yeast and filamentous fungi include, but are not limited to Pichia pastoris, Pichiafinlandica, Pichia trehalophila, Pichia koclamae, Pichia membranaefaciens, Pichia minuta (Ogataea minuta, Pichia lϊndnerϊ), Pichia opuntiae, Pichia thermotolerans, Pichia salictaria, Pichia guercuum, Pichia pijperi, Pichia stipiis, Pichia methanolica, Pichia sp., Saccharomyces cerevisiae, Saccharomyces sp., Hansenula polymorpha, Kluyveromyces sp,, Kluyveromyces lactis, Candida albicans, Aspergillus nidulans, Aspergillus niger, Aspergillus oryzae, Trichoderma reesei, Chrysosporium lucknowense, Fusarium sp., Fusarium gramineum, Fusarium venenatum, Physcomitrella patens and Neurospora crassa. Pichia sp., any Saccharomyces sp., Hansenula polymorpha, any Kluyveromyces sp., Candida albicans, any Aspergillus sp., Trichoderma reesei, Chrysosporium lucknowense, any Fusarium sp. and Neurospora crassa. The term "specific binding pair" refers to a pair of molecules (each being a member of a specific binding pair) which are naturally derived or synthetically produced. One of the pair of molecules, has an area on its surface, or a cavity which specifically binds to, and is therefore, defined as complementary with a particular spatial and polar organisation of the other molecule, so that the pair have the property of binding specifically to each other. Examples of types of specific binding pairs are antigen-antibody, biotin-avidin, hormone-hormone receptor, receptor- ligand, enzyme-substrate, IgG-protein A.

As used herein, the terms "antibody," "immunoglobulin," "immunoglobulins" and "immunoglobulin molecule" are used interchangeably. Each immunoglobulin molecule has a unique structure that allows it to bind its specific antigen, but all immunoglobulins have the same overall structure as described herein. The basic immunoglobulin structural unit is known to comprise a tetramer of subunits. Each tetramer has two identical pairs of polypeptide chains, each pair having one "light" chain (LC) (about 25 kDa) and one "heavy" chain (HC) (about 50-70 kDa). The amino-terminal portion of each chain includes a variable region of about 100 to 110 or more amino acids primarily responsible for antigen recognition. The carboxy-terminal portion of each chain defines a constant region primarily responsible for effector function. Light chains (LCs) are classified as either kappa or lambda. Heavy chains (HCs) are classified as gamma, mu, alpha, delta, or epsilon, and define the antibody's isotype as IgG, IgM, IgA, IgD and IgE, respectively.

The light and heavy chains are subdivided into variable regions and constant regions (See generally, Fundamental Immunology (Paul, W., ed., 2nd ed. Raven Press, N.Y., 1989), Ch. 7. The variable regions of each light/heavy chain pair form the antibody binding site. Thus, an intact antibody has two binding sites. Except in bifunctional or bispecific antibodies, the two binding sites are the same. The chains all exhibit the same general structure of relatively conserved framework regions (FR) joined by three hypervariable regions, also called complementarity determining regions or CDRs. The CDRs from the two chains of each pair are aligned by the framework regions, enabling binding to a specific epitope. The terms include naturally occurring forms, as well as fragments and derivatives. Included within the scope of the term are classes of immunoglobulins (Igs), namely, IgG, IgA, IgE, IgM, and IgD. Also included within the scope of the terms are the subtypes of IgGs, namely, IgGl, IgG2, IgG3 and IgG4. The term is used in the broadest sense and includes single monoclonal antibodies (including agonist and antagonist antibodies) as well as antibody compositions which will bind to multiple epitopes or antigens. The terms specifically cover monoclonal antibodies (including full length monoclonal antibodies), polyclonal antibodies, multispecific antibodies (for example, bispecifϊc antibodies), and antibody fragments so long as they contain or are modified to contain at least the portion of the CH2 domain of the heavy chain immunoglobulin constant region which comprises an N-linked glycosylation site of the CH2 domain, or a variant thereof. Included within the terms are molecules comprising only the Fc region, such as immunoadhesins (U.S. Published Patent Application No. 20040136986), Fc fusions, and antibody-like molecules. Alternatively, these terms can refer to an antibody fragment of at least the Fab region that at least contains an N- linked glycosylation site.

The term "Fc" fragment refers to the 'fragment crystallized' C-terminal region of the antibody containing the CH2 and CR3 domains (Figure 1). The term "Fab" fragment refers to the 'fragment antigen binding' region of the antibody containing the VH, CH I₅ VL and CL domains (See Figure 1).

The term "monoclonal antibody" (mAb) as used herein refers to an antibody obtained from a population of substantially homogeneous antibodies, i.e., the individual antibodies comprising the population are identical except for possible naturally occurring mutations that may be present in minor amounts. Monoclonal antibodies are highly specific, being directed against a single antigenic site. Furthermore, in contrast to conventional (polyclonal) antibody preparations which typically include different antibodies directed against different determinants (epitopes), each mAb is directed against a single determinant on the antigen. In addition to their specificity, monoclonal antibodies are advantageous in that they can be synthesized by hybridoma culture, uncontaminated by other immunoglobulins. The term "monoclonal" indicates the character of the antibody as being obtained from a substantially homogeneous population of antibodies, and is not to be construed as requiring production of the antibody by any particular method. For example, the monoclonal antibodies to be used in accordance with the present invention may be made by the hybridoma method first described by Kohler et ai, (1975) Nature, 256:495, or may be made by recombinant DNA methods (See, for example, U.S. Pat. No. 4,816,567 to Cabilly et ai).

The term "fragments" within the scope of the terms "antibody" or "immunoglobulin" include those produced by digestion with various proteases, those produced by chemical cleavage and/or chemical dissociation and those produced recombinantly, so long as the fragment remains capable of specific binding to a target molecule. Among such fragments are Fc, Fab, Fab', Fv, F(ab')2, and single chain Fv (scFv) fragments. Hereinafter, the term "immunoglobulin" also includes the term "fragments" as well.

Immunoglobulins further include immunoglobulins or fragments that have been modified in sequence but remain capable of specific binding to a target molecule, including: interspecies chimeric and humanized antibodies; antibody fusions; heteromeric antibody complexes and antibody fusions, such as diabodies (bispecific antibodies), single-chain diabodies, and intrabodies (See, for example, Intracellular Antibodies: Research and Disease Applications, (Marasco, ed., Springer- Verlag New York, Inc., 1998).

The term "catalytic antibody" refers to immunoglobulin molecules that are capable of catalyzing a biochemical reaction. Catalytic antibodies are well known in the art and have been described in U.S. Patent Application Nos. 7205136; 4888281; 5037750 to Schochetman et al, U.S. Patent Application Nos. 5733757; 5985626; and 6368839 to Barbae, III et al.

As used herein, the term "consisting essentially of will be understood to imply the inclusion of a stated integer or group of integers; while excluding modifications or other integers which would materially affect or alter the stated integer. With respect to species of N-glycans, the term "consisting essentially of a stated N-glycan will be understood to include the N-glycan whether or not that N-glycan is fucosylated at the N-acetylglucosamine (GIcNAc) which is directly linked to the asparagine residue of the glycoprotein.

As used herein, the term "predominantly" or variations such as "the predominant" or "which is predominant" will be understood to mean the glycan species that has the highest mole percent (%) of total neutral N-glycans after the glycoprotein has been treated with PNGase and released glycans analyzed by mass spectroscopy, for example, MALDI-TOF MS or HPLC. In other words, the phrase "predominantly" is defined as an individual entity, such as a specific glycoform, is present in greater mole percent than any other individual entity. For example, if a composition consists of species A in 40 mole percent, species B in 35 mole percent and species C in 25 mole percent, the composition comprises predominantly species A, and species B would be the next most predominant species. Some host cells may produce compositions comprising neutral N-glycans and charged N-glycans such as mannosylphosphate. Therefore, a composition of glycoproteins can include a plurality of charged and uncharged or neutral N-glycans. In the present invention, it is within the context of the total plurality of neutral N-glycans in the composition in which the predominant N-glycan determined. Thus, as used herein, "predominant N-glycan" means that of the total plurality of neutral N-glycans in the composition, the predominant N-glycan is of a particular structure.

As used herein, the term "essentially free of a particular sugar residue, such as fucose, or galactose and the like, is used to indicate that the glycoprotein composition is substantially devoid of N-glycans which contain such residues. Expressed in terms of purity, essentially free means that the amount of N-glycan structures containing such sugar residues does not exceed 10%, and preferably is below 5%, more preferably below 1%, most preferably below 0.5%, wherein the percentages are by weight or by mole percent. Thus, substantially all of the N- glycan structures in a glycoprotein composition according to the present invention are free of fucose, or galactose, or both. As used herein, a glycoprotein composition "lacks" or "is lacking" a particular sugar residue, such as fucose or galactose, when no detectable amount of such sugar residue is present on the N-glycan structures at any time. For example, in preferred embodiments of the present invention, the glycoprotein compositions are produced by lower eukaryotic organisms, as defined above, including yeast (for example, Pichia sp.; Saccharomyces sp.; Kluyveromyces sp.; Aspergillus sp.)_; and will "lack fucose," because the cells of these organisms do not have the enzymes needed to produce fucosylated N-glycan structures. Thus, the term "essentially free of fucose" encompasses the term "lacking fucose." However, a composition may be "essentially free of fucose" even if the composition at one time contained fucosylated N-glycan structures or contains limited, but detectable amounts of fucosylated N-glycan structures as described above. The interaction of antibodies and antibody-antigen complexes with cells of the immune system and the variety of responses, including antibody-dependent cell-mediated cytotoxicity (ADCC) and complement-dependent cytotoxicity (CDC), clearance of immunocomplexes (phagocytosis), antibody production by B cells and IgG serum half-life are defined respectively in the following: Daeron et al, 1997, Annu. Rev. Immunol. 15: 203-234; Ward and Ghetie, 1995, Therapeutic Immunol. 2:77-94; Cox and Greenberg, 2001, Semin. Immunol 13: 339-345;

Heyman, 2003, Immunol. Lett. 88:157-161; and Ravetch, 1997, Curr. Opin. Immunol. 9: 121- 125.

BRIEF DESCRIPTION OF THE DRAWINGS Figure 1 illustrates a yeast cell engineered to express a potential cell wall anchor - coiled coil peptide fusion protein that dimerizes with a gene of interest - coiled coil peptide fusion. The two fusion protein heterodimers are locked by an artificial disulfide bond.

Figure 2A illustrates the fusion protein expression construct consisting of a signal sequence, GR2 coiled coil peptide, Myc tag, and GPI anchor fusion protein. Figure 2B shows the amino acid sequences of the SEDl fusion protein (SEQ ID NO:20) wherein the alpha amylase signal peptide is underlined, the GR2 coiled coil peptide sequence is in bold-faced type, the myc- tag sequence in italics, and the S, cerevisiae SEDl sequence in normal font.

Figure 3 shows a map of pGLY3033 expressing a fusion protein consisting of a Pichia pastoris signal sequence, a myc tag, the GR2 coiled coil peptide, and the cell wall anchor protein

SED under the control of the AOXl promoter. Figure 4 illustrates three different constructs that can be used to create a monoclonal antibody or antibody fragment fused to coiled coil peptide GRl for displaying the antibody or antibody fragment on the host glycoengineered yeast. Construct A is a Fab fragment display format, Construct B is a full-length antibody display format, and Construct C contains a stop codon between the antibody heavy chain and the coiled coil (GRl). AOXl is the AOXl promoter, SS is signal peptide, Lc is light chain, TT is transcription termination sequence, Fd is heavy chain variable fragment, CH2 and CH3 are heavy chain constant domains, HA is haemaglutinin, His is poly histidine, GRl is coiled coil peptide.

Figure 5 shows a map of Fab fragment display plasmid pGLY3915. Figure 6 shows a map of full-length antibody display plasmid pGLY3941.

Figure 7 illustrates detection of a displayed antibody or antibody fragment on the yeast cell surface by goat anti-human H+L IgGs Alexa 488 or fluorophor conjugated antigen.

Figures 8A-J show fluorescent microscopy photographs of glycoengineered Pichia pastoris YGLY 4102 overexpressing anti-Her2 Fab fragment and transformed with expression plasmids of GR2 coiled coil peptide fused to different GPI cell wall anchor proteins pGLY3015 (CWP2), pGLY3033 (ScSEDl), pGLY3034 (ScSEDl truncation), pGLY3035 (PpSPIl), pGLY3036 (PpGASl), pGLY3037 (ScGASl), pGLY3038 (ScGASl truncation), pGLY3039 (HpTIP) and pGLY3040 (HpTIP truncation).

Figures 9A-C show fluorescent microscope images of glycoengineered Pichia pastoris displaying different Fab fragment: Fig. 9A, Pichia pastoris GS2.0anti-DKKl Fab fragment displaying cell; Fig. 9B, Pichia pastoris GS2.0 anti-Her2 secreted Fab fragment without overexpression of the SEDl GPI-protein cell wall anchor; Fig. 9C, Pichia pastoris GS2.0 anti- Her2 Fab fragment displaying cell. All these cells were labeled with anti-human H&L Alexa 488 and photographed using the same exposure time. Figure 10 shows an overlay of several fluorescence-activated cell sorting (FACS) experiments. Flow cytometry analysis was conducted using fluorescently labeled cells expressing anti -Her2 Fab fragment and α-DKKl Fab fragment. X axis: Fluorescence intensity, Y axis: Number of sorted events. The fluorescence mean of the anti -Her2 Fab fragment is significantly higher than that of the anti -DKKl population. Non-labeled cells are to the left marking the area of background detection.

Figure 11 shows FACS analysis of different population ratios expressing cell-surface displayed anti~Her2 and anti-DKKl, respectively. Only the ratios of anti-Her2: anti-DKKl of 1 : 1 (red) and 1 :10 (green) allow for the detection of two distinct populations. At the higher ratios of 1 : 100 (blue) and 1 : 1000 (brown) no distinct subpopulations can be observed. Figures 12A-C show the distribution of anti-Her2 and anti-DKKl expressing cells. Fig.

12A: Cells were isolated from five areas of decreasing fluorescence (Cl through C5). Fig. 12B shows two population of cells with different fluorescent intensity when we mix anti-Her2 and anti-DKKl as 1:1 ratio. Fig. 12C: Cells were plated and analyzed by colony PCR to determine their identity.

Figure 13 illustrates enrichment of anti-Her2 Fab fragment expressing cells over three rounds of FACS. Figures 14A-F show fluorescence microscopy and FACS analysis of P. pastoris strain

YGLY6724, displaying anti-Her2 full length mAb using a read-through stop codon construct and YGLY6722, displaying anti-Her2 full length mAb. Cells are labeled with goat anti-human H&L Alexa 488. G418 is an antibiotic which increases stop codon read-through. YGLY6732 is a non-displaying non-labeled yeast used to determine the level of background fluorescence. Figures 14A-E shows fluorescence microscopy of samples A-E in the FACS analysis shown in Figure 14F.

Figures 15 A-D show fluorescence of several clones of host cells of the YGLY2696 background co-expressing SED1-GR2 fusion protein and anti-CD20 Fab fragment (heavy chain fused to GRl) and labeled with goat anti-human IgG (H+L)- Alexa 488. YGLY5149 15 second exposure (Fig. 15A), YGLY5152 15 second exposure (Fig. 15B), YGLY6693 30 second exposure (Fig. 15C), and YGLY6694 30 second exposure (Fig. 15D).

Figure 16 shows the genealogy of humanized chaperone strain YGLY2696. Figures 17A-B show that the heavy and light chains of the displayed Fab fragment in the Fab fragment-displaying cells were properly assembled. Cells displaying Fab fragments were labeled with light and heavy chain specific fluorophore-conjugated antibodies. Flow cytometric analysis shows that the displayed Fab fragment heavy chain expression corresponds with displayed light chain expression indicating proper assembly of the Fab fragments. Fig. 17A YGLY7762 cells (1D05); Fig. 17B YGLY7764 cells (1H23).

Figure 18 shows a flow cytometric analysis of antigen- labeled cells. Anti-CD20 and anti-PCSK9 (1D05 and 1H23) Fab fragments were displayed on the yeast surface using the methods described herein. The cells were labeled by fluorophore-conjugated antigen and generic antibody detection. Panel A shows the profile of α-CD20 and anti-PCSK9 (1D05) displayed cells when mixed in a 1 : 1 ratio prior to labeling. In panel B, high affinity (1D05) and low affinity(lH23) anti-PCSK9 Fab fragment expressing cells flow cytometric profiles were overlayed in the same picture.

Figures 19A-B show FACS sorting of mixed Fab fragment displaying cell populations based on antigen affinity. Fig. 19A shows FACS sorting of cells displaying binding Fab fragment (1D05) from cells displaying non-binding Fab fragment (α-CD20) when using fluorophore-labeled PCSK-9 antigen. The cells were mixed at a 1D05: α-CD20 ratio of 1: 1,000; 1 :10_f000; and 1 : 100,000 and then sorted for up to two rounds. Fig. 19B shows FACS sorting of cells displaying high affinity binding Fab fragment (1D05) from cells displaying low affinity Fab fragment (1H23) when using fluorophore-labeled PCSK-9 antigen. The cells were mixed at a 1D05: 1H23 ratio of 1 :10,000 and 1 :100,000 and sorted two rounds.

Figure 20 shows a map of pGLY3958, which encodes light chain Cl kappa and the heavy chain CHl fused to GR2, and which targets the plasmid to the TRP2 locus of P ¹IcMa pastoris, This plasmid was used to make plasmids pGLYS108, pGLY5110, and pGLY5107.

Figure 21 shows a map of plasmid pGLY5108, which encodes the 1D05 anti~PCSK9 Fab fragment.

Figure 22 shows a map of plasmid pGLY5110, which encodes the 1H23 anti-PCSK9 Fab fragment. Figure 23 shows a map of plasmid pGLYSl 07, which encodes the anti-CD20 Genmab antibody.

DETAILED DESCRIPTION OF THE INVENTION

The present invention provides a protein display system that is capable of displaying diverse libraries of proteins on the surface of a eukaryote host cell such as a lower eukaryote host cell (e.g., yeast or filamentous fungal cells). The compositions and methods are particularly useful for the display of collections of proteins in the context of discovery (that is, screening) or molecular evolution protocols. A salient feature of the method is that it provides a display system in which proteins of interest can be displayed on the surface of a host cell without having to express the protein of interest as a fusion protein in which the protein of interest is fused to a surface anchor protein.

In general, provided is a method for selecting proteins for displayability on a lower eukaryote cell surface, comprising providing a host cell that expresses a capture moiety comprising a cell surface anchoring protein fused to a first binding moiety; transforming the host cell with a nucleic acid encoding proteins fused to a second binding moiety that is capable of specifically interacting with the first binding moiety fused to the cell surface anchoring protein, wherein mutagenesis is used to generate a plurality of host cells encoding a variegated population of mutants of the proteins; contacting the plurality of host cells with a detection means that specifically binds to proteins that are displayed on the surface of the host cell and does not bind to proteins that are not displayed on the surface of the host cell; and isolating the host cells with which the detection means is bound, wherein the presence of the detection means bound to a protein on the surface of the host cells indicates the proteins are displayable on the lower eukaryote cell surface.

Further provided is a method for selecting recombinant lower eukaryote host cells that display a desired protein on the surface of the host cells, comprising providing host cells that expresses a capture moiety comprising a cell surface anchoring protein fused to a first binding moiety; transforming the host cells with nucleic acids encoding proteins fused to a second binding moiety that is capable of specifically interacting with the first binding moiety fused to the cell surface anchoring protein to produce a plurality of host cells wherein at least one host cell is suspected of displaying the desired protein on the cell surface; contacting the transformed host cells with a detection means that specifically binds to the desired proteins that are displayed on the cell surface; and isolating the host cells with which the detection means is bound to select the host cells that display the desired protein.

Thus, as shown in Figure 1, the system comprises at least two components. The first component is a helper vector that expresses a capture moiety that in particular embodiments comprises a cell surface anchoring protein that is capable of binding or integrating to the surface of the host cell fused to a first binding moiety, which in particular aspects comprises an adapter peptide that is capable of pairwise binding to the second adapter peptide. The first binding moiety or adapter peptide is located at the end of the cell surface anchoring protein that is exposed to the extracellular environment such that the first adapter peptide is capable of interacting with the second adapter peptide. The second component is a vector that expresses a protein of interest or libraries of which the protein of interest is to be selected (for example, a library of vectors expressing antibodies or fragments thereof). The vector expresses the proteins of interest as fusion proteins in which a second adapter peptide is fused to the N- or C-terminus of the proteins of interest.

Both of the components can be provided in vectors which integrate the nucleic acids into the genome of the host cell by homologous recombination. Homologous recombination can be double crossover or single crossover homologous recombination. RoIHn single crossover homologous recombination has been described in Nett et ah, Yeast 22: 295-304 (2005). Each component can be integrated in the same locus in the genome or in separate loci in the genome. Alternatively, one or both components can be transiently expressed in the host cell. The method enables selection of proteins with desirable binding properties including but not limited to antibodies or fragments thereof (e.g., Fab fragments) of a desired affinity or avidity, enzymes with a particular enzymatic activity or substrate specificity, including catalytic antibodies, receptors with a particular specificity for particular ligands, and fusion proteins including but not limited to those comprising the Fc region of antibody fused to a heterologous protein. In general, the method comprises transforming lower eukaryote host cells with a first nucleic acid expressing a host cell wall binding protein fused at its N- or C-terminus to a first binding moiety such as an adapter peptide capable of pairwise binding to the second adapter peptide and a second nucleic acid expressing a protein to be tested fused at its N- or C-terminus to a second binding moiety such as an adapter peptide capable of pairwise binding to the first adapter peptide. The first and second nucleic acids can be operably linked to the same promoter or to different promoters that are separately inducible. Preferably, the protein of interest is fused to a cellular signal peptide that facilitates shuttle of the fusion protein through the secretory pathway to the cell surface. Expression of first nucleic acids results in the production of the cell wall binding fusion protein, which is transported to the cell surface where it then binds to the surface of the cell with the first binding moiety exposed to the extracellular environment. Expression of the second nucleic acid results in the production of the protein of interest fusion protein, which is transported through the secretory pathway and secreted from the cell. However, as the protein of interest fusion protein is secreted, it is retained on the cell surface because the second binding moiety fused to the protein of interest forms a specific interaction with the first binding moiety fused to the cell wall binding protein.

Further provided is a library method for identifying and selecting cells that produce a particular member of a specific binding pair including but not limited to antibodies and Fab fragments. Therefore, in further aspects, a method of producing a protein that is a member of a specific binding pair, wherein the specific binding pair member is an antibody or antibody fragment, comprising an antibody VH domain and an antibody VL domain, and having an antigen binding site with binding specificity for an antigen of interest. The method comprises providing a library of lower eukaryote host cells displaying on their surface a specific binding pair member, which specific binding pair member is an antibody or antibody fragment comprising a synthetic human antibody VH domain and a human antibody VL domain. The library is created by providing lower eukaryote host cells that express a capture moiety comprising a cell surface anchoring protein fused to a first binding moiety and providing a library of nucleic acid sequences encoding a genetically diverse population of the specific binding pair member, wherein the VH domains of the genetically diverse population of the specific binding pair member are biased for one or more VH gene families and wherein the specific binding pair member includes a second binding moiety that is capable of specifically interacting with the first binding moiety fused to the cell surface anchoring protein. The library of nucleic acid sequences is expressed in the lower eukaryote host cells so that each specific binding pair member is displayed at the surface of a lower eukaryote host cell. Then, cells that produce one or more specific binding pair members having a binding specificity for the antigen of interest are selected by binding the one or more specific binding pair members with the antigen of interest.

The further aspects, the specific binding pair member comprises a synthetic human antibody VH domain and a synthetic human antibody VL domain and wherein the synthetic human antibody VH domain and the synthetic human antibody VL domain comprise framework regions and hypervariable loops, wherein the framework regions and first two hypervariable loops of both the VH domain and VL domain are essentially human germ line, and wherein the VH domain and VL domain have altered CDR3 loops. In further still aspects in addition to having altered CDR3 loops, the human synthetic antibody VH and VL domains contain mutations in other CDR loops. In further aspects, each human synthetic antibody VH domain CDR loop is of random sequence. In further still aspects, the human synthetic antibody VH domain CDR loops are of known canonical structures and incorporate random sequence elements. The binding pair member can be a full-sized or whole antibody or a fragment such as a single-chain Fv antibody fragment.

Detection of host cells that express the desired protein of interest can be achieved by labeling the host cells with a first label, wherein the first label associates with or binds to the protein of interest and does not associate with or bind to host cells which do not express the protein of interest. For example, in the case when the protein of interest is an antibody, the first label can be an antigen that is specifically recognized by the antibody of interest. The host cells with which the first label is associated are selected and the amount of first label associated with the host cell is quantitated. A high occurrence of the first label indicates the protein of interest has desirable binding properties and a low occurrence of the first label indicates the protein of interest does not have desirable binding properties.

A further aspect includes the steps of labeling the above host cells with a second label, wherein the second label associates with or binds to host cells expressing an epitope tag fused to the protein of interest and does not associate with or bind to host cells which do not express the epitope tag. The amount of second label associated with the host cells is quantitated. The amount of the second label associated with the host cell indicates a number of expressed copies of the epitope-tagged protein of interest on the host cell surface and by comparing the quanititation of the first label to the quantitation of the second label enables the amount of the first label normalized for the amount of the second label, wherein a high occurrence of the first label relative to the occurrence of the second label indicates the protein to be tested has desirable binding properties.

Another aspect includes the steps of labeling the above host cells with a third label that competes with the first label for binding to the protein of interest. In this aspect, the host cells are labeled with the first label and the amount of first label associated with host cells is quantitated. Then the host cells are labeled with the second label and the amount of second label associated with host cells is quantitated. Comparing the quantitation of the first label to the quantitation of the second label is performed to determine the occurrence of the first label normalized for the occurrence of the second label, wherein a low occurrence of the first label relative to the occurrence of the second label indicates the protein of interest has desirable binding properties.

In further aspects, the first label is a fluorescent label attached to a ligand specific for the protein of interest and the second label is a fluorescent label attached to an antibody specific for the protein of interest. When the labels are fluorescein the quantitation step is performed by flow cytometry or confocal fluorescence microscopy. In a further still aspect, the first label is a fluorescent label attached to a ligand specific for the protein of interest and fluorescence- activated cell sorting (FACS) is used to separate the host that express the protein of interest from host cells that do not produce the protein of interest.

Further provided is a method for selecting antibodies and fragments thereof with desirable binding properties, performed as described above using a vector in which a single stop codon is place between the nucleic acid encoding the antibody sequence and the nucleic acid encoding the second adapter peptide. The vector is transformed into lower eukaryote host cells comprising nucleic acids expressing a host cell wall binding protein fused at its N- or C-terminus to a first adapter peptide that is capable of pairwise binding to the second adapter peptide. Translation of mRNAs encoded by the vector is performed under conditions that increases translational readthrough through the stop codon thereby producing antibodies that are fused to the second adapter. Labeling the host cells with a first label, wherein the first label associates with or binds to host cells expressing the desired antibodies and does not associate with or bind to host cells which do not express the desired antibodies enables identification and selection of those host cells that produce the desired antibodies. After the host cells that produce the desired antibodies have been selected and isolated, the host cells are grown under conditions that do result in an increase in translational readthrough through the stop codon. Under the second conditions, the host cells produce antibodies or fragments thereof that are not fused to the second adapter peptide.

Figure 4 shows expression cassette C, which had been designed for use in lower eukaryotes such as yeast such that when introduced into a host cell and the host cell is grown under appropriate conditions, the host cell is capable of producing full-length antibodies that include a second adapter peptide fused to the heavy chain for selection of a desired full-length antibody. However, under production conditions, the construct enables production of antibodies in which the heavy chain is not fused to the second adapter peptide. Thus, expression cassette C avoids the need to redone the nucleic acid encoding the antibody to remove the nucleic acϊd that encodes the second adapter peptide. In expression cassette C, the second ORF that encodes a fusion protein comprising the heavy chain fused at the C-terminus to the N-terminus of the second adapter peptide further includes a single stop codon between the end of the nucleic acid sequence encoding the heavy chain and the nucleic acid encoding the second adapter peptide, in which readthrough of the stop codon is inducible. Under most conditions, translation of an mRNA transcribed from the construct predominantly terminates at the single stop codon and thus results in the production of a full-length antibody that is not fused to the second adapter peptide. However, in the presence of the antibiotic G418, translational readthrough through the stop codon is increased; however, even in the presence of the antibiotic, expression of full-length antibody not fused to the GRl coiled coil peptide is the predominant species. In general, the mistranslation results in the insertion of a random amino acid. This proportional readthrough can reflect the expressability of the full-length antibody; by monitoring both the secreted full-length antibody and the full-length antibody fusion captured at the cell surface, one can screen for high producing host cells. Thus, in the presence of the antibiotic, a population of the full-length antibodies will include the heavy chain-adapter peptide fusion protein. Therefore, when screening a library of antibodies for a desired antibody, the host cells are grown in the presence of the antibiotic. The full-length antibodies comprising the heavy chain-adapter peptide fusion protein are captured at the cell surface by heterodimerization to the first adapter peptide fused to the cell surface anchoring protein on the surface of the cell. Desired antibodies can then be detected by a suitable detection means. However, for production of full-length antibodies in which the heavy chain is not fused to the adapter peptide, host cells that have been identified to produce the desired antibody are grown in the absence of the antibiotic. The premise behind expression cassette C can be adapted to produce Fab fragments that are not fused to the GRl coiled coil peptide and can be adapted to use with other protein species such as enzymes and receptor proteins.

I. General Characteristics of the Adapters

A further consideration in constructing the display system is to select a pair of adapter peptides that encode two adapters capable of pairwise interaction. Whereas a nucleic acid encoding one of the adapter peptides is inserted in-frame with the nucleic acid encoding an exogenous protein of interest carried by the vector, a nucleic encoding the other is fused in-frame with a nucleic acid encoding a cell surface anchoring protein capable of attaching to the outer wall or membrane of the host cell. By "pairwise interaction" is meant that the two adapters can interact with and bind to each other to form a stable complex. The stable complex must be sufficiently long-lasting to permit detecting the protein of interest on the outer surface of the host cell. The complex or dimer must be able to withstand whatever conditions exist or are introduced between the moment of formation and the moment of detecting the displayed polypeptide, these conditions being a function of the assay or reaction which is being performed. The stable complex or dimer may be irreversible or reversible as long as it meets the other requirements of this definition. Thus, a transient complex or dimer may form in a reaction mixture, but it does not constitute a stable complex if it dissociates spontaneously and yields no detectable polypeptide displayed on the outer surface of a genetic package.

The pairwise interaction between the first and second adapters may be covalent or non- covalent interactions. Non-covalent interactions encompass every exiting stable linkage that do not result in the formation of a covalent bond. Non-limiting examples of noncovalent interactions include electrostatic bonds, hydrogen bonding, Van der Waal's forces, steric interdigitation of amphiphilic peptides. By contrast, covalent interactions result in the formation of covalent bonds, including but not limited to disulfide bond between two cysteine residues, C-- C bond between two carbon-containing molecules, C-O or C-H between a carbon and oxygen- or hydrogen-containing molecules respectively, and O— P bond between an oxygen-and phosphate-containing molecule.

Adapter peptides applicable for constructing the expression and helper vectors of the display system can be derived from a variety of sources. Generally, any protein sequences involved in the formation of stable multimers are candidate adapter peptides. As such, these peptides may be derived from any homomultimeric or heteromultimeric protein complexes. Representative homomultimeric proteins are homodimeric receptors (e.g., platelet-derived growth factor homodimer BB (PDGF), homodimeric transcription factors (e.g. Max homodimer, NF-kappaB p65 (ReIA) homodimer), and growth factors (e.g., neurotrophin homodimers). Non- limiting examples of heteromultimeric proteins are complexes of protein kinases and SH2- domain-containing proteins (Cantley et al., Cell 72: 767-778 (1993); Cantley et ai., J. Biol. Chem. 270: 26029-26032 (1995)), heterodimeric transcription factors, and heterodimeric receptors.

Currently used heterodimeric transcription factors are α-Pal/Max complexes and Hox/Pbx complexes. Hox represents a large family of transcription factors involved in patterning the anterior-posterior axis during embryogenesis. Hox proteins bind DNA with a conserved three alpha helix homeodomain. In order to bind to specific DNA sequences, Hox proteins require the presence of hetero-partners such as the Pbx homeodomain. Wolberger et al. solved the 2.35 A crystal structure of a HoxBl-Pbxl-DNA ternary complex in order to understand how Hox-Pbx complex formation occurs and how this complex binds to DNA. The structure shows that the homeodomain of each protein binds to adjacent recognition sequences on opposite sides of the DNA. Heterodimerization occurs through contacts formed between a six amino acid hexapeptide N-terminal to the homeodomain of HoxBl and a pocket in Pbxl formed between helix 3 and helices 1 and 2. A C-terminal extension of the Pbxl homeodomain forms an alpha helix that packs against helix 1 to form a larger four helix homeodomain (Wolberger et al., Cell 96: 587-597 (1999); Wolberger et al., J MoI Biol. 291: 521-530).

A vast number of heterodimeric receptors have also been identified. They include but are not limited to those that bind to growth factors (e.g. heregulin), neurotransmitters (e.g. γ- Aminobutyric acid), and other organic or inorganic small molecules (e.g. mineralocorticoid, glucocorticoid). Currently used heterodimeric receptors are nuclear hormone receptors (Belshaw et al., Proc. Natl. Acad. Sci. U.S. A 93:4604-4607 (1996)), erbB3 and erbB2 receptor complex, and G-protein-coupled receptors including but not limited to opioid (Gomes et al., J. Neuroscience 20: RCl 10 (2000)); Jordan et al. Nature 399: 697-700 (1999)), muscarinic, dopamine, serotonin, adenosine/dopamine, and GABA_B families of receptors. For majority of the known heterodimeric receptors, their C-terminal sequences are found to mediate heterodimer formation. Peptides derived from antibody chains that are involved in dimerizing the L and H chains can also be used as adapters for constructing the subject display systems. These peptides include but are not limited to constant region sequences of an L or H chain. Additionally, adapter peptides can be derived from antigen-binding site sequences and its binding antigen. In such case, one adapter of the pair contains antigen-binding site amino acid residues that is recognized (i.e. being able to stably associate with) by the other adapter containing the corresponding antigen residues.

Based on the wealth of genetic and biochemical data on vast families of genes, one of ordinary skill will be able to select and obtain suitable adapter peptides for constructing the subject display system without undue experimentation.

Where desired, sequences from novel hetermultimeric proteins can be employed as adapters. In such situation, the identification of candidate peptides involved in formation of heteromultimers can be determined by any genetic or biochemical assays without undue experimentation. Additionally, computer modeling and searching technologies further facilitates detection of heteromultimeric peptide sequences based on sequence homologies of common domains appeared in related and unrelated genes. Non-limiting examples of programs that allow homology searches are Blast (http://www.ncbi.nlm.nih.gov/BLAST/), Fasta (Genetics Computing Group package, Madison, Wis.), DNA Star, Clustlaw, TOFFEE, COBLATH, Genthreader, and MegAlign. Any sequence databases that contains DNA sequences corresponding to a target receptor or a segment thereof can be used for sequence analysis.

Commonly employed databases include but are not limited to GenBank, EMBL, DDBJ, PDB, SWISS-PROT, EST, STS, GSS₅ and HTGS.

The subject adapters that are derived from heterodimerization sequences can be further characterized based on their physical properties. Current heterodimerization sequences exhibit pairwise affinity resulting in predominant formation of heterodimers to a substantial exclusion of homodimers. Preferably, the predominant formation yields a heteromultimeric pool that contains at least 60% heterodimers, more preferably at least 80% heterodimers, more preferably between 85-90% heterodimers, and more preferably between 90-95% heterodimers, and even more preferably between 96-99% heterodimers that are allowed to form under physiological buffer conditions and/or physiological body temperatures. In certain embodiments of the present invention, at least one of the heterodimerization sequences of the adapter pair is essentially incapable of forming a homodimer in a physiological buffer and/or at physiological body temperature. By "essentially incapable" is meant that the selected heterodimerization sequences when tested alone do not yield detectable amounts of homodimers in an in vitro sedimentation experiment as detailed in Kammerer et al., Biochemistry 38: 13263-13269 (1999)), or in the in vivo two-hybrid yeast analysis (see e.g. White et al., Nature 396: 679-682 (1998)). In addition, individual heterodimerization sequences can be expressed in a host cell and the absence of homodimers in the host cell can be demonstrated by a variety of protein analyses including but not limited to SDS-PAGE, Western blot, and immunoprecipitation. The in vitro assays must be conducted under a physiological buffer conditions, and/or preferably at physiological body temperatures. Generally, a physiological buffer contains a physiological concentration of salt and at adjusted to a neutral pH ranging from about 6.5 to about 7.8, and preferably from about 7.0 to about 7.5.

An illustrative adapter pair exhibiting the above-mentioned physical properties is GABAB-RI /GABAB-R2 receptors. These two receptors are essentially incapable of forming homodimers under physiological conditions (e.g. in vivo) and at physiological body temperatures. Research by Kuner et al. and White et al. (Science 283 : 74-77 (1999)); Nature 396: 679-682 (1998)) has demonstrated the heterodimerization specificity of GABAβ-Rl and GABAβ-R2 in vivo. In fact, White et al. were able to clone GABAβ~R2 from yeast cells based on the exclusive specificity of this heterodimeric receptor pair. In vitro studies by Kammerer et al. supra has shown that neither GABAβ-Rl nor GABAg-R2 C-terminal sequence is capable of forming homodimers in physiological buffer conditions when assayed at physiological body temperatures. Specifically, Kammerer et al. have demonstrated by sedimentation experiments that the heterodimerization sequences of GABAg receptor 1 and 2, when tested alone, sediment at the molecular mass of the monomer under physiological conditions and at physiological body temperatures (e.g., at 37°C). When mixed in equimolar amounts, GABAB receptor 1 and 2 heterodimerization sequences sediment at the molecular mass corresponding to the heterodimer of the two sequences (see Table 1 of Kammerer et al.). However, when the GABAβ-Rl and GABAβ-R2 C-terminal sequences are linked to a cysteine residue, homodimers may occur via formation of disulfide bond.

Adapters can be further characterized based on their secondary structures. Current adapters consist of amph philic peptides that adopt a coiled-coil helical structure. The helical coiled coil is one of the principal subunit oligomerization sequences in proteins. Primary sequence analysis reveals that approximately 2-3% of all protein residues form coiled coils (Wolf et al., Protein Sci. 6: 1179-1189 (1997)). Well-characterized coiled-coil-containing proteins include members of the cytoskeletal family (e.g,. α-keratin, vimentin), cytoskeletal motor family (e.g., myosine, kinesins, and dyneins), viral membrane proteins (e.g. membrane proteins of Ebola or HIV), DNA binding proteins, and cell surface receptors (e.g. GABAβ receptors 1 and 2). Coiled-coil adapters of the present invention can be broadly classified into two groups, namely the left-handed and right-handed coiled coils. The left-handed coiled coils are characterized by a heptad repeat denoted "abcdefg" with the occurrence of apolar residues preferentially located at the first (a) and fourth (d) position. The residues at these two positions typically constitute a zigzag pattern of "knobs and holes" that interlock with those of the other stand to form a tight-fitting hydrophobic core. In contrast, the second (b), third (c) and sixth (f) positions that cover the periphery of the coiled coil are preferably charged residues. Examples of charged amino acids include basic residues such as lysine, arginine, histidine, and acidic residues such as aspartate, glutamate, asparagine, and glutamine. Uncharged or apolar amino acids suitable for designing a heterodimeric coiled coil include but are not limited to glycine, alanine, valine, leucine, isoleucine, serine and threonine. While the uncharged residues typically form the hydrophobic core, inter-helical and intra-helical salt-bridge including charged residues even at core positions may be employed to stabilize the overall helical coiled-coiled structure (Burkhard et al (2000) J. Biol Chem. 275:11672-11677). Whereas varying lengths of coiled coil may be employed, the subject coiled coil adapters preferably contain two to ten heptad repeats. More preferably, the adapters contain three to eight heptad repeats, even more preferably contain four to five heptad repeats.

In designing optimal coiled-coil adapters, a variety of existing computer software programs that predict the secondary structure of a peptide can be used. An illustrative computer analysis uses the COILS algorithm which compares an amino acid sequence with sequences in the database of known two-stranded coiled coils, and predicts the high probability coiled-coil stretches (Kammerer et al., Biochemistry 38:1326343269 (1999)).

While a diverse variety of coiled coils involved in multimer formation can be employed as the adapters in the subject display system. Currentcoiled coils are derived from heterodimeric receptors. Accordingly, the present invention encompasses coiled-coil adapters derived from G AB Ag receptors 1 and 2. In one aspect, the subject coiled coils adapters comprise the C- terminal sequences of GABAβ receptor 1 and GABAβ receptor 2. In another aspect, the subject adapters are composed of two distinct polypeptides of at least 30 amino acid residues, one of which is essentially identical to a linear sequence of comparable length depicted in SEQ ΪD NO: 13 (GRl), and the other is essentially identical to a linear peptide sequence of comparable length depicted in SEQ ID NO: 11 (GR2).

Another class of current coiled coil adapters are leucine zippers. The leucine zipper have been defined in the art as a stretch of about 35 amino acids containing 4-5 leucine residues separated from each other by six amino acids (Maniatis and Abel, Nature 341 :24 (1989)). The leucine zipper has been found to occur in a variety of eukaryotic DNA-binding proteins, such as GCN4, C/EBP, c-fos gene product (Fos), c-jun gene product (Jun), and c-Myc gene product. In these proteins, the leucine zipper creates a dimerization interface wherein proteins containing leucine zippers may form stable homodimers and/or heterodimers. Molecular analysis of the protein products encoded by two proto-oncogenes, c-fos and c-jun, has revealed such a case of preferential heterodimer formation (Gentz et al., Science 243: 1695 (1989); Nakabeppu et al., Cell 55: 907 (1988); Cohen et al., Genes Dev. 3: 173 (1989)). Synthetic peptides comprising the leucine zipper regions of Fos and Jun have also been shown to mediate heterodimer formation, and, where the amino-termini of the synthetic peptides each include a cysteine residue to permit intermolecular disulfide bonding, heterodimer formation occurs to the substantia! exclusion of homodimerization.

The leucine-zipper adapters of the present invention have the general structural formula known as the heptad repeat (Leucine-Xj-X2-X3-X4-X5-X6) n_> where X may be any of the conventional 20 amino acids, but are most likely to be amino acids with alpha-helix forming potential, for example, alanine, valine, aspartic acid, glutamic acid, and lysine, and n may be 2 or greater, although typically n is 3 to 10, preferably 4 to 8, more preferably 4 to 5, Currently, the sequences are the Fos or Jun leucine zippers.

As used herein, a linear sequence of peptide is "essentially identical" to another linear sequence, if both sequences exhibit substantial amino acid or nucleotide sequence homology. Generally, essentially identical sequences are at least about 60% identical with each other, after alignment of the homologous regions. Generally, the sequences are at least about 70% identical; more specifically, they are at least about 80% identical; more specifically, they are at least about 90% identical; more specifically, the sequences are at least about 95% identical; still more specifically, the sequences are 100% identical.

In determining whether polypeptide sequences are essentially identical, a sequence that preserves the functionality of the polypeptide with which it is being compared is particularly preferred. Functionality may be established by different criteria, such as ability to form a stable complex with a pairing adapter, and ability to facilitate display of polypeptides fused in-frame with the adapter.

The subject adapters include modified leucine zippers and GABAβ heterodimerization peptide sequences which are functionally equivalent to the polypeptide sequences exemplified herein. In particular embodiments, modified polypeptides providing improved stability to the paired adapters and/or display efficiency are used. Examples of modified polypeptides include those with conservative substitutions of amino acid residues, and one or more deletions or additions of amino acids which do not significantly deleteriously alter the heterodimerization specificity. Substitutions can range from changing or modifying one or more amino acid residues to complete redesign of a region as long as the pairwise interaction is maintained. Amino acid substitutions, if present, are preferably conservative substitutions that do not deleteriously affect folding or functional properties of the peptide. Groups of functionally related amino acids within which conservative substitutions can be made are glycine/alanine; valine/isoleucine/leucine; asparagine/glutamine; aspartic acid/glutamic acid; serine/threonine/methionine; lysine/arginine; and phenylalanine/tryosine/tryptophan. Polypeptides of this invention can be in glycosylated or unglycosylated form, can be modified post-translationally (e.g., acetylation, and phosphorylation) or can be modified synthetically (e.g., the attachment of a labeling group). One c-fos zipper is: LQAETDQLEDEKSALQTEIANLLKEKEKL (SEQ ID NO: 1). One c-Jun zipper is LEEKVKTLKAQNSELASTANMLREQVAQL (SEQ ID NO: 2). Longer forms of these zippers are as follows: c-fos: LTDTLQ AETDQLEDEKSALQ TEIANLLKEKEKLEFILA (SEQ ID NO: 3). c-Jun: RIARLEEKVKTLKAQNSELAS TANMLREQVAQLKQKVMN (SEQ ID NO: 4).

Alternative c-Jun zippers may also be used. These zippers have reduced ability to form homodimers, but still heterodimerize with c-Fos (Smeal et al. (1989) Genes & Development 3:2091-2100).

Some c-Jun zippers with reduced heterodimerϊzation ability include: LEEKVKTLKAQNSELASTFNMLREQFAQL (SEQ ID NO:5); LEEKVKTLKAQNSELASTANMLREQVAQF (SEQ ID NO:6); LEEKVKTFKAQNSELASTANMLREQVAQF (SEQ ID NO:7); LEEKVKSFKAQNSEHASTANMLREQVAQL (SEQ ID NO:8)

The adapter sequences of the present invention can be obtained using conventional recombinant cloning methods and/or by chemical synthesis. Using well-established restriction and ligation techniques, the appropriate adapter sequences can be excised from various DNA sources and integrated in-frame with the exogenous gene sequences and the outer-surface sequences to generate the expression and helper vectors, respectively.

Preferably, the second adapter sequence is inserted into the expression vector in such a way to minimize structural interference, if any, on the resulting exogenous fusion polypeptide. Whereas the first adapter can be fused to the 5 ' or 3' of the exogenous gene sequence, Figure 4 depicts a construct in which the adapter peptide sequence (i.e., hererodimerization sequence derived from GABAβ receptor 1) is fused in-frame to the 3^! end of the exogenous gene sequence.

Similarly, the first adapter peptide sequence is inserted into the second vector in a position where the integrity of the cell surface anchoring protein is not undermined. The adapter sequence can be fused to the 5' or 3' end of an outer-surface sequence without disrupting the coding region. Figure 2 depicts a vector in which the adapter sequence (i.e. heterodimerizeration sequence derived from GABA B receptor 2) is placed in-frame to the 5' end of the cell surface anchoring protein SEDl .

II. Host Cells

In general, lower eukaryotes such as yeast are used for expression of the proteins, particularly glycoproteins because they can be economically cultured, give high yields, and when appropriately modified are capable of suitable glycosylation. Yeast particularly offers established genetics allowing for rapid transformations, tested protein localization strategies and facile gene knock-out techniques. Suitable vectors have expression control sequences, such as promoters, including 3-phosphoglycerate kinase or other glycolytic enzymes, and an origin of replication, termination sequences and the like as desired.

While the invention has been demonstrated herein using the methylotrophic yeast Pichia pastoris, other useful lower eukaryote host cells include Pichia pastor is, Pichia finlandica, Pichia trehalophila, Pichia koclamae, Pichia membranaefaciens, Pichia minuta (Ogataea minuta, Pichia lindneri), Pichia opuntiae, Pichia thermotolerans, Pichia salictaria, Pichia guercuum, Pichia pijperi, Pichia stiptis, Pichia methanolica, Pichia sp., Saccharomyces cerevisiae, Saccharomyces sp,, Hansenula polymorpha, Kluyveromyces sp,, Kluyveromyces lactis, Candida albicans, Aspergillus nidulans, Aspergillus niger, Aspergillus oryzae, Trichoderma reesei, Chrysosporium lucknowense, Fusarium sp., Fusarium gramineum,

Fusarium venenatum and Neurospora crassa. Various yeasts, such as K. lactis, Pichia pastoris, Pichia methanolica, and Hansenula polymorpha are particularly suitable for cell culture because they are able to grow to high cell densities and secrete large quantities of recombinant protein. Likewise, filamentous fungi, such as Aspergillus niger, Fusarium sp, Neurospora crassa and others can be used to produce glycoproteins of the invention at an industrial scale. In the case of lower eukaryotes, cells are routinely grown from between about 1.5 to 3 days under conditions that induce expression of the capture moiety. The induction of immunoglobulin expression while inhibiting expression of the capture moiety is for about 1 to 2 days. Afterwards, the cells are analyzed for those cells that display the immunoglobulin of interest. Lower eukaryotes, particularly yeast and filamentous fungi, can be genetically modified so that they express glycoproteins in which the glycosylation pattern is human-like or humanized. In this manner, glycoprotein compositions can be produced in which a specific desired glycoform is predominant in the composition. Such can be achieved by eliminating selected endogenous glycosylation enzymes and/or genetically engineering the host cells and/or supplying exogenous enzymes to mimic all or part of the mammalian glycosylation pathway as described in US

2004/0018590. If desired, additional genetic engineering of the glycosylation can be performed, such that the glycoprotein can be produced with or without core fucosylation. Use of lower eukaryotic host cells is further advantageous in that these cells are able to produce highly homogenous compositions of glycoprotein, such that the predominant glycoform of the glycoprotein may be present as greater than thirty mole percent of the glycoprotein in the composition. In particular aspects, the predominant glycoform may be present in greater than forty mole percent, fifty mole percent, sixty mole percent, seventy mole percent and, most preferably, greater than eighty mole percent of the glycoprotein present in the composition.

Lower eukaryotes, particularly yeast, can be genetically modified so that they express glycoproteins in which the glycosylation pattern is human-like or humanized. Such can be achieved by eliminating selected endogenous glycosylation enzymes and/or supplying exogenous enzymes as described by Geragross et al., US 20040018590. For example, a host cell can be selected or engineered to be depleted in 1,6-mannosyl transferase activities, which would otherwise add mannose residues onto the Λf-glycan on a glycoprotein.

In one embodiment, the host cell further includes an αl,2-mannosidase catalytic domain fused to a cellular targeting signal peptide not normally associated with the catalytic domain and selected to target the αl ,2-mannosidase activity to the ER or Golgi apparatus of the host cell. Passage of a recombinant glycoprotein through the ER or Golgi apparatus of the host cell produces a recombinant glycoprotein comprising a Mans GIcN Ac2 glycoform, for example, a recombinant glycoprotein composition comprising predominantly a Ma^GIcNAc₂ glycoform. For example, U.S. Patent No, 7,029,872 and U.S. Published Patent Application Nos. 2004/0018590 and 2005/0170452 disclose lower eukaryote host cells capable of producing a glycoprotein comprising a MansGlcNAc2 glycoform.

In a further embodiment, the immediately preceding host cell further includes a GIcNAc transferase I (GnT I) catalytic domain fused to a cellular targeting signal peptide not normally associated with the catalytic domain and selected to target GIcNAc transferase I activity to the ER or Golgi apparatus of the host cell. Passage of the recombinant glycoprotein through the ER or Golgi apparatus of the host cell produces a recombinant glycoprotein comprising a GIcN AcMans GIcN Ac2 glycoform, for example a recombinant glycoprotein composition comprising predominantly a GIcN AcMans GIcN Ac2 glycoform. U.S. Patent No, 7,029,872 and

U.S. Published Patent Application Nos. 2004/0018590 and 2005/0170452 disclose lower eukaryote host cells capable of producing a glycoprotein comprising a GIcN AcMans GIcN Ac2 glycoform. The glycoprotein produced in the above cells can be treated in vitro with a hexaminidase to produce a recombinant glycoprotein comprising a MansGlcNAc2 glycoform.

In a further embodiment, the immediately preceding host cell further includes a mannosidase II catalytic domain fused to a cellular targeting signal peptide not normally associated with the catalytic domain and selected to target mannosidase II activity to the ER or Golgi apparatus of the host cell. Passage of the recombinant glycoprotein through the ER or Golgi apparatus of the host cell produces a recombinant glycoprotein comprising a GlcNAcMan3GlcNAc2 glycoform, for example a recombinant glycoprotein composition comprising predominantly a GlcNAcMan3GlcNAc2 glycoform. U.S. Patent No₇ 7,029,872 and U.S. Published Patent Application No. 2004/0230042 discloses lower eukaryote host cells that express mannosidase II enzymes and are capable of producing glycoproteins having predominantly a GIcN Ac2Man3 GIcN Ac2 glycoform. The glycoprotein produced in the above cells can be treated in vitro with a hexaminidase to produce a recombinant glycoprotein comprising a Man3GlcNAc2 glycoform. In a further embodiment, the immediately preceding host cell further includes GIcNAc transferase II (GnT II) catalytic domain fused to a cellular targeting signal peptide not normally associated with the catalytic domain and selected to target GIcNAc transferase II activity to the ER or Golgi apparatus of the host cell. Passage of the recombinant glycoprotein through the ER or Golgi apparatus of the host cell produces a recombinant glycoprotein comprising a GIcN Ac2Man3 GIcN Ac2 glycoform, for example a recombinant glycoprotein composition comprising predominantly a GIcN Ac2Man3 GIcN Ac2 glycoform. U.S. Patent No, 7,029,872 and U.S. Published Patent Application Nos. 2004/0018590 and 2005/0170452 disclose lower eukaryote host cells capable of producing a glycoprotein comprising a GIcN Ac2Man3 GIcN Ac2 glycoform. The glycoprotein produced in the above cells can be treated in vitro with a hexaminidase to produce a recombinant glycoprotein comprising a Man₃ GIcNAc₂ glycoform. In a further embodiment, the immediately preceding host cell further includes a galactosyltransferase catalytic domain fused to a cellular targeting signal peptide not normally associated with the catalytic domain and selected to target galactosyltransferase activity to the ER or Golgi apparatus of the host cell. Passage of the recombinant glycoprotein through the ER or Golgi apparatus of the host cell produces a recombinant glycoprotein comprising a GalGlcNAc2Man3GlcNAc2 or Gal2GlcNAc2Man3GlcNAc2 glycoform, or mixture thereof for example a recombinant glycoprotein composition comprising predominantly a

GalGlcNAc2Man3GlcNAc2 glycoform or Gal2GlcNAc2Man3GlcNAc2 glycoform or mixture thereof. U.S. Patent No, 7,029,872 and U.S. Published Patent Application No. 2006/0040353 discloses lower eukaryote host cells capable of producing a glycoprotein comprising a GaI₂GIcNAc₂MaH₃GIcNAc₂ glycoform. The glycoprotein produced in the above cells can be treated in vitro with a galactosidase to produce a recombinant glycoprotein comprising a GIcN Ac2Man3 GIcN Ac2 glycoform, for example a recombinant glycoprotein composition comprising predominantly a GlcNAc2Man3GlcNAc2 glycoform.

In a further embodiment, the immediately preceding host cell further includes a sialyl transferase catalytic domain fused to a cellular targeting signal peptide not normally associated with the catalytic domain and selected to target sialytransferase activity to the ER or Golgi apparatus of the host cell. Passage of the recombinant glycoprotein through the ER or Golgi apparatus of the host cell produces a recombinant glycoprotein comprising predominantly a NANA2Gal2GlcNAc₂Man3GIcNAc2 glycoform or NANAGal2GlcNAc2Man3GlcNAc2 glycoform or mixture thereof. For lower eukaryote host cells such as yeast and filamentous fungi, it is useful that the host cell further include a means for providing CMP-sialic acid for transfer to the jV-glycan. U.S. Published Patent Application No. 2005/0260729 discloses a method for genetically engineering lower eukaryotes to have a CMP-sialic acid synthesis pathway and U.S. Published Patent Application No. 2006/0286637 discloses a method for genetically engineering lower eukaryotes to produce sialylated glycoproteins. The glycoprotein produced in the above cells can be treated in vitro with a neuraminidase to produce a recombinant glycoprotein comprising predominantly a Gal2GlcNAc2Man3GlcNAc2 glycoform or GalGlcNAc2Man3GlcNAc2 glycoform or mixture thereof. Any one of the preceding host cells can further include one or more GIcNAc transferase selected from the group consisting of GnT HI, GnT IV, GnT V, GnT VI₅ and GnT IX to produce glycoproteins having bisected (GnT HI) and/or multiantennary (GnT IV, V, VI, and IX) ϊV-glycan structures such as disclosed in U.S. Published Patent Application Nos. 2004/074458 and 2007/0037248.

In further embodiments, the host cell that produces glycoproteins that have predominantly GIcN AcMans GIcN Ac2 Λf-glycans further includes a galactosyl transferase catalytic domain fused to a cellular targeting signal peptide not normally associated with the catalytic domain and selected to target Galactosyltransferase activity to the ER or Golgi apparatus of the host cell. Passage of the recombinant glycoprotein through the ER or Golgi apparatus of the host cell produces a recombinant glycoprotein comprising predominantly the GalGlcNAcMan5GlcNAc2 glycoform.

In a further embodiment, the immediately preceding host cell that produced glycoproteins that have predominantly the GalGlcNAcMan5GlcNAc2 JV-glycans further includes a sialyltransferase catalytic domain fused to a cellular targeting signal peptide not normally associated with the catalytic domain and selected to target sialytransferase activity to the ER or Golgi apparatus of the host cell. Passage of the recombinant glycoprotein through the ER or Golgi apparatus of the host cell produces a recombinant glycoprotein comprising a NANAGalGlcNAcMan5GlcNAc2 glycoform. Various of the preceding host cells further include one or more sugar transporters such as UDP-GIcNAc transporters (for example, Kluyveromyces lactis and Mns musculus UDP- GIcNAc transporters), UDP-galactose transporters (for example, Drosophila melanogaster UDP- galactose transporter), and CMP-sialic acid transporter (for example, human sialic acid transporter). Because lower eukaryote host cells such as yeast and filamentous fungi lack the above transporters, it is preferable that lower eukaryote host cells such as yeast and filamentous fungi be genetically engineered to include the above transporters.

Host cells further include lower eukaryote cells (e.g., yeast such as Pichia pastoris) that are genetically engineered to eliminate glycoproteins having α-mannosidase-resistant 7V-glycans by deleting or disrupting one or more of the β-mannosyltransferase genes (e.g., BMTl, BMT2, BMT3, and BMT4)(See, U.S. Published Patent Application No. 2006/0211085) and glycoproteins having phosphomannose residues by deleting or disrupting one or both of the phosphomannosyl transferase genes PNOl and MNN4B (See for example, U.S. Patent Nos. 7,198,921 and 7,259,007), which in further aspects can also include deleting or disrupting the MNN4A gene. Disruption includes disrupting the open reading frame encoding the particular enzymes or disrupting expression of the open reading frame or abrogating translation of RNAs encoding one or more of the β-mannosyltransferases and/or phosphomannosy .transferases using interfering RNA, antisense RNA, or the like. The host cells can further include any one of the aforementioned host cells modified to produce particular JV-glycan structures.

Host cells further include lower eukaryote cells (e.g., yeast such as Pichia pastoris) that are genetically modified to control O-glycosylation of the glycoprotein by deleting or disrupting one or more of the protein O-mannosyltransferase (Dol-P-Man: Protein (Ser/Thr) Mannosyl Transferase genes) (PMTs) (See U.S. Patent No. 5,714,377) or grown in the presence of Pmtp inhibitors and/or an alpha-mannosidase as disclosed in Published International Application No. WO 2007061631, or both. Disruption includes disrupting the open reading frame encoding the Pmtp or disrupting expression of the open reading frame or abrogating translation of RNAs encoding one or more of the Pmtps using interfering RNA, antisense RNA, or the like. The host cells can further include any one of the aforementioned host cells modified to produce particular iV-glycan structures..

Pmtp inhibitors include but are not limited to a benzylidene thiazolidinediones. Examples of benzylidene thiazolidinediones that can be used are 5-[[3,4-bis(phenyImethoxy) phenyl]methylene]-4-oxo-2-thioxo-3-thiazolidineacetic Acid; 5-[[3-(l-Phenylethoxy)-4-(2- phenylethoxy)]phenyl]methylene]-4-oxo-2-thioxo-3-thiazolidineacetic Acid; and 5-[[3-(l- Phenyl-2-hydroxy)ethoxy)-4-(2-phenylethoxy)]phenyl]methylene]-4-oxo-2-thioxo-3- thiazolidineacetic Acid.

In particular embodiments, the function or expression of at least one endogenous PMT gene is reduced, disrupted, or deleted. For example, in particular embodiments the function or expression of at least one endogenous PMT gene selected from the group consisting of the PMTl, PMT2, PMT3, and PMT4 genes is reduced, disrupted, or deleted; or the host cells are cultivated in the presence of one or more PMT inhibitors. In further embodiments, the host cells include one or more PMT gene deletions or disruptions and the host cells are cultivated in the presence of one or more Pmtp inhibitors. In particular aspects of these embodiments, the host cells also express a secreted alpha- 1,2-mannosidase.

PMT deletions or disruptions and/or Pmtp inhibitors control O-glycosylation by reducing O-glycosylation occupancy, that is by reducing the total number of O-glycosylation sites on the glycoprotein that are glycosylated. The further addition of an alpha- 1 ,2-mannsodase that is secreted by the cell controls O-glycosylation by reducing the mannose chain length of the O- glycans that are on the glycoprotein. Thus, combining PMT deletions or disruptions and/or Pmtp inhibitors with expression of a secreted alpha- 1₅2-mannosidase controls O-glycosylation by reducing occupancy and chain length. In particular circumstances, the particular combination of PMT deletions or disruptions, Pmtp inhibitors, and alpha- 1 ,2-mannosidase is determined empirically as particular heterologous glycoproteins (Fabs and antibodies, for example) may be expressed and transported through the Golgi apparatus with different degrees of efficiency and thus may require a particular combination of PMT deletions or disruptions, Pmtp inhibitors, and alpha- 1,2-mannosidase. In another aspect, genes encoding one or emore endogenous mannosyltransferase enzymes are deleted. This deletion(s) can be in combination with providing the secreted alpha- 1 ,2-mannosidase and/or PMT inhibitors or can be in lieu of providing the secreted alpha- 1,2-mannosidase and/or PMT inhibitors. Thus, the control of 0-glycosylaιion can be useful for producing particular glycoproteins in the host cells disclosed herein in better total yield or in yield of properly assembled glycoprotein. The reduction or elimination of O-glycosylation appears to have a beneficial effect on the assembly and transport of whole antibodies and Fab fragments as they traverse the secretory pathway and are transported to the cell surface. Thus, in cells in which O-glycosylation is controlled, the yield of properly assembled antibodies or Fab fragments is increased over the yield obtained in host cells in which O-glycosylation is not controlled.

In addition, O-glycosylation may have an effect on an antibody or Fab fragment's affinity and/or avidity for an antigen. This can be particularly significant when the ultimate host cell for production of the antibody or Fab is not the same as the host cell that was used for selecting the antibody. For example, O-glycosylation might interfere with an antibody's or Fab fragment's affinity for an antigen, thus an antibody or Fab fragment that might otherwise have high affinity for an antigen might not be identified because O-glycosylation may interfere with the ability of the antibody or Fab fragment to bind the antigen. In other cases, an antibody or Fab fragment that has high avidity for an antigen might not be identified because O-glycosylation interferes with the antibody's or Fab fragment's avidity for the antigen. In the preceding two cases, an antibody or Fab fragment that might be particularly effective when produced in a mammalian cell line might not be identified because the host cells for identifying and selecting the antibody or Fab fragment was of another cell type, for example, a yeast or fungal cell (e.g., a Pichia pastoris host cell). It is well known that O-glycosylation in yeast can be significantly different from O-glycosylation in mammalian cells. This is particularly relevant when comparing wild type yeast o-glycosylation with mucin-type or dystroglycan type O-glycosylation in mammals. In particular cases, O-glycosylation might enhance the antibody or Fab fragments affinity or avidity for an antigen instead of interfere. This effect is undesirable when the production host cell is to be different from the host cell used to identify and select the antibody or Fab fragment (for example, identification and selection is done in yeast and the production host is a mammalian cell) because in the production host the O-glycosylation will no longer be of the type that caused the enhanced affinity or avidity for the antigen. Therefore, controlling O- glycosylation can enable use of the materials and methods herein to identify and select antibodies or Fab fragments with specificity for a particular antigen based upon affinity or avidity of the antibody or Fab fragment for the antigen without identification and selection of the antibody or Fab fragment being influenced by the O-glycosylation system of the host cell. Thus, controlling 0-glycosylation further enhances the usefulness of yeast or fungal host cells to identify and select antibodies or Fab fragments that will ultimately be produced in a mammalian cell line.

Yield of antibodies and Fabs can in some situations be improved by overexpressing nucleic acid molecules encoding mammalian or human cliaperone proteins or replacing the genes encoding one or more endogenous chaperone proteins with nucleic acid molecules encoding one or more mammalian or human chaperone proteins. In addition, the expression of mammalian or human chaperone proteins in the host cell may control <9-glycosylation in the cell. Thus, further included are the host cells herein wherein the function of at least one endogenous gene encoding a chaperone protein has been reduced or eliminated, and a vector encoding at least one mammalian or human homolog of the chaperone protein is expressed in the host cell. Also included are host cells in which the endogenous host cell chaperones and the mammalian or human chaperone proteins are expressed. In further aspects, the lower eukaryotic host cell is a yeast or filamentous fungi host cell. Examples of the use of chaperones of host cells in which human chaperone proteins are introduced to improve the yield and reduce or control O- glycosylation of recombinant proteins has been disclosed in U.S. Provisional Application Nos. 61/066409 filed February 20, 2008 and 61/188,723 filed August 12, 2008. Like above, further included are lower eukaryotic host cells wherein, in addition to replacing the genes encoding one or more of the endogenous chaperone proteins with nucleic acid molecules encoding one or more mammalian or human chaperone proteins or overexpressing one or more mammalian or human chaperone proteins as described above, the function or expression of at least one endogenous gene encoding a protein (9-mannosyltransferase (PMT) protein is reduced, disrupted, or deleted.

In particular embodiments, the function of at least one endogenous PMT gene selected from the group consisting of the PMTl, PMT2, PMT3, and P MT 4 genes is reduced, disrupted, or deleted.

Therefore, the methods disclosed herein can use any host cell that has been genetically modified to produce glycoproteins that have no iV-glycan compositions wherein the predominant iV-glycan is selected from the group consisting of complex 7V-glycans, hybrid iV-glycans, and high mannose iV-glycans wherein complex JV-glycans are selected from the group consisting of Man3GlcNAc2, GIcN AC(i-4)Man3 GIcN Ac2, Gal(i_4)GlcNAc(i_4)Man3GlcNAc2, and NANA(i_4)Gal(i-4)Man3GlcNAc2; hybrid iV-glycans are selected from the group consisting of GlcNAcMan5GlcNAc₂, GalGlcNAcMansGlcNAc2, and NANAGalGlcNAcMan₅GlcNAc2; and high Mannose iV-glycans are selected from the group consisting of Mans GIcN Ac2, Man6GlcNAc2, Man7GlcNAc2, Man8GlcNAc2, and Man9GlcNAc2- In particular aspects, the composition of ΪV-glycans comprises about 39% GlcNAC2Man3GlcNAc2; 40% GaliGlcNAC2Man3GlcNAc₂; and 6% GatøGlcN AC2Man3 GIcN Ac2 or about 60% GlcNAC2Man3GlcNAc2; 17% GaIi GIcN AC2Man3 GIcNAc₂; and 5% Gal2GlcNAC2Man3GlcNAc2, or mixtures in between. In the above embodiments in which the yeast cell does not display 1 ,6-mannosyI transferase activity (that is, the OCHl gene encoding ochlp has been disrupted or deleted or the activity of Ochlp has been disabled), the host cell is not capable of mating. Thus, depending on the efficiency of transformation, the potential library diversity of light chains and heavy chains appears to be limited to a heavy chain library of between about \Φ to 10ό diversity and a light chain library of about 10³ to 10<> diversity. However, in a yeast host cell that is capable of mating, the diversity can be increased to about 106 to 1012 because the host cells expressing the heavy chain library can be mated to host cells expressing the light chain library to produce host cells that express heavy chain/light chain library. Therefore, in particular embodiments, the host cell is a yeast cell such as Pichia past oris that displays 1,6-mannosyl transferase activities (that is, has an OCH/ gene encoding a functional ochlp) but which is modified as described herein to display antibodies or fragments thereof on the cell surface. In these embodiments, the host cell can be a host cell with its native glycosylation pathway.

In embodiments that express whole antibodies or the Fc region of an antibody heavy chain (e.g., Fab fragments), the nucleic acid molecule encoding the antibody or heavy chain fragment thereof is modified to replace the codon encoding an asparagine residue at position 297 of the molecule (the glycosylation site) with a codon encoding any other amino acid residue. Common replacements include but are not limited to alanine, glutamine, and aspartate. Thus, the antibody or fragment thereof that is produced in the host cell is not glycosylated at asparagine-297. In this embodiment, the host cell displaying the heavy chain library is mated to the host cell displaying the light chain library and the resulting combinatorial library is screened as taught herein. Because the antibodies or fragments thereof lack JV~glycosylatϊon at asparagine- 297, the non-human yeast iV-glycans of the host cell linked to asparagine-297 which might interfere with antibody affinity for a desired antigen are not present on the recombinant antibodies or fragments thereof. Cells producing antibodies or fragments that have desired affinity for an antigen of interest are selected. The nucleic acid molecules encoding the heavy and light chains of the antibody or fragments thereof are removed from the cells and the nucleic acid molecule encoding the heavy chain is modified to reintroduce an asparagine residue at position 297. This enables appropriate human-like glycosylation at position 297 of the antibody or fragment thereof when the nucleic acid molecule encoding the antibody or fragment thereof is introduced into a mammalian cell line (e.g., CHO or the like) or lower eukaryote (e.g., Pichia pastoris) host cell that has been engineered to make glycoproteins that have human-like JV- glycans (e.g., high mannose, hybrid, or complex JV-glycans as discussed previously.

While in general the host cells used to practice the present invention are lower eukaryote host cells (e.g., yeast or filamentous fungal cells), it is envisioned that the methods herein can be adapted to use higher eukaryote cells. Thus, in particular embodiments, the cell systems used for recombinant expression and display of the immunoglobulin can also be any higher eukaryote cell, tissue, organism from the animal kingdom, for example transgenic goats, transgenic rabbits, CHO cells, insect cells, and human cell lines. Examples of animal cells include, but are not limited to, SC-I cells, LLC-MK cells, CV-I cells, CHO cells, COS cells, murine cells, human cells, HeLa cells, 293 cells, VERO cells, MDBK cells, MDCK cells, MDOK cells, CRPK cells, RAF cells, TCMK cells, LLC-PK cells, PKl 5 cells, WI-38 cells, MRC-5 cells, T-FLY cells, BHK cells, SP2/0, NSO cells, and derivatives thereof. Insect cells include cells of Drosophila melanogaster origin. In addition, these cells can be genetically engineered to render the cells capable of making immunoglobulins that have particular JV-glycans or predominantly particular /V-glycans. For example, U.S. Patent No. 6,949,372 discloses methods for making glycoproteins in insect cells that are sialylated. Yamane-Ohnuki et al, Biotechnol. Bioeng. 87: 614-622 (2004), Kanda et al., Biotechnol. Bioeng. 94: 680-688 (2006), Kanda et al, Glycobiol. 17: 104-118 (2006), and U.S. Pub. Application Nos. 2005/0216958 and 2007/0020260 disclose mammalian cells that are capable of producing immunoglobulins in which the iV-glycans thereon lack fucose or have reduced fucose. In particular embodiments, the higher eukaryote cell, tissue, organism can also be from the plant kingdom, for example, wheat, rice, corn, tobacco, and the like. Alternatively, bryophyte cells can be selected, for example from species of the genera Physcomitrella, Funaria, Sphagnum, Cemtodon, Marchantia, and Sphaerocarpos. Exemplary of plant cells is the bryophyte cell of Physcomitrella patens, which has been disclosed in WO 2004/057002 and WO2008/006554. Expression systems using plant cells can further manipulated to have altered glycosylation pathways to enable the cells to produce immunoglobulins that have predominantly particular jV-glycans. For example, the cells can be genetically engineered to have a dysfunctional or no core fucosyltransferase and/or a dysfunctional or no xylosyltransferase, and/or a dysfunctional or no βl,4-galactosyltransferase. Alternatively, the galactose, fucose and/or xylose can be removed from the immunoglobulin by treatment with enzymes removing the residues. Any enzyme resulting in the release of galactose, fucose and/or xylose residues from iV-glycans which are known in the art can be used, for example α-galactosidase, β- xylosidase, and α-fucosidase. Alternatively an expression system can be used which synthesizes modified iV-glycans which can not be used as substrates by 1,3 -fucosyltransferase and/or 1 ,2- xylosyltransferase, and/or 1 ,4-galactosyltransferase. Methods for modifying glycosylation pathways in plant cells has been disclosed in U.S. Published Application No. 2004/0018590.

The methods disclosed herein can be adapted for use in mammalian, insect, and plant cells. The regulatable promoters selected for regulating expression of the expression cassettes in mammalian, insect, or plant cells should be selected for functionality in the cell-type chosen. Examples of suitable regulatable promoters include but are not limited to the tetracycline- regulatable promoters (See for example, Berens & Hillen, Eur. J. Biochem. 270: 3109-3121 (2003)), RU 486-inducible promoters, ecdysone-inducible promoters, and kanamycin-regulatable systems. These promoters can replace the promoters exemplified in the expression cassettes described in the examples. The capture moiety can be fused to a cell surface anchoring protein suitable for use in the cell-type chosen. Cell surface anchoring proteins including GPI proteins are well known for mammalian, insect, and plant cells. GPI-anchored fusion proteins has been described by Kennard et ah, Methods Biotechnol. Vo. 8: Animal Cell Biotechnology (Ed.

Jenkins. Human Press, Inc., Totowa, NJ) pp. 187-200 (1999). The genome targeting sequences for integrating the expression cassettes into the host cell genome for making stable recombinants can replace the genome targeting and integration sequences exemplified in the examples. Transfection methods for making stable and transiently transfected mammalian, insect, plant host cells are well known in the art. Once the transfected host cells have been constructed as disclosed herein, the cells can be screened for expression of the immunoglobulin of interest and selected as disclosed herein.

HI. Glycosylphosphatidylinositol-anchored (GPI) proteins Lower eukaryotic cells have systems of GPI proteins that are involved in anchoring or tethering expressed proteins to the cell wall so that they are effectively displayed on the cell wall of the cell from which they were expressed. For example, 66 putative GPI proteins have been identified in Saccharomyces cerevisiae (See, de Groot et al., Yeast 20: 781-796 (2003)). GPI proteins which may be used in the methods herein include, for example Saccharomyces cerevisiae CWPl; CWP2; SEDl; GASl; Pichia pastoris SPl; GASl; and K pofymorpha TIPl. Additional GPI proteins may also be useful. Suitable GPI proteins can be identified using the methods and materials of the invention described and exemplified herein.

The selection of the appropriate GPI protein will depend on the particular recombinant protein to be produced in the host cell and the particular post-translation modifications to be performed on the recombinant protein. For example, production of antibodies or fragments thereof with particular glycosylation patterns will entail the use of recombinant host cells that produce glycoproteins having particular glycosylation patterns. The GPI protein most suitable in a system for producing antibodies or fragments thereof that have predominantly Man5GlcNAc2

N-glycosylation many not necessarily be the GPI protein most suitable in a system for producing antibodies or thereof having predominantly Gal2GlcNAc2Man3GlcNAc2 N-glycosylation. In addition, the GPI most suitable in a system for producing antibodies or fragments thereof specific for one epitope or antigen may not necessarily be the most suitable GPI protein in a system for producing antibodies or fragments thereof specific for another epitope or antigen. Furthermore, the GPI most suitable in a system for producing antibody fragments such as scFv or the like may not necessarily be the most suitable GPI protein in a system for producing full-length antibodies. Therefore, further provided is a library method for constructing the host cell that is to be used for producing a particular recombinant protein. In general, the host that is desired to produce the recombinant proteins is selected based on the desired characteristics that will be imparted to the recombinant protein produced by the host cell. For example, a host cell that produces glycoproteins having predominantly Mans GIcN Ac2 or Gal2GlcNAc2Man3GlcNAc2 N- glycosylation is selected. A library of vectors encoding GPI proteins fused to one or more adapters is then provided, A library of host cells is then constructed wherein each host cell to make up the library is transfected with one of the vectors in the library of vectors encoding GPI- adapter fusion proteins such that each host cell species in the library will express one particular GPI-adapter fusion protein. Each host cell species of the library is then transformed with a vector encoding the desired protein or a protein similar in function or structure to the desired protein. The host cell that results in the best presentation of recombinant protein on the surface of the host cell is selected as the host cell for producing the desired recombinant protein.

In general, the GPI protein used in the methods disclosed herein is a chimeric protein or fusion protein comprising the GPI protein fused at its N-terminus to the C-terminus of a binding moiety or adapter peptide. The N-terminus of the binding moiety or adapter peptide is fused to the C-terminus of a signal sequence that enables the GPI fusion protein to be transported through the secretory pathway to the cell surface where the GPI fusion protein is secreted and then bound to the cell surface. In some aspects, the GPI fusion protein comprises the entire GPI protein and in other aspects, the GPI fusion protein comprises the portion of the GPI protein that is capable of binding to the cell surface.

V. Regulatory Sequences

Regulatory sequences which may be used in the practice of the methods disclosed herein include signal sequences, promoters, and transcription terminator sequences. It is generally preferred that the regulatory sequences used be from a species or genus that is the same as or closely related to that of the host cell or is operational in the host cell type chosen. Examples of signal sequences include those of Saccharomyces cerevisiae invertase; the Aspergillus niger amylase and glucoamylase; human serum albumin; Kluyveromyces maxianus inulinase; and Pichia pastor is mating factor and Kar2. Signal sequences shown herein to be useful in yeast and filamentous fungi include, but are not limited to, the alpha mating factor presequence and preprosequence from Saccharomyces cerevisiae; and signal sequences from numerous other species.

Examples of promoters include promoters from numerous species, including but not limited to alcohol-regulated promoter, tetracycline-regulated promoters, steroid-regulated promoters (e.g., glucocorticoid, estrogen, ecdysone, retinoid, thyroid), metal -regulated promoters, pathogen-regulated promoters, temperature-regulated promoters, and light-regulated promoters. Specific examples of regulatable promoter systems well known in the art include but are not limited to metal-inducible promoter systems (e.g., the yeast copper-metallothionein promoter), plant herbicide safner-activated promoter systems, plant heat-inducible promoter systems, plant and mammalian steroid-inducible promoter systems, Cym repressor-promoter system (Krackeler Scientific, Inc. Albany, NY), RheoSwitch System (New England Biolabs, Beverly MA), benzoate-inducible promoter systems (See WO2004/043885), and retroviral-inducible promoter systems. Other specific regulatable promoter systems well-known in the art include the tetracycline-regulatable systems (See for example, Berens & Hillen, Eur J Biochem 270: 3109- 3121 (2003)), RU 486-inducibIe systems, ecdysone-inducible systems, and kanamycin- regulatable system. Lower eukaryote-specific promoters include but are not limited to the Saccharomyces cerevisiae TEF-I promoter, Pichia pastoris GAPDH promoter, Pichia pastoris GUTl promoter, PMA-I promoter, Pichia pastoris PCK-I promoter, and Pichia pastoris AOX-I and AOX- 2 promoters. For temporal expression of the GPI-IgG capture moiety and the immunoglobulins, the Pichia pastoris GUTl promoter operably linked to the nucleic acid molecule encoding the GPI-IgG capture moiety and the Pichia pastoris GAPDH promoter operably linked to the nucleic acid molecule encoding the immunoglobulin are shown in the examples herein to be useful.

Examples of transcription terminator sequences include transcription terminators from numerous species and proteins, including but not limited to the Saccharomyces cerevisiae cytochrome C terminator; and Pichia pastoris ALG3 and PMAl terminators.

VI. Nucleic Acid Sequences Encoding the Protein of Interest

The methods of the present invention can be employed with any gene of interest for further study. Because of the particular advantages afforded by the methods disclosed herein, the methods and materials will utilize genes encoding glycoproteins. Of particular interest are human glycoproteins with known therapeutic utility, including but not limited to monoclonal antibodies and functional fragments thereof such as Fab fragments; immunoglobulins including but not limited to IgG, IgM, IgD, antibody fragments such as scFv, Fab fragments, or the like; Fc fusion proteins; catalytic antibodies, camel or lama antibodies; erythropoietin; cytokines such as interferon-alpha, interferon-beta, interferon-gamma, interferon-omega, and granulocyte-CSF; coagulation factors such as factor VIII, factor IX, and human protein C; soluble IgE receptor alpha-chain; urokinase; chymase and urea trypsin inhibitor; IGF-binding protein; epidermal growth factor; growth hormone-releasing factor; annexin V fusion protein; angiostatin; vascular endothelial growth factor-2; myeloid progenitor inhibitory factor- 1; and osteoprotegerin.

Nucleic acids encoding desired glycoproteins can be obtained from several sources. cDNA sequences can be amplified from cell lines known to express the glycoprotein using primers to conserved regions (See, e.g., Marks et al., J. MoI. Biol. 581-596 (1991)). Nucleic acids can also be synthesized de novo based on sequences in the scientific literature. Nucleic acids can also be synthesized by extension of overlapping oligonucleotides spanning a desired sequence (See, e.g., Caldas et al., Protein Engineering, 13: 353-360 (2000)). Production of active glycoproteins requires proper folding of the protein when it is produced and secreted by the cells. The presence of effective molecular chaperone proteins may be required, or may enhance the ability of the cell to produce and secrete properly folded proteins. Nucleic acid molecules encoding immunoglobulins can be obtained from any suitable source including spleen and liver cells and antigen-stimulated antibody producing cells, obtained from either in vivo or in vitro sources. Regardless of source, the cellular VH and VL mRNAs are reverse transcribed into VH and VL cDNA sequences. Reverse transcription may be performed in a single step or in an optional combined reverse transcription/PCR procedure to produce cDNA libraries containing a plurality of immunoglobulin-encoding DNA molecules. (See, for example, Marks et al, J. MoI. Biol. 222: 581-596 (1991)). Nucleic acid molecules can also be synthesized de novo based on sequences in the scientific literature. Nucleic acid molecules can also be synthesized by extension of overlapping oligonucleotides spanning a desired sequence (See, e.g., Caldas et al, Protein Engineering, 13: 353-360 (2000)). Humanized immunoglobulin- encoding cDNA libraries can be constructed by PCR amplifying the complementary-determining regions (CDR) from the cDNAs in one or more libraries from any source and integrating the PCR amplified CDR-encoding nucleic acid molecules into nucleic acid molecules encoding a human immunoglobulin framework to produce a cDNA library encoding a plurality of humanized immunoglobulins (See, for example, U.S. Patent Nos. 6,180,370; 6,632,927; and, 6,872,392). Chimeric immunoglobulin-encoding cDNA libraries can be constructed by PCR amplifying the variable regions from the cDNAs in the cDNA library from one species and integrating the nucleic acid molecules encoding the PCR-amplified variable regions onto nucleic acid molecules encoding immunoglobulin constant regions from another species to produce a cDNA library encoding a plurality of chimeric immunoglobulins (See, for example, U.S. Patent No. 5,843,708). Various methods that have been developed for the creation of diversity within protein libraries, including random mutagenesis (Daugherty et al., Proc. Natl Acad. Sd. USA, 97:, 2029-2034 (2000);Boder et al, Proc. Natl Acad. Sci. USA, 97:, 10701-10705 (2000); Holler et al, Proc. Natl Acad. Sci. USA, 97:, 5387-5392 (2000)), in vitro DNA shuffling (Stemmer, Nature, 370:, 389-391 (1994); Stemmer, Proc. Natl Acad. Sci. USA, 91 :, 10747- 10751 (1994)), in vivo DNA shuffling (Swers et al., Nucl. Acid Res. 32: e36 (2004)), and site- specific recombination (Rehberg et al, J. Biol. Chem., 257:, 11497-11502 (1982); Streuli et al, Proc. Natl Acad. Sci. USA, 78:, 2848-2852 (1981); Waterhouse et al,. (1993) Nucl. Acids Res., 21:, 2265-2266 (1993); Sblattero & Bradbury, Nat. BiotechnoL, 18:, 75-80 (2000)) can be used or adapted to produce the plurality of host cells disclosed herein that express immunoglobulins and the capture moiety comprising a cell surface anchoring protein fused to a binding moiety that is capable of specifically binding an immunoglobulin. Production of active immunoglobulins requires proper folding of the protein when it is produced and secreted by the cells. In E.coli, the complexity and large size of an antibody presents an obstacle to proper folding and assembly of the expressed light and heavy chain polypeptides, resulting in poor yield of intact antibody. The presence of effective molecular chaperone proteins may be required, or may enhance the ability of the cell to produce and secrete properly folded proteins. The use of molecular chaperone proteins to improve production of immunoglobulins in yeast has been disclosed in U.S. Patent No. 5,772,245; U.S. Patent Nos. 5,700,678 and 5,874,247; U.S. Application Publication No. 2002/0068325; Toman et al., J. Biol. Chem. 275: 23303-23309 (2000); Keizer-Gunnink et al, Martix Biol. 19: 29-36 (2000); Vad et al, J- Biotechnol. 116: 251-260 (2005); Inana et al, Biotechnol. Bioengineer. 93: 771-778 (2005); Zhang et al, Biotechnol. Prog. 22: 1090-1095 (2006); Damasceno et al., Appl. Microbiol. Biotechnol. 74: 381-389 (2006); Huo et al, Protein Express. Purif. 54: 234-239 (2007); and copending application Serial No. 61/066,409, filed 20 February 2008.

As used herein, the methods can use host cells from any kind of cellular system which can be modified to express a capture moiety comprising a cell surface anchoring protein fused to a binding moiety capable of binding an immunoglobulin and whole, intact immunoglobulins. Within the scope of the invention, the term "cells" means the cultivation of individual cells, tissues, organs, insect cells, avian cells, reptilian cells, mammalian cells, hybridoma cells, primary cells, continuous cell lines, stem cells, plant cells, yeast cells, filamentous fungal cells, and/or genetically engineered cells, such as recombinant cells expressing and displaying a glycosylated immunoglobulin.

VII. Uses of the Adapter-Directed Display Systems

The adapter-directed display systems disclosed herein allows the display of monomeric and multimeric polypeptides on the surface of suitable lower eukaryote host cells. The subject display systems also can be used to create libraries of random or predetermined polypeptides, full-length proteins, and protein domains for a variety of purposes. For instance, the displayed libraries can be employed for mapping epitopes and mimotopes, identifying antagonists and agonists of various target proteins, engineering antibodies, optimizing antibody specificities and creating novel binding activities.

Accordingly, provided is a method of detecting the presence of a specific interaction between a test agent and an exogenous polypeptide that is displayed on the surface of a suitable lower eukaryote host cell. The method involves the steps of: (a) providing a lower eukaryote host cell of the subject display system that presents the exogenous polypeptide; (b) contacting the lower eukaryote host cell with the test agent under conditions suitable to produce a stable polypeptide-agent complex; and (c) detecting the formation of the stable polypeptide-agent complex on the surface of the lower eukaryote host cell, thereby detecting the presence of the specific interaction.

The term "test agent" is intended to include, but not be limited to a biological or chemical compound such as a simple or complex organic or inorganic molecule, a protein, carbohydrate, lipid, polynucleotide or combinations thereof. A vast array of compounds can be synthesized, for example oligomers, such as oligopeptides and oligonucleotides, and synthetic organic compounds based on various core structures, and these are also included in the term "agent." In addition, various natural sources can provide compounds for screening, such as plant or animal extracts, and the like. It should be understood, although not always explicitly stated that the agent is used alone or in combination with another agent, having the same or different biological activity as the agents identified by the inventive screen. In particular embodiments, the agents are candidate diagnostics and/or therapeutics, such as those capable of modulating the signal transduction pathways of a cell.

In a separate embodiment, the present invention provides a method of obtaining a polypeptide with desired property. The method comprises the steps of (a) providing a selectable library of the subject display system; and (b) screening the selectable library to obtain at least one lower eukaryote host cell displaying a polypeptide on its surface with the desired property. The method may further comprise the step of isolating the lower eukaryote host cell that displays a polypeptide having the desired property. Such isolation of the lower eukaryote host cell may involve obtaining a nucleotide sequence from the lower eukaryote host cell that encodes the desired polypeptide. The desired property encompasses the ability of the polypeptide to specifically bind to an agent of interest. The selected polypeptide with the desired property may fall within one or more classes of the following molecules, namely antigen-binding unit, cell surface receptor, receptor Hgand, cytosolic protein, secreted protein, nuclear protein, and functional motif thereof. The choice of specific agent to be tested and the libraries of exogenous polypeptides to be displayed will depend on the intended purpose of the screening assay.

VIII. Isolating Antibodies Exhibiting Desired Binding Specificity or Affinity

One of the most powerful applications of display system herein is its use in the arena of antibody engineering. It has been shown that scFv antigen-binding units can be expressed on the surface of lower eukaryote host cells with no apparent loss of binding specificity and affinity (See for example, U.S. Patent No. 6,300,065). It has also been shown that full-length antibodies can be captured and bound to the surface of hybridomas and CHO cells, for example (See U.S. Patent Nos. 6,919,183 and 7,166,423 ). While antibodies and fragments thereof to many diverse antigens have been successfully isolated using phage display technology, there is still a need for a robust display system for producing antibodies and fragments thereof in lower eukaryotic host cell. It is particularly desirable to have a robust display system for producing antibodies and fragments thereof that have human-like glycosylation patterns. Genetically engineered lower eukaryotes that produce glycoproteins that have various human-like glycosylation patterns has been described in U.S. Patent No. 7,029,872 and for example in Choi et al., Hamilton, et al, Science 313; 1441 1443 (2006); Wildt and Gerngross, Nature Rev. 3: 119-128 (2005); Bobrowicz et al., GlycoBioL 757-766 (2004); Li et al., nature Biotechnol. 24: 210-215 (2006); Chiba et al., J. Biol. Chem. 273: 26298-26304 (1998); and, Mara et al., Glycoconjugate J. 16: 99- 107 (1999).

The subject display system is particularly suited for this application because the system allows presentation of a vast diverse repertoire of antibodies having particular glycosylation patterns. In many respects the subject display system mimics the natural immune system.

Antigen-driven stimulation can be achieved by selecting for high-affinity binders from a display library of cloned antibody H and L chains. The large number of chain permutations that occur during recombination of H and L chain genes in developing B cells can be mimicked by shuffling the cloned H and L chains as DNA, and protein and through the use of site-specific recombination (Geoffory et al. Gene 151 : 109-1 13 (1994)). The somatic mutation can also be matched by the introduction of mutations in the CDR regions of the H and L chains.

Antibodies or fragments thereof with desired binding specificity or affinity can be identified using a form of affinity selection known as "panning" (Parmley and Smith (1988) Gene 73:305-318). The library of Antibodies or fragments thereof is first incubated with an antigen of interest followed by the capture of the antigen with the bound antibodies or fragments thereof. The antibodies or fragments thereof recovered in this manner can then be amplified and again gain selected for binding to the antigen, thus enriching for those antibodies or fragments thereof that bind the antigen of interest. After one or more rounds of selection isolation will enable isolation of antibodies or fragments thereof with the desired specificity or avidity. Thus, rare host cells expressing a desired antibody or fragment thereof can easily be selected from greater than 104 different individuals in one experiment. The primary structure of the binding Antibody or fragment thereof is then deduced by nucleotide sequence of the individual host cell clone. When human VH and VL regions are employed in the displayed antibodies or fragments thereof, the subject display systems allow selection of human antibodies without further manipulation of a non-human antibodies or fragments thereof.

IX. Generating Novel Proteins Including Antibodies and fragments thereof with Improved Binding Specificity or Affinity

Using the subject display systems, one can obtain a replicable host cells that displays a polypeptide, such as an antibody or fragment thereof, having high affinity and specificity for a target protein. Such a host cells carries a first polynucleotide encoding the antibody or fragment thereof fused to a second adapter peptide and a second polynucleotide encoding the cell surface anchoring protein fused to a first adapter peptide that is capable of pairwise interaction with the second adapter peptide. The presence of the first polynucleotide facilitates recombinant expression and subsequent manipulation of the binding protein. For instance, the first polynucleotide can be mutagenized by cassette mutagenesis, error-prone PCR, or shuffling to generate a refined repertoire of altered sequences that resemble the parent polynucleotide. Upon screening the refined repertoire of novel antibodies or fragments thereof, those exhibiting improved binding specificity or affinity can be identified.

X. Mapping Antigenic Epitopes Traditionally, epitope mapping of an antigen has relied heavily on physical chemical analysis. These approaches have included: (1) fragmenting the purified antigen with various proteases, identifying reactive fragments, and sequencing them; (2) chemical modification experiments in which residues interaction with the antigen-binding unit are protected from modification; (3) synthesizing a series of peptides corresponding to the primary structure of the antigen; and (4) direct physical characterization using NMR or X-ray crystallography. All of these methods are labor intensive and generally not amenable to high-throughput analyses. Lower eukaryote display as disclosed herein provides a highly efficient and robust alternative for localizing the antigenic epitope. Fragments of DNA that encode portions of the antigen can be expressed as the exogenous polypeptides by the subject expression vectors. The lower eukaryote host cells can then be tested with the antibody to determine which displayed fragments react with the antibody. This application of display technology has been widely used in the art and has been shown to be successful for determining the antigenic epitopes of a variety of molecules.

XI. Mapping Binding Epitopes The subject display system also can be used to present random peptide libraries for mapping the specificity of the antigen-binding sites. Random peptide libraries represent a source of sequences from which epitopes and mimotopes can be operationally defined. With such a library, one can identity and obtain peptide competitors for antigen-antibody interactions, and thus map accessible and/or functional sites of numerous antibodies or fragments thereof.

XIL Kits Comprising the Vectors of the Present Invention

The present invention also encompasses kits containing the expression and helper vectors of this invention in suitable packaging. Each kit necessarily comprises the reagents which render the delivery of vectors into a host cell possible. The selection of reagents that facilitate delivery of the vectors may vary depending on the particular transfection or infection method used. The kits may also contain reagents useful for generating labeled polynucleotide probes or proteinaceous probes for detection of exogenous sequences and the protein product. Each reagent can be supplied in a solid form or dissolved/suspended in a liquid buffer suitable for inventory storage, and later for exchange or addition into the reaction medium when the experiment is performed. Suitable packaging is provided. The kit can optionally provide additional components that are useful in the procedure. These optional components include, but are not limited to, buffers, capture reagents, developing reagents, labels, reacting surfaces, means for detection, control samples, instructions, and interpretive information.

In the following examples, heterologous human proteins are expressed in host cells of the species Pichia pastof is. The following examples are intended to promote a further understanding of the present invention.

EXAMPLE 1

The objective was to develop a novel yeast display method especially designed for Pichia pastor is strains genetically engineered to produce glycoproteins with various mammalian glycosylation patterns. In this example, a nucleic acid encoding the N-terminus of a cell surface anchoring protein that inherently contains an attached glycophosphotidylinositol (GPI) post- trans lational modification that anchors the protein in the cell wall was linked to a nucleic acid that encodes a first coiled coil peptide that is capable of forming a heterodimer with a second coiled coil peptide fused to a test protein. The specific cell surface anchoring protein that was used was Sedl p, which had been identified by screening a panel of cell wall or plasma membrane proteins that had been identified using GPI protein prediction software.

Expression cassettes encoding the GPI protein and the test antibodies and Fab fragments were constructed using as the adapter peptides the coiled coil peptides GABAB-R2 (AEQ ID NO:19) fused to the N-terminus of the GPI protein and the GABAB-Rl (SEQ ID NO:21) fused to the C-terminus of the antibody or Fab fragment. GABAB-Rl (GRl ) and GABAB-R2 (GR2) are derived from the γ-Aminobutyric acid (GABA) receptors GABAB-Rl and GABAB-R2, Heterodimerization of GABAB-Rl and GABAB-R2 subunits is a prerequisite for the formation of a functional GABAB receptor. Each individual subunit contains one stretch of 30 amino acid residues within its intracellular C-terminal domain that mediates heterodimer formation. (Kammerer et al, J. Biochern. 38:13263-9 (1999)). Heterodimerization of a functional GABAB receptor is mediated by parallel coiled-coil alpha-helices. Three additional amino acid residues, GIy, GIy, and Cys were attached at the end of GRl. The Cys at the end of the GRl creates a disulfide bond with the Cys at the end of GR2, which is fused at the C-terminal of the display Fab fragment CHl. The two Glys are believed to increase the flexibility of the heterodimer. Construction of expression cassettes encoding the cell surface anchoring protein library was as follows. Candidate cell surface anchoring proteins were selected from S. cerevϊsiae, P. pastoris and H. polymorpha according to the literature and further identified as cell surface anchoring proteins using GPI protein prediction software available al IMP (Research Institute of Molecular Pathology), Bioinformalics Group, Dr. Bohr-Gasse 7, 1030 Vienna, Austria.. Ten proteins were selected for analysis.

Table 1 below shows the amino acid sequences for the relevant portion of ten GPI proteins and truncated variants of the proteins that were selected for analysis. Because highly expressed genes are desirable, truncation of the 3' end of the candidate nucleic acid sequences was made for several of the proteins in an attempt to improve expression. For all of the GPI proteins, the nucleic acid encoding the endogenous signal sequence for the GPI protein was removed. Therefore, the amino acid sequences shown in Table 1 do not include the amino acid sequences for the endogenous signal peptides. The bold-faced amino acids in the amino acid sequences shown in Table 1 signify the omega site. The omega site is the region at which GPI is attached to the protein. The GPI proteins were separated into two types based upon site of anchoring: GPI-anchored plasma membrane proteins (GPI-PMP) and GPI-dependent cell surface anchoring proteins (GPI-CWP).

Table 1

GPI Source Type Sequence SEQ ID protein NO:

CWP2 & cerevisiae CWP VDESAAAISQITDGQIQATTTATTEATTTAAP 9

SSTVETVSPS STETISQQTE NGAAKAAVGM

GAGALAAAAM LL

CWP2* S. cerevisiae CWP VDTTEATTTAAPSSTVETVSPSSTETISQQTE 10

Truncated NGAAKAAVGMGAGALAAAAMLL

SEDl S. cerevisiae CWP VDQFSNSTSASSTDVTSSSSISTSSGSVTITSS 11

EAPESDNGTSTAAPTETSTEAPTTAIPTNGTS

TEAPTTAIPTNGTSTEAPTDTTTEAPTTALPT

NGTSTEAPTDTTTEAPTTGLPTNGTTSAFPPT

TSLPPSNTTTTPPYNPSTDYTTDYTVVTEYTT

YCPEPTTFTTNGKTYTVTEPTTLTITDCPCTIE

KPTTTSTTEYTVVTEYTTYCPEPTTFTTNGK

TYTVTEPTTLTITDCPCTIEKSEAPESSVPVTE

SKGTTTKETGVTTKQTTANPSLTVSTVVPVS

SSASSHSVVINSNGANV VVPGALGLAGVAM

LFL

SEDl* S. cerevisiae CWP VDLTVSTVVPVSSSASSHSVVINSNGANVVV 12

Truncated PGALGLAGVAMLFL

SPIl P .pastoris CWP VDLVSNSSSSVIVVPSSDATIAGNDTATPAPE 13

PSSAAPIFYNSTATATQYEVVSEFTTYCPEPT

TFVTNGATFTVTAPTTLTITNCPCTIEKPTSET

SVSSTHDVETNSNAANARAIPGALGLAGAV MMLL

GASl S. cerevisiae PMP VDDVP AIEVVGNKFFYSNNGSQFYIRGV AY 14

QADTANETSGSTVNDPLANYESCSRDIPYLK KXNTNVIRVYAINTTLDHSECMKALNDADIY VIADLAAPATSINRDDPTWTVDLFNSYKTVV DTFANYTNVLGFFAGNEVTNNYTNTDASAF VKAAIRDVRQYISDKNYRKIPVGYSSNDDED TRVKMTDYFACGDDDVKADFYGINMYEWC GKSDFKTSGYADRTAEFKNLSIPVFFSEYGC NEVTPRLFTEVEALYGSNMTDVWSG GIVYMYFEET

NKYGLVSIDGNDVKTLDDFNNYSSEINKISPT SANTKSYSATTSDVACPATGKYWSAATELP PTPNGGLCSCMNAANSCVVSDDVDSDDYET LFNWICNEVDCSGISANGTAGKYGAYSFCTP KEQLSFVMNLYYEKSGGSKSDCSFSGSATL QTATTQASCSSALKEIGSMGTNSASGSVDLG SGTESSTASSNASGSSSKSNSGSSGSSSSSSSS SASSSSSSKKNAATNVKANLAOVVFTSHSLS IAAGVGFALV

GASl P. pastoris CWP VDADFPTIE VTGNKFF YSNNGSQF YIKGV AY 15

QKDTSGLSSDATFVDPLADKSTCERDIPYLE

ELGTNVIRVYAVDADADHDDCMQMLQDAG IYVIADLSQPNNSIITTDPEWTVDLYDGYTAV LDNLQKYDNILGFFAGNEVITNKSNTDTAPF

VKAAIRDMKTYMEDKGYRSIPVGYSANDDE LTRVASADYFACGDSDVKADFYGINMYEW CGKATFSNSGYKDRTAEFKNLSIPVFFSEYG CNEVQPRLFTEVQSLYGDDMTDVWSGGIVY

MYFEETNNYGLVTIKSDGDVSTLEDFNNLK

TELASISPSIATQSEVSATATEIDCPATGSNW

KASTDLPPVPEQAACQCMADALSCVVSEDV

DTDDYSDLFSYVCENVSSCDGVSADSESGE

YGSYSFCSSKEKLSFLLNLYYSENGAKSSAC

DFSGSATLVSGTTASECSSILSAAGTAGTGSI

TGITGSVEAATQSGSNSGSSKSSSASQSSSSN

AGVGGGASGSSWAMTGLVSISVALGMIMSF

GASl" P. pastoris CWP VDSILS AAGTAGTGSITGITGSVEAATQSGSN 16 Truncated SGSSKSSSASQSSSSNAGVGGGASGSSWAM

TGLVSISVALGMIMSF

TIPl H. CWP VDAAATSSVAAAASEVSSSSAAASSTQAAA 17 polymorpha AASTSAAASTEATTSAAAAATSSSEAASSSA HVHSHAAESTSAVESTSAAHSHAAESSSAA HSHAVESSSAAHVHSHAAESSSAAHSHAAG SSSAASNSSGHISTFSGAGAKLAVGAGAGIV GLAALLM

TIPl H. CWP VDSSAAHSHAVESSSAAHVHSHAAESSSAA 18 polymorpha HSHAAGSSSAASNSSGHISTFSGAGAKLAVG

Truncated AGAGIVGLAALLM

The nucleic acids encoding each of the anchoring proteins was codon-optimized according to Pichia pastoris codon usage. A nucleic acid encoding a valine and aspartic acid dipeptide (VD) was added to the 5' end of the nucleic acid encoding the proteins to create a Sail restriction site at the 5' end of the nucleic acid. The endogenous signal peptides of each of these GPI proteins was replaced with the Aspergillus niger alpha-amylase signal peptide. The DNA encoding the signal peptide is ATGGTTGCTT GGTGGTCCTT GTTCTTGTAC GGATTGCAAG TTGCTGCTCC AGCTTTGGCT (SEQ ID NO:33) and the signal peptide has the amino acid sequence MV AWWSLFLY GLQV AAPALA (SEQ ID NO:34). Further optimization of anchor protein expression and cell surface localization may be achieved through screening a library of N-lerminal signal peptides fused to the n-terminus of the anchoring proteins to identify signal peptides that best localize the GPI protein to the cell surface. For each construct, a nucleic acid encoding a GR2 coiled coil peptide having the amino acid sequence TSRLEGLQSE NHRLRMKITE LDKDLEEVTM QLQDVGGC (SEQ ID NO: 19) was inserted between the nucleic acid encoding the signal peptide and the nucleic acid encoding the GPI protein. The cassettes further included a nucleic acid encoding a myc epitope which was inserted between the nucleic acid encoding the GR2 coiled coil peptide and the GPI protein. The myc epitope is optional but had been included in the expression cassettes in order to provide an epitope to facilitate detecting the expressed GPI protein attached to the cell surface using a commercially available anti-myc antibody.

Figure 2 A shows an example of the S. cerevisaie SEDl GPI protein fused to the GR2 coiled coil peptide. The fusion protein consists of the Aspergillus niger alpha-amylase signal peptide followed by the GR2 coiled coil peptide followed by a Myc tag, and ending with the SEDl GFP anchor protein (without its endogenous signal peptide). Figure 2B shows the amino acid sequence of the fusion protein (SEQ ID NO:20). Figure 3 shows a representative plasmid map encoding the SEDl fusion protein. The SEDl fusion protein when expressed in the cell is transported to the surface of the cell where it is bound at the cell wall such that the GR2 coiled coil peptide is oriented extracellularly and rendered accessible to binding any protein in the extracellular environment that contains a GRl coiled coil peptide accessible to the GR2 coiled coil peptide. All of the above nucleic acid sequences were codon optimized according to Pichia pastoris codon usages and synthesized by Gene Art AG. Table 2 shows a representative number of plasmids containing cell surface anchoring expression cassettes in which the GPI protein was fused to GR2 that were constructed. The Pichia pastoris URA6 locus was chosen as an integrating site for the GPI anchoring protein expression cassettes. The URA6 gene was PCR amplified from Pichia pastoris genomic DNA and cloned into pCR2.1 TOPO to produce plasmid pGLY1849. The BgI2 and EcoRl sites within the gene were mutated by silent mutation for cloning purposes. The TRP2 targeting nucleic acid of plasmid pGLY2184 was replaced with the Pichia pastoris URA6 gene from pGLYl 849. In addition, the Pichia pastoris ARGl selection marker was replaced with the with Arsenite marker cassette from plasmid pGFI8. The final plasmid was named pGFBOt and was used to make the plasmids shown in Table 2.

Table 2

Plasmids Containing Cell Surface Anchoring Expression Cassettes

Plasmid Description pGLY3015 S. cerevisiae CWP 2-GKl fusion protein pGLY3033 S. cerevisiae SED1-GR2 fusion protein pGLY3034 S. cerevisiae SEDl truncated-GR2 fusion protein pGLY3035 P. pastoris SPIl-GKl fusion protein pGLY3036 P. pastoris GAS1-GK2 fusion protein pGLY3037 S. cerevisiae GAS1-GK2 fusion protein pGLY3038 S. cerevisiae GASl truncated-GR2 fusion protein pGLY3039 H. polymorpha TIPl -GKZ fusion protein

PGLY3040 Ii. polymorpha TIPl truncated-GR2 fusion protein

The antibody and Fab fragment expression cassettes were constructed as follows. Figure 4 illustrates three different antibody expression cassettes that have been constructed for producing antibodies or fragments thereof. Expression cassette A comprises two separately expressed open reading frames (ORFs). The first ORF encodes the light chain and the second ORF encodes a fusion protein comprising the Fd region of the heavy chain fused at the C- terminus to the GRl coiled coil peptide. Each ORF is operably linked to an AOXl promoter, which enables expression of the fusion proteins to be inducibly expressed. When expression is induced, this expression cassette is capable of producing an Fab fragment consisting of the light chain and Fd fragment fused at its C-terminus to a GRl coiled coil peptide. The Fab fragment can be captured by heterodimerization by the GR2 coiled coil peptide fused to the GPI protein, which is on the surface of the cell. Desired Fab fragments can then be detected by a suitable detection means. Figure 5 shows a plasmid map of a plasmid that was constructed in which expression cassette A encodes an Fab that is specific for Her2 antigen.

Expression cassette B is capable of producing a full-length antibody fused to a GRl coiled coil peptide. The first ORF encodes the light chain and the second ORF encodes a fusion protein comprising the heavy chain fused at the C-terminus to the GRl coiled coil protein. Each ORF is operably linked to an AOXl promoter. When expression is induced, this expression cassette is capable of producing full-length antibody consisting of the light chain and heavy chain fused at its C-terminus to a GRl coiled coil peptide. The full-length antibody can be captured by heterodimerization by the GR2 coiled coil peptide fused to the GPI protein, which is on the surface of the cell. Desired antibodies can then be detected by a suitable detection means.

The limitation of expression cassette B is that the full-length antibodies produced will always include the GRl coiled coil peptide fused to the heavy chain. This limitation may not be desirable for antibodies that are intended for therapeutic purposes. Thus, a new expression cassette must be constructed by isolating from the host cell that produces the desired antibody the nucleic acid that encodes the desired antibody and recloning the nucleic in an expression cassette that does not include the nucleic acid encoding the GRl coiled coli peptide and which, therefore, produces the full-length antibody without the GRl coiled coli peptide fused to the C-terminus of the heavy chain. To get around the limitation, expression cassette C was designed. Expression cassette C under appropriate conditions is capable of producing full-length antibodies that include the GRl coiled coil peptide fused to the heavy chain for selection of a desired full-length antibody; however, under production conditions, the expression cassette produces the desired antibody in which the heavy chain is not fused to the GRl coiled coil peptide. Thus, expression cassette C avoids the need to redone the nucleic acid encoding the desired antibody. In expression cassette C, the second ORF that encodes a fusion protein comprising the heavy chain fused at the C-terminus to the N-terminus of the GRl coiled coil peptide further includes a single stop codon between the end of the nucleic acid sequence encoding the heavy chain and the nucleic acid encoding the GRl coiled coil peptide, in which readthrough of the stop codon is inducible. Normally, stop codons signal the ribosome to terminate the decoding of an mRNA template. In yeast, inefficient termination will allow translation to continue; the frequency of read-through varies depending on the yeast strain and stop codon chosen. The cassette is designed with a stop codon in frame with the nucleic acid encoding the full length antibody and separating it from the nucleic acid encoding the coiled coil peptide GRl. therefore, under most conditions, translation of an mRNA transcribed from the expression cassette predominantly terminates at the single stop codon and thus results in production of a full-length antibody that is not fused to the GRl coiled coil peptide. However, in the presence of the antibiotic G418, translation readthrough through the stop codon is increased, which results in the production of full-length antibodies fused to GRl coiled coil peptide; however, even in the presence of the antibiotic, expression of full-length antibody not fused to the GRl coiled coil peptide is the predominant species. This proportional readthrough can reflect the expressability of the full-length antibody; by monitoring both the secreted full-length antibody and the full-length antibody fusion captured at the cell surface, one can screen for high producing host cells. Thus, in the presence of the antibiotic, a population of the full-length antibodies will include the heavy chain-GRl coiled coil peptide fusion protein. Therefore, when screening a library of antibodies for a desired antibody, the host cells are grown in the presence of the antibiotic. The full-length antibodies comprising the heavy chain GRl fusion protein are captured at the cell surface by heterodimerization to the GR2 coiled coil peptide fused to the GPI protein on the surface of the cell. Desired antibodies can then be detected by a suitable detection means. However, for production of full-length antibodies in which the heavy chain is not fused to the GRl coiled coil peptide, host cells that have been identified to produce the desired antibody are grown in the absence of the antibiotic. The premise behind expression cassette C can be adapted to produce Fab fragments that are not fused to the GRl coiled coil peptide. Figure 6 shows a map of a plasmid that was constructed in which expression cassette C encodes a full-length antibody that is specific for Her2. Table 3 shows representative number of plasmids that were constructed that contain expression cassettes encoding Fabs (A) or antibodies (B) fused to GRl. Also shown in Table 3 are plasmids comprising expression cassette C.

Table 3

Plasmid Containing Antibody or Fab Expression Cassettes Plasmid Cassette Type Description pGLY3028 A Anti-Her2 Fab-GRl fusion protein pGLY3915 A Anti-Her2 Fab-GRl fusion protein pGLY302ό A Anti-DKKl Fab-GRl fusion protein pGLY391ό A Anti-CD20, C2B8 Fab-GRl fusion protein pGLY3917 A Anti-CD20, Frame grafted Fab-GRl fusion protein pGLY3918 A Anti-CD20, Frame grafted Fab-GRl fusion protein pGLY3919 A Anti-CD20, Frame grafted Fab-GRl fusion protein pGLY3920 A Anti-CD20, Frame grafted Fab-GRl fusion protein pGLY3939 B Anti-Her2 full-length antibody-GRl fusion protein pGLY3941 C Anti-her2 full-length antibody-GRl fusion protein with single stop codon between antibody ORF and GRl ORF pGLY3942 C Anti CD20 C2B8 full length antibody-GRl fusion protein single stop codon between antibody ORF and GRl ORF pGLY3943 C Anti-CD20 Genmab antibody-GRl fusion protein single stop codon between antibody ORF and GRl ORF pGLY3944 C Anti-CD20 full length antibody-GRl fusion protein single stop codon between antibody ORF and GRl ORF Plasmids pGLY3028 and pGLY3915. The amino acid sequences for the heavy and light chains of the anti-her2 antibody are shown in SEQ ID NOs :22 and 23, respectively. The nucleic acid sequence encoding the anti-her2 Fab heavy chain fused to GRl and the ScαMTprepro signal sequence is shown in SEQ ID N0:51. The nucleic acid sequence encoding the anti-her2 light chain fused to the ScαMTprepro signal sequence (SEQ ID NO:49) is shown in SEQ ID NO:52.

Plasmid pGLY3926. The amino acid sequences for the heavy and light chains of the anti-DKKl antibody are shown in SEQ ID NOs:24 and 25, respectively. The nucleic acid sequence encoding the anti-DKKl Fab heavy chain fused to GRl and the Aspergillus niger alpha amylase signal sequence is shown in SEQ ID NO:53. The nucleic acid sequence encoding the anti-DKKl light chain fused to the Aspergillus niger alpha amylase signal sequence (SEQ ID NO:33) is shown in SEQ ID NO:54.

Plasmid pGLY3916. The amino acid sequences for the heavy and light chains of the anti-CD20 antibody are shown in SEQ ID NOs:26 and 27, respectively. The nucleic acid sequence encoding the anti-CD20, C2B8, Fab heavy chain fused to GRl and the Aspergillus niger alpha amylase signal sequence is shown in SEQ ID NO:55. The nucleic acid sequence encoding the anti-CD20, C2B8, light chain fused to the Aspergillus niger alpha amylase signal sequence is shown in SEQ ID NO: 56.

Plasmids pGLY3917-3920. The amino acid sequences for frame-grafted heavy and light chains of the anti-C20 Fab antibody are shown in SEQ ID NOs:28 and 29, respectively. The nucleic acid sequence encoding the anti-CD20, frame-grafted, Fab heavy chain fused to GRl and the Aspergillus niger alpha amylase signal sequence is shown in SEQ ID NO: 57. The nucleic acid sequence encoding the anti-CD20, frame-grafted, light chain fused to the Aspergillus niger alpha amylase signal sequence is shown in SEQ ID NO:58.

Plasmids pGLY3939 and 41. The amino acid sequences for the heavy and light chains of the anti-her2 antibody are shown in SEQ ID NOs:22 and 23, respectively. The nucleic acid sequence encoding the anti-her2 full length heavy chain fused to GRl and the ScαMTprepro signal sequence is shown in SEQ ID NO:59 (pGLY3939). The nucleic acid sequence encoding the anti-her2 full length heavy chain with single stop codon between the heavy chain-encoding ORF and GRl encoding ORF fused to GRl and the ScαMTprepro signal sequence is shown in SEQ ID NO:60 (pGLY3941). The nucleic acid sequence encoding the anti-her2 light chain fused to the ScαMTprepro signal sequence in both plasmids is shown in SEQ ID NO: 52.

Plasmid pGLY3942. The amino acid sequences for the heavy and light chains of the anti-CD20 antibody are shown in SEQ ID NOs:26 and 27, respectively. The nucleic acid sequence encoding the anti-CD-20, C2B8, full length heavy chain with single stop codon between the heavy chain-encoding ORF and GRl encoding ORF fused to GRl and the

Aspergillus niger alpha amylase signal sequence is shown in SEQ ID N0:61. The nucleic acid sequence encoding the anti-CD20, C2B8, light chain fused to the Aspergillus niger alpha amylase signal sequence is shown in SEQ ID NO: 56.

Plasmid pGLY3943. The amino acid sequences for Genmab heavy and light chains of the anti-CD20 antibody are shown in SEQ ID NOs:30 and 31, respectively. The nucleic acid sequence encoding the anti-CD-20, Genmab, full length heavy chain with single stop codon between the heavy chain-encoding ORF and GRl encoding ORF fused to GRl and the Aspergillus niger alpha amylase signal sequence is shown in SEQ ID NO:62. The nucleic acid sequence encoding the anti-CD20, Genmab, light chain fused to the Aspergillus niger alpha amylase signal sequence is shown in SEQ ID NO:63. Plasmid pGLY3944. The nucleic acid sequence encoding the anti-CD-20 full length heavy chain with single stop codon between the heavy chain-encoding ORF and GRl encoding ORF fused to GRl and the Aspergillus niger alpha amylase signal sequence is shown in SEQ ID NO:64. The nucleic acid sequence encoding the anti-CD20 light chain fused to the Aspergillus niger alpha amylase signal sequence is shown in SEQ ID NO:65.

Co-expression of Fab- and antibody-GRl fusion protein expression cassettes and GPI protein- GR2 fusion protein expression cassettes in yeast.

Two different methods were used for transforming the plasmids containing expression cassettes encoding the GPI protein-GR2 fusion proteins and Fab- or antibody-GRl fusion proteins into glycoengineered yeast.

In the first approach, plasmid vectors containing the GPI protein-GR2 fusion protein expression cassettes and containing a first selection marker is transformed into P. pastoris and plated on medium with the selection means to select for colonies carrying the GPI protein-GR2 expression cassettes. Then, colony PCR is used to screen the positive colonies for the presence of the GPI protein-GR2 fusion proteins. Finally, these cells are transformed with plasmids containing the Fab- or antibody-GRl fusion expression cassette and containing a gene for conferring a second selection marker and recombinant cells identified by growing the cells in the presence of a second selection means. In the second approach, the plasmids containing the antibody or Fab-GRl fusion protein expression cassettes are transformed first into the glycoengineered Pichia pastoris followed by transformation with plasmids containing the GPI protein-GR2 fusion protein expression cassettes.

Figure 3 shows plasmid pGLY3033, which is an example of a plasmid vector that contains a GPI protein- GR2 expression cassette and the S. cerevisiae ARR3 gene as the marker gene. The ARR3 gene from S. cerevisiae confers arsenite resistance to cells that are grown in the presence of arsenite (Bobrowicz et al., Yeast, 13:819-828 (1997); Wysocki el al, J. Biol. Chem. 272:30061-066 (1997)). Figures 5 and 6 show examples of plasmids that contain Fab-GRl (Figure 5) or antibody-GRl (Figure 6) fusion protein expression cassettes. The plasmids shown in Figures 5 and 6 containing the Fab-GRl or whole antibody-GRl fusion expression cassette also contain a nucleic acid homologous to a portion of the TRP 2 locus in Pichia pastoris to target the vector for integration into the TRP2 locus and a gene that confers resistance to the antibiotic Zeocin. When the vector is linearized within this nucleic acid, the plasmid vector is capable of single crossover homologous recombination into the TRP2 locus. Thus, the vectors shown in Figures 5 and 6 enable recombinant Pichia pastoris strains to be made with the Fab- or antibody- GRl fusion expression cassette integrated into the genome of the cell.

Table 4 shows a representative number of yeast strains that were made. All the strains were in a GS2.0 background. GS2.0 strains are glycoengiiieered Pichia pastoris strains that produce glycoproteins having predominantly Mans GIcN Ac2 iV-glycans (strains YGLY638 and YGLY2696. Strains that produce glycoproteins that have predominantly MansGlcNAc2 JV- glycans have been described in for example, U.S. Patent No. 7,029,872 and in Choi et al., Proc. Natl. Acad. Sci. USA 100: 5022-5027 (2003). Strain YGLY2696 is a GS2.0 strain that further has the gene encoding the endogenous chaperone protein PDI deleted and expresses a nucleic acid encoding a human PDI chaperone protein and further includes a nucleic acid encoding the human GRP94 protein inserted into the PEP4 locus (See Example 6 below).

Table 4 Yeast Strains

Strain Description

YGLY638 GS2.0 glycoengineered Pichia pastoris

YGLY2696 GS2.0 glycoengineered and humanized chaperones Pichia pastoris

YGLY2966 YGLY638/pGLY3026 - expresses anti-DKKl Fab

YGLY4105 YGLY638/pGLY3028 - expresses anti-Her2 Fab

YGLY4145 YGLY4102/pGLY3033 - expresses anti Her2 Fab and SEDl anchor

YGLY4146 YGLY2966/pGLY3033 - expresses anti-DKKl Fab and SEDl anchor

YGLY5079 YGLY2696/SEDlpGLY3033 #1 SEDl anchor

YGLY5147 YGLY269ό/pGLY3916 Patch #1 - expresses anti-CD20 Fab

YGLY5148 YGLY2696/pGLY3917 Patch #4 - expresses anti-CD20 Fab

YGLY5149 YGLY5079/pGLY3916 Patch #16 - expresses anti-CD20 Fab and SEDl anchor

YGLY515O YGLY5079/pGLY3916 Patch #18 - expresses anti-CD20 Fab and SEDl anchor

YGLY5151 YGLY5079/pGLY3917 Patch #19 -expresses anti-CD20 Fab and SEDl anchor

YGLY5152 YGLY5079/pGLY3917 Patch #20 - expresses anti-CD20 Fab and SEDl anchor

YGLY5153 YGLY5079/pGLY3918 Patch #22 - expresses anti-CD20 Fab and SEDl anchor

YGLY6693 YGLY5079/pGLY3918 Patch #23 - expresses anti-CD20 Fab and SEDl anchor

YGLY6694 YGLY5079/pGLY3919 Patch #25 - expresses anti-CD20 Fab and SEDl anchor The above Pichia pastoris strains are grown in 50 mL BMGY media until OD 600 = 2. The cells are washed three times with 1 M sorbitol and resuspended in 1 mL 1 M sorbitol. About 1 to 2 μg linearized plasmid are mixed with these competent cells. Transformation is performed with a BioRad electroporation apparatus using the manufacturer's program specific for electroporation of nucleic acids into Pichia pastoris. One mL recovery media is added to the cells, which are then plated out on MG with 300 μg/mL zeocin or YPG with 50 μg/mL arsenite.

Growth and induction of Fab displaying yeast.

Glycoengineered yeast transformed with both Fab-GRl fusion protein expression cassette and GPI protein-GR2 expression cassette was inoculated using 600 μL BMGY in a 96 deep well plate or 50 mL BMGY in a 250 mL shake flask for two days. The cells were collected by centrifugation and the supernatant was discarded. The cells are induced by incubation in 300 μL or 25 mL BMMY with Pmti-3 inhibitor overnight following the methods taught in WO2007/061631. Pmti-3 is 3-hydroxy-4-(2-phenylethoxy)benzaldehyde; 3-(l-ρhenylethoxy)-4- (2-phenylethoxy)-benzaldehyde, which as been described in U.S. Patent No. 7,105,554 and Published International Application No. WO 2007061631.

Induced cells were labeled with goat anti-human heavy and light chain (H+L) Alexa 488 conjugated antibody and viewed using fluorescence microscopy (as illustrated in Figure 7). After induction, 0.5-1 OD600 cells were collected by centrifugation in a 1.5-mL tube. The cells were rinsed twice with 1 mL PBS and 0.5 mL goat anti-human IgG (H+L)- Alexa 488 (1 :500 in 1% BSA in PBS) is added. Alternatively, fiuoresin labeled secondary antibody can be used to detect the antigen. The tubes were rotated for one hour at 37°C, centrifuged, and rinsed 3X with ImL PBS to remove the detection antibody. About 50-100 μL of PBS was added to the tube, the cells were mixed, and a 10 μL aliquot viewed with a fluorescence microscope and photographed (See Figure 7).

Following the above in which the expression cassette encoded the anti-Her2 Fab-GRl fusion protein, it was determined that of the nine GPI anchored proteins in the library, cells that expressed the full length Saccharomyces cerevisiae SEDJ had the most intense signal followed by S. cerevisiae CWP2 {See Figure 8A- J). Figures 15 A-D shows that YGLY5149, YGLY5152, YGLY6693, and YGLY6694 all expressed anti-CD20 Fab which was captured to the cell surface using GR2 fused to SEDl anchor. Thus, SED1-GR2 fusion protein was selected the cell surface anchoring protein for the remainder of the examples.

EXAMPLE 2 Expression levels of two different Fab-GRl fusion proteins displayed on the surface of glycoengineered Pichia pastoris correlated with the expression levels of their full length counterparts. Expression levels of Anti-Her2 full length monoclonal antibodies are generally five times greater than anti-DKKl full length monoclonal antibodies when both are expressed in glycoengineered Pichia pastoris. Pichia pastoris expressing full-length anti-Her2 antibodies can produce about 1.3 g/L of antibody whereas Pichia pastoris expressing full-length anti-DKKl antibodies produces about 200mg/L in 3 L fermentors. In this Example, anti-Her2 Fab-GRl fusion protein and anti-DKKl -GRl fusion protein Fab were expressed and displayed on the surface of glycoengineered Pichia pastoris strain 2.0 expressing the SED1-GR2 fusion protein as described in Example 1. The amino acid sequences of the anti-her2 heavy and light chains are shown in SEQ ID NOs:22 and 23, respectively. The amino acid sequences of the anti-DKKl heavy and light chains are shown in SEQ ID NOs:24 and 25, respectively.

To determine the expression levels of the two Fabs, cells were labeled with goat anti- Human H+L Alexa 488 and photographed according to the method described in the Example 2. Figure 9 shows the difference in fluorescence intensity between the anti-Her2 Fab and anti- DKKl Fab. The cells expressing the anti-Her2 Fab displayed a much stronger signal on the surface of the cells than the cells expressing the anti-DKKl Fab. In Figure 9 A shows a Pichia pastoris GS2.0 strain expressing both SED1-GR2 fusion protein and anti-DKKl Fab-GRl fusion protein. Figure 9B shows a Pichia pastoris GS2.0 strain expressing the anti-Her2 Fab . but not the SEDl -GR2 fusion protein. Figure 9C shows a Pichia pastoris GS2.0 strain expressing both SEDl -GR2 fusion protein and anti-Her2 Fab-GRl fusion protein. All these cells were labeled with anti-human H&L Alexa 488 and photographed using the same exposure time. Figure 9B clearly shows that without the GPI protein anchor, cells cannot display the Fab. Figures 9A-C also show that the intensity of the fluorescent signal reflects the expression level of the Fab: the weakly expressed anti-DKKl Fab had a weak signal and the higher expression antibody has a stronger signal. This result correlates with the expression levels observed for full-length anti- DKKl and anti-Her2 monoclonal antibodies.

EXAMPLE 3

Flow cytometry analysis was conducted using the cells expressing anti-Her2 Fab and anti-DKKl Fab displayed on the surface of the cells. Glycoengineered yeast displaying fluorescently labeled anti-Her2 or anti-DKKl Fab respectively were prepared as described in Examples 1 and 2. Controls were prepared in which both cell types were not labeled with the detection antibody. Using flow cytometry analysis, anti-Her2 Fab displaying cells were found to have a stronger fluorescence intensity compared to anti-DKKl Fab displaying cells and both cell types had a stronger signal compared to the signal produced in their corresponding unlabeled controls. In Figure 10_? the fluorescent intensities from these experiments were combined. The Figure shows the difference of fluorescence intensity between the anti-Her2 Fab displaying cells and the anti-DKKl Fab displaying cells and the same cells in the absence of detection label: anti- Her2 Fab displaying cells showed significantly higher fluorescence intensity than the anti-DKKl Fab displaying cells. These results are in congruence with fluorescence microscopy observations in Example 3.

Fluorescence-activated cell sorting (FACS) profile of a mixture of cells displaying anti- Her2 Fab (strain YGLY4145) and anti-DKKl Fab (strain YGLY4146) was performed as follows. The cells displaying anti-Her2 Fab and cells displaying anti-DKKl Fab were mixed together in the following ratios: 1:1, 1 :10, 1 : 100 and 1 : 1000. Cells were labeled with goat anti-human H+L Alexa 488 prior to mixing. Figure 11 shows that at a 1 :1 ratio there are two separate populations of cells visible: anti-Her2 Fab displaying cells and anti-DKKl Fab displaying cells. As the mixing ratio goes up, the florescent intensity of the anti-Her2 Fab population decreases. At the higher ratios of 1 : 100 and 1 : 1000, there are no longer two separate populations of cells visible. A second experiment was performed to gain better insight into cell diversity across the observed distribution of high to low levels of fluorescence. Anti-Her2 Fab and anti -DKKl Fab displaying cells (strains YGLY4145 and YGLY 4146, respectively) were mixed at a ratio of 1 :1 (See Figure 12A). Across the intensity spectrum, cell populations were isolated from five areas of decreasing fluorescence (Figure 12B: areas Cl through C5) and plated. For each population of cells, 96 colonies were analyzed by colony PCR using PCR primers specific for each antibody to determine the predominant Fab in the area. The results obtained confirmed that sorting for high- fluorescence signal will almost exclusively result in enrichment of high-expressing anti-Her2 Fab cells, whereas isolating for low-fluorescence will result in enrichment of low expressing cells, here anti-DKKl Fab expressing cells (Figure 12C).

EXAMPLE 4 This example illustrates the use of FACS to isolate and enrich for a population of high Fab producing cells from a larger population of low level Fab producing cells.

Fluorescently labeled anti-Her2 Fab and anti-DKKl Fab displaying cells were labeled, mixed at a ratio of 1 : 1000, and analyzed by flow cytometry. The cells of highest 1 % of fluorescence were isolated (far right of left histogram in Figure 13). The cells were plated out on selection media and incubated three to four days. The cells were then collected by washing the plate with BMGY media and re-induced with BMMY. The re-induced cells were labeled and subsequently sorted. This second round of sorting resulted in two distinct populations of cells (Figure 13, center histogram). Cells with the highest and lowest fluorescence were isolated and, as in the first round of sorting, grown, collected, induced, and labeled again. These two population were again analyzed using flow cytometry (Figure 13, histogram on right). Cells from extremes of high and low fluorescence intensity were then isolated and grown up. The higher fluorescence signal population gave rise to a typical anti-Her2 Fab fragment profile whereas the lower fluorescence intensity population displayed an anti-DKKl Fab fragment-profile very similar to that shown in Figure 12B. Thus, in three rounds of sorting, we were able to isolate and enrich out of a 1 : 1000 (anti-Her2 Fab fragment: anti-DKKl Fab fragment) dilution a distinct population of cells expressing high levels of anti-Her2 Fab fragments. These experiments clearly demonstrate the versatility and power of a cell-sorting based approach to isolate and enrich for particular population of Fab fragments. The methods herein can be used to isolate and enrich for cells expressing particular populations of antibodies.

EXAMPLE 5 This example illustrates surface display of full-length antibodies using the methods disclosed herein.

Figure 6 shows plasmid pGLY3941 which comprises an expression cassette encoding anti-Her2 antibody fused to GRl wherein there is a single stop codon inserted in frame after the last codon encoding the full-length anti-Her2 antibody and which can be used to a display full- length antibody on the yeast cell surface using stop codon read-through method as discussed in Example 1 for expression cassette C.

Pichia pastoris strain YGLY6724 containing pGLY3941 displays a full length anti- Her2 antibody-GRl coiled coil fusion protein when the protein is produced under conditions that results in translational readthrough of the stop codon {See SEQ ID 32). Pichia pastoris strain YGLY6722 containing pGLY3939 (no stop codon between the coding sequences for the Her2 antibody and the GRl peptide) also displays a full length anti-Her2 antibody-GRl coiled coil fusion. YGLY6724 was grown with increasing amounts of the antibiotic G418 in the medium. G418 inhibits translational termination, thereby increasing stop codon readthrough and increasing fluorescence intensity. To determine the expression levels of the two antibodies, cells were labeled with goat anti-Human H+L Alexa 488 and photographed according to the method described in Example 2. Figures 14A-F show by microscopy observation and FACS that anti- Her2 full length antibody can be displayed on the surface and detected using fluorescence and FACS analysis.

EXAMPLE 6 In strain YGLY2696, the gene encoding the endogenous PDI replaced with a nucleic acid molecule encoding the human PDI and a nucleic acid molecule encoding the human GRP94 protein inserted into the PEP4 locus. The strain was further engineered to alter the endogenous glycosylation pathway to produce glycoproteins that have predominantly Man5GlcNAc2 N- glycans. Strain YGLY2696 has been disclosed in co-pending Application Serial Nos. 61/066,409, filed 20 February 2008, and 61/188,723, filed 12 August 2008, both of which are incorporated herein in their entirety. This strain was shown to be useful for producing Immunoglobulins and for producing immunoglobulins that have reduced 0-glycosylation. Construction of strain yGLY2696 involved the following steps.

Construction of expression/integration plasmid vector pGLY642 comprising an expression cassette encoding the human PDI protein and nucleic acid molecules to target the plasmid vector to the Pichia pastor is PDIl locus for replacement of the gene encoding the Pichia pastoris PDIl with a nucleic acid molecule encoding the human PDI was as follows and is shown in Figure 8. cDNA encoding the human PDIl was amplified by PCR using the primers hPDI/UPl: 5¹ AGCGC TGACG CCCCC GAGGA GGAGG ACCAC 3' (SEQ ID NO:35) and hPDI/LP-PacI: 5' CCTTA ATTAA TTACA GTTCA TCATG CACAG CTTTC TGATC AT 3' (SEQ ID NO: 36), Pfu turbo DNA polymerase (Stratagene, La Jolla, CA)₅ and a human liver cDNA (BD Bioscience, San Jose, CA). The PCR conditions were 1 cycle of 95⁰C for two minutes, 25 cycles of 95⁰C for 20 seconds, 58°C for 30 seconds, and 72°C for 1.5 minutes, and followed by one cycle of 72°C for 10 minutes. The resulting PCR product was cloned into plasmid vector pCR2.1 to make plasmid vector pGLY618. The nucleotide and amino acid sequences of the human PDIl are shown in SEQ ID NOs:37 and 38, respectively.

The nucleotide and amino acid sequences of the Pichia pastoris PDIl are shown in SEQ ID NOs:39 and 40, respectively. Isolation of nucleic acid molecules comprising the Pichia pastoris PDIl 5' and 3' regions was performed by PCR amplification of the regions from Pichia pastoris genomic DNA. The 5' region was amplified using primers PB248: 5' ATGAA TTCAG GCCAT ATCGG CCATT GTTTA CTGTG CGCCC ACAGT AG 3' (SEQ ID NO: 41); PB249: 5' ATGTT TAAAC GTGAG GATTA CTGGT GATGA AAGAC 3¹ (SEQ ID NO: 42). The 3^r region was amplified using primers PB250: 5' AGACT AGTCT ATTTG GAGAC ATTGA CGGAT CCAC 3' (SEQ ID NO: 43); PB251: 5' ATCTC GAGAG GCCAT GCAGG CCAAC CACAA GATGA ATCAA ATTTT G-3' (SEQ ID NO: 44). Pichia pastoris strain NRRL-11430 genomic DNA was used for PCR ampliøcation. The PCR conditions were one cycle of 95°C for two minutes, 25 cycles of 95°C for 30 seconds, 55°C for 30 seconds, and 72°C for 2.5 minutes, and followed by one cycle of 72⁰C for 10 minutes. The resulting PCR fragments, PpPDIl (5') and PpPDIl (3'), were separately cloned into plasmid vector pCR2.1 to make plasmid vectors pGLY620 and pGLY617, respectively. To construct pGLY678_> DNA fragments PpARG3-5' and PpARG-3' of integration plasmid vector pGLY24, which targets the plasmid vector to Pichia pastoris ARG3 locus, were replaced with DNA fragments PpPDI (5^!) and PpPDI (3^*), respectively, which targets the plasmid vector pGLY678 to the PDIl locus and disrupts expression of the PDIl locus.

The nucleic acid molecule encoding the human PDI was then cloned into plasmid vector pGLY678 to produce plasmid vector pGLY642 in which the nucleic acid molecule encoding the human PDI was placed under the control of the Pichia pastoris GAPDH promoter (PpGAPDH). Expression/integration plasmid vector pGLY642 was constructed by Ii gating a nucleic acid molecule encoding the Saccharomyces cerevisiaβ alpha mating factor (MF) presequence signal peptide (ScαMFpre-signal peptide) having a Notl restriction enzyme site at the 5' end and a blunt 3' end and the expression cassette comprising the nucleic acid molecule encoding the human PDI released from plasmid vector pGLY618 withAfel and Pad to produce a nucleic acid molecule having a blunt 5' end and a Pad site at the 3' end into plasmid vector pGLY678 digested with Notl and Pad. The resulting integration/expression plasmid vector pGLY642 comprises an expression cassette encoding a human PDIl /ScαMFpre-signal peptide fusion protein operably linked to the Pichia pastoris promoter and nucleic acid molecule sequences to target the plasmid vector to the Pichia pastoris PDIl locus for disruption of the PDIl locus and integration of the expression cassette into the PDIl locus. Figure 2 illustrates the construction of plasmid vector pGLY642. The nucleotide and amino acid sequences of the ScαMFpre-signal peptide are shown in SEQ ID NOs:49 and 50, respectively.

Construction of expression/integration vector pGLY2233 encoding the human GRP94 protein was as follows and is shown in Figure 3. The human GRP94 was PCR amplified from human liver cDNA (BD Bioscience) with the primers hGRP94/UP 1 : 5'-AGCGC TGACG

ATGAA GTTGA TGTGG ATGGT ACAGT AG-3¹; (SEQ ID NO: 45); and hGRP94/LPl : 5'- GGCCG GCCTT ACAAT TCATC ATGTT CAGCT GTAGA TTC 3'; (SEQ ID NO: 46). The PCR conditions were one cycle of 95°C for two minutes, 25 cycles of 95⁰C for 20 seconds, 55°C for 20 seconds, and 72°C for 2.5 minutes, and followed by one cycle of 72⁰C for 10 minutes. The PCR product was cloned into plasmid vector pCR2.1 to make plasmid vector pGLY2216. The nucleotide and amino acid sequences of the human GRP94 are shown in SEQ ID NOs :47 and 48, respectively.

The nucleic acid molecule encoding the human GRP94 was released from plasmid vector pGLY221ό with Af el and Psel. The nucleic acid molecule was then Ii gated to a nucleic acid molecule encoding the ScαMPpre-signal peptide having Notl and blunt ends as above and plasmid vector pGLY2231 digested with Notl and Fsel carrying nucleic acid molecules comprising the Pichia pastoris PEP4 5' and 3' regions (PpPEP4-5' and PpPEP4-3' regions, respectively) to make plasmid vector pGLY2229. Plasmid vector pGLY2229 was digested with BgRl and Notl and a DNA fragment containing the PpPDIl promoter was removed from plasmid vector pGLY2187 with BgHl and Notl and the DNA fragment ligated into pGLY2229 to make plasmid vector pGLY2233. Plasmid vector pGLY2233 encodes the human GRP94 fusion protein under control of the Pichia pastoris PDI promoter and includes the 5' and 3' regions of the Pichia pastoris PEP 4 gene to target the plasmid vector to the P EP 4 locus of genome for disruption of the PEP 4 locus and integration of the expression cassette into the PEP 4 locus, Figure 3 illustrates the construction of plasmid vector pGLY2233.

Construction of plasmid vectors pGLYl 162, pGLY1896, and pGFI207t was as follows. All Trichoderma reesei α-l,2-mannosidase expression plasmid vectors were derived from pGFI165, which encodes the T. reesei α-l,2-mannosidase catalytic domain (See published International Application No. WO2007061631) fused to S. cerevisiae αMATpre signal peptide (ScαMPpre-signal peptide) herein expression is under the control of the Pichia pastoris GAP promoter and wherein integration of the plasmid vectors is targeted to the Pichia pastoris PROl locus and selection is using the Pichia pastoris URA5 gene. A map of plasmid vector pGFΪl 65 is shown in Figure 4.

Plasmid vector pGLYl 162 was made by replacing the GAP promoter in pGFI165 with the Pichia pastoris AOXl (PpAOXl) promoter. This was accomplished by isolating the PpAOXl promoter as an EcoRl (made blunt)-i?g/II fragment from pGLY2028_> and inserting into pGFI165 that was digested with Not! (made blunt) and BgRl. Integration of the plasmid vector is to the Pichia pastoris PROl locus and selection is using the Pichia pastoris URA5 gene. A map of plasmid vector pGLYl 162 is shown in Figure 5.

Plasmid vector pGLY1896 contains an expression cassette encoding the mouse α-l,2~ mannosidase catalytic domain fused to the S. cerevisiae MNN2 membrane insertion leader peptide fusion protein (See Choi et ah, Proc. Natl. Acad. Sci. USA 100: 5022 (2003)) inserted into plasmid vector pGFI165 (Figure 5). This was accomplished by isolating the GAPp- ScMNN2-mouse MNSI expression cassette from pGLY1433 digested with Xhoϊ (and the ends made blunt) and Pmel, and inserting the fragment into pGFI165 that digested with Pmeϊ. Integration of the plasmid vector is to the Pichia pastoris PROl locus and selection is using the Pichia pastoris URA5 gene. A map of plasmid vector pGLYl 896 is shown in Figure 4.

Plasmid vector pGFI207t is similar to pGLY1896 except that the URA5 selection marker was replaced with the S. cerevisiae ARR3 (ScARR3) gene, which confers resistance to arsenite. This was accomplished by isolating the ScARR3 gene from pGFIl 66 digested with Ascl and the Ascl ends made blunt) and BgUl, and inserting the fragment into pGLY1896 that digested with Spel and the Speϊ ends made blunt and BgUl. Integration of the plasmid vector is to the

Pichia pastoris PROl locus and selection is using the Saccharomyces cerevisiae ARR3 gene. A map of plasmid vector pGFI2007t is shown in Figure 4. The ARR3 gene from S. cerevisiae confers arsenite resistance to cells that are grown in the presence of arsenite (Bobrowicz et al, Yeast, 13:819-828 (1997); Wysocki et al, J. Biol. Chem. 272:30061-066 (1997)). Yeast transfections with the above expression/integration vectors were as follows.

Pichia pastoris strains were grown in 50 mL YPD media (yeast extract (1%), peptone (2%), dextrose (2%)) overnight to an OD of between about 0.2 to 6. After incubation on ice for 30 minutes, cells were pelleted by centrifugation at 2500-3000 rpm for 5minutes. Media was removed and the cells washed three times with ice cold sterile 1 M sorbitol before resuspending in 0.5 ml ice cold sterile 1 M sorbitol. Ten μL linearized DNA (5-20 μg) and 100 μL cell suspension was combined in an electroporation cuvette and incubated for 5 minutes on ice. Electroporation was in a Bio-Rad GenePulser Xcell following the preset Pichia pastoris protocol (2 kV, 25 μF, 200 Ω)_? immediately followed by the addition of 1 niL YPDS recovery media (YPD media plus 1 M sorbitol). The transfected cells were allowed to recover for four hours to overnight at room temperature (26° C) before plating the cells on selective media.

Generation of Cell Lines was as follows and is shown in Figure 6. The strain yGLY24- 1 (ura5A: :MET1 ochlA:;IacZ

mnn4LlA: ύacZ! MmSLC35A3 pnol Amnn4A::ϊacZ metl6&::lacZ), was constructed using methods described earlier (See for example, Nett and Gerngross, Yeast 20:1279 (2003); Choi et al, Proc. Natl. Acad. Sci. USA 100:5022 (2003); Hamilton et al, Science 301:1244 (2003)). The BMT2 gene has been disclosed in Mille et al., J. Biol. Chem. 283: 9724-9736 (2008) and U.S. Published Application No. 20060211085. The PNOl gene has been disclosed in U.S. Patent No. 7,198,921 and the mnn4Ll gene (also referred to as mnn4b) has been disclosed in U.S. Patent No. 7,259,007. The mnn4 refers to mnn4L2 or mnn4a. In the genotype, K1MNN2-2 is the Kluveromyces lactis GIcNAc transporter and MmSLC35A3 is the Mus musculus GIcNAc transporter. The URA5 deletion renders the yGLY24-l strain auxotrophic for uracil (See U.S. Published application No. 2004/0229306) and was used to construct the humanized chaperone strains that follow. While the various expression cassettes were integrated into particular loci of the Pichia pastoris genome in the examples herein, it is understood that the operation of the invention is independent of the loci used for integration. Loci other than those disclosed herein can be used for integration of the expression cassettes. Suitable integration sites include those enumerated in U.S. Published application No. 20070072262 and include homologs to loci known for Saccharomyces cerevisiae and other yeast or fungi.

Strains yGLY702 and yGLY704 were generated in order to test the effectiveness of the human PDIl expressed in Pichia pastoris cells in the absence of the endogenous Pichia pastoris PDI gene. Strains yGLY702 and yGLY704 (huPDI) were constructed as follows. Strain yGLY702 was generated by transfecting yGLY24-l with plasmid vector pGLY642 containing the expression cassette encoding the human PDI under control of the constitutive PpGAPDH promoter. Plasmid vector pGLY642 also contained an expression cassette encoding the Pichia pastoris URA5, which rendered strain yGLY702 prototrophic for uracil. The URA5 expression cassette was removed by counterselecting yGLY702 on 5-FOA plates to produce strain yGLY704 in which, so that the Pichia pastoris PDIl gene has been stably replaced by the human PDI gene and the strain is auxotrophic for uracil.

Strain yGLY733 was generated by transfecting with plasmid vector pGLY1162, which comprises an expression cassette that encodes the Trichoderma Reesei marmosidase (TrMNSl) operably linked to the Pichia pastoris AOXl promoter (PpAOXl-TrMNSl) and the Saccharomyces cerβvisiea αMAT pre signal sequence, into the PROl locus of yGLY704. This strain has the gene encoding the Pichia pastoris PDl replaced with the expression cassette encoding the human PDIl \ has the PpAOXl-TrMNSl expression cassette integrated into the PROl locus, and is a URA 5 auxotroph. The PpAOXl promoter allows overexpression when the cells are grown in the presence of methanol.

Strain yGLY762 was constructed by integrating expression cassettes encoding TrMNSl and mouse mannosidase IA (MuMNS IA)₇ each operably linked to the Pichia pastoris GAPDH promoter in plasmid vector pGFΪ207t into control strain yGLY733 at the 5' PROl locus UTR in Pichia pastoris genome. This strain has the gene encoding the Pichia pastoris PDl replaced with the expression cassette encoding the human PDIl, has the PpGAPDH-TrMNSl and PpGAPDH-MuMNSlA expression cassettes integrated into the PROl locus, and is a URA5 auxotroph. Strain yGLY2677 was generated by counterselecting yGLY762 on 5-FOA plates. This strain has the gene encoding the Pichia pastoris PDl replaced with the expression cassette encoding the human PDIl, has the PpAOXl-TrMNSl expression cassette integrated into the PROl locus, has the PpGAPH-TrMNSl and PpGAPDH-MuMNSlA expression cassettes integrated into the PROl locus, and is a URA5 prototroph. Strains yGLY2696 was generated by integrating plasmid vector pGLY2233, which encodes the human GRP94 protein, into the PEP4 locus. This strain has the gene encoding the Pichia pastoris PDl replaced with the expression cassette encoding the human PDIl, has the PpAOXl-TrMNSl expression cassette integrated into the PROl locus, has the PpGAPDH- TrMNSl and PpGAPDH-MuMNSlA expression cassettes integrated into the PROl locus, has the human GRP64 integrated into the P EP 4 locus, and is a URA5 prototroph. The genealogy of this chaperone-humanized strain is shown in Figure 16.

EXAMPLE 7

Construction of plasmid pGLY5107, pGLY5108 and pGLYSl 10 encoding various antibody heavy and light chains to make Fab fragments 1H23 and 1D05 (low and high affinity Fab fragments speciøc to PCSK9, Proprotein convertase subtilisin/kexin type 9) and anti-CD20 Fab fragment Genmab was as follows.

Fab display vector pGLY3958 (Figure 20) was constructed using Zeocin as a marker for Pichia transformation selection. Nucleic acid molecules encoding the IgGi CHl domain, linker, and GRl coiled coil peptide and the constant region of IgGj kappa light chain were codon optimized and synthesized by GeneArt according to Pichia pastoris codon usage. Both the nucleic acid molecules encoding the heavy chain and light chains are under Pichia AOXl promoter. Unique sites EcoRl and Xho 1 were made for different antibody variable regions cloning. Likewise, Pst\ and Kpn 1 sites were added between the AOXl promoter and the constant region of light chain to facilitate variable region of light chain cloning.

Nucleic acid molecules encoding the variable regions of the heavy and light chains of IDO 5, 1H23, and anti-CD20 Genmab were codon optimized, reverse translated, and synthesized by Gene Art based on their amino acid sequences. Nucleic acid molecules encoding the Aspergillus amylase signal sequence (SEQ ID NO:33) was added in-frame to the 5' end of the open reading frames encoding the 1H23 heavy and light chains (SEQ ID NO:70 and 72, respectively) during the gene synthesis. The open reading frame encoding the heavy chain also included the nucleotide sequence encoding GRl . Nucleic acid molecules encoding the

Saccharomyces cerevisiae mating factor pre-signal peptide (alpha-MAT-pro; SEQ ID NO:49) signal sequence was added in-frame to the 5' end of the open reading frames encoding the 1D05 heavy and light chains (SEQ ID NO:66 and 68, respectively) during the gene synthesis. The open reading frame encoding the heavy chain also included the nucleotide sequence encoding GRl . During synthesis, EcoRl site was introduced at the 5' of the nucleic acid molecules encoding the heavy chains and Pstl sites were introduced at the 5' ends of the nucleic acid molecules encoding the light chains. Xhol and Kpnl sites were created at the 3' ends of the heavy and light chains, respectively, using nucleic acid molecules encoding heavy chain and light chain constant regions conserved amino acids. The nucleic acid molecules encoding the variable regions of the heavy chain and light chains were cloned into pGLY3958 (Figure 21) at the same time using four-piece ligation. Plasmids pGLY5108 encoding the 1D05 heavy and light chains (Figure 21), pGLY5109 encoding the 1H23 heavy and light chains (Figure 22), and pGLY5107 encoding the light and heavy chains of anti-CD20 Genmab (Figure 23), Colony PCR, enzymatic digestion, and DNA sequencing were applied to confirm the identity of the created plasmids. The amino acid sequence of 1D05 heavy chain with signal peptide encoded by SEQ ΪD NO:66 is shown in SEQ ID NO:67. The amino acid sequence of 1D05 light chain with signal peptide encoded by SEQ ID NO:68 is shown in SEQ ID NO:69. The amino acid sequence of 1H23 heavy chain with signal peptide encoded by SEQ ID NO: 70 is shown in SEQ ID NO:71. The amino acid sequence of 1H23 light chain with signal peptide encoded by SEQ ID NO:72 is shown in SEQ ID NO:73. The amino acid sequence of PCSK9 is shown in SEQ ID NO:74. Yeast transformation for making 1D05, 1H23 and anti-CD20 Genmab Fab display strains were as follows. Plasmids pGLY5107, pGLYS 108 and pGLY5110 were linearized by Spel digestion at 37°C and linearization was confirmed by gel electrophoresis, DNA was precipitated down using standard procedure using cold ethanol. Grew Pichia host YGL Y5079 (expresses ScSEDl -GR2 fusion protein in YGLY2696) in 50 mL BMGY media overnight to a cell density of between 1-2 of ODgOO- Cells were washed three times with cold sterile water and

1 M sorbitol to render the cells competent for transformation. The linearized DNA was mixed with competent cells and shocked using the Bio-Rad electroporation machine. Then 1 mL recovery media was added to the shocked cells and the cells incubated at room temperature for 1 to 2 hours. Then the cells were plated on YPG plates with appropriate Zeocin concentration to select for transformants. The strains produced are shown in Table 5.

EXAMPLE 8

It has been reported for Saccharomyces cerevisiae that assembly of heavy and light chains expressed in yeast can be problematic. Therefore, the ratio of heavy chain to light chain in the Fab fragments displayed on the cell surface was measured to determine the intactness of the Fab fragments displayed on the cell surface.

Strain YGLY7762 (expresses 1D05 Fab fragment heavy and light chains) and strain YGLY7764 (expresses 1H23 Fab fragment heavy and light chains) were grown in 200 niL BMGY and expression induced in a Micro24 bioreactor according to the description of Micro24 cell culture and induction. Then remove about 20-40 uL of induced yeast culture, add 1 mL of blocking solution to the sample, centrifuge at 10,000 rpm for 30 seconds and wash the cell pellet three times with 1 mL blocking solution. Measure ODgOO ^d calculate the volume needed to get an OD600 of 1 in desired final volume. (Usually the final volume is about 200 uL). Blocking solution: 60 g BSA from Omni Pur, 200 mL 0.5 % Tween 2O₅ 200 mL 1 Ox PBS (from Omni Pur), and dH2θ up to two liters.

Anti-human IgG2 Fd biotin-conjugated antibody (CALTAG Laboratories, code #MH1522, lot#443408A: anti-heavy chain antibody) coupled with strepavidin Alexa Fluor 488 (2mg/mL, Invitrogen, lot#53729A) was used for detecting the displayed Fab via the Fd region of the heavy chain and anti-human kappa allophycocyanin-conjugated antibody (CALTAG

Laboratories, code# MH10515, lot #358897A: anti-light chain antibody) was used for detecting the light chain. In general, three uL of anti-heavy chain antibody was incubated with the cells at room temperature for 30 minutes on a rotator kept in the dark. Then the cells were washed four times with 3% BSA-0.05% Tween 20-PBS buffer. After this, three uL of Strepavidin Alexa Fluor 488 and 3 uL of anti-light chain antibody were added and the mixtures incubated at room temperature for 30 minutes on a rotator in the dark. Then, cells were wash three times with 3%BSA-O.O5% Tween 20-PBS buffer. The cells were analyzed by FACS. Fluorescent intensity of light chain and heavy chain (Fd) were plotted using FluoJo.

Flow cytometric analysis showed that displayed heavy chains corresponded with displayed light chains. This is shown in Figure 17A and 17B, which show that the cells properly assembled and displayed the heavy and light chains of the Fab fragment. In contrast to results reported for Saccharomyces cerevisiae, the nucleic acid molecules encoding the Fab fragments were integrated into a specific locus which results in constant expression f the heavy and light chains as opposed to be provided in autonomously replicating plasmids. Furthermore, the cells were grown under conditions that controlled O-glycosylation of the heavy and light chains, i.e., the presence of human chaperone proteins in place of host cell chaperone proteins and/or Pmti-3 inhibitor of O-glycosylation. The Pmti-3 inhibitor reduces the O-glycosylation occupancy, that is the number of total O-glycans on the Fab or antibody molecule. The cell further express a T. reesei alpha- 1 ,2-mannsodase catalytic domain linked to the Saccharomyces cerevisiea αMAT pre signal peptide to control the chain length of those O-glycans that are on the Fab or antibody molecule.

Cell culture and induction in Micro 24

Yeast display cells are grown in 200 mL BMGY medium in regular shake flask for two days at room temperature. The yeast culture is centrifuged and the spent supernatant is decanted. The remaining cell pellet is suspended in fresh induction media (see below for recipe) to an ODeoo of between 100 and 200 depending on the experiment. About 4.5 mL of the resulting culture is inoculated into a well of an Applikon Microreactor cassette and a gas-permeable, low evaporation adhesive membrane is used to seal the cassette. The induced cells are run using a constant agitation rate of 800 rpm with a pH set-point of 6.5. Each well is aerated with a continuous flow of lwm (4.5 mL/min). Under these conditions the culture will typically consume 2.5% methanol in about 16-20 hours. After 16-20 hours or when a dissolved oxygen spike is observed and additional bolus of l%-2.5% methanol will be added so the cells remain in an induction start. Once the desired length of induction is achieved the Microreactor is stopped and the culture can be removed from the well for labeling.

BMGY Medium

EXAMPLE 8

This example shows that the method can sort cells that display the antibody or Fab fragments of interest from cells that do not display the antibody of Fab of interest.

In a first experiment, Pichia pastoris cells engineered to display anti-CD20 Fab fragments (YGLY7761) were mixed with Pichia pastoris cells engineered to display anti-PC SK- 9 Fab fragments (YGLY7762).

Strains YGLY 7762, YGLY 7761, and YGLY 7764 were incubated at 24°C for 24 hours and expression induced in Micro24 with BMMY and PMT inhibitor as described previously for 18 hours. Induced cells were harvested and transferred into 50 mL tubes; centrifuged at 2500 rpm for five minutes at 4⁰C. Supernatant fractions were decanted and the pellets resuspended in 50 mL of blocking solution. The cells were pelleted as before and the cell pellet washed once more in 50 mL of blocking solution and cells pelleted. The pellet was resuspended in blocking solution and the OD 600 was adjusted with blocking solution to give about three OD units. Then the cells were mixed in a 1 : 1 ratio and then labeled sequentially with fluorophore-conjugated PCSK9 antigen (Alexa 647-conjugated) for one hour at room temperature and fluorophore-conjugated generic H+L antibody (Alexa Fluor488-conjugated) for

30 minutes at room temperature. Afterwards, the cells were washed and the flow cytometric profile was determined.

Figure 18 A, shows the FACS profile of anti-CD20 Fab displaying cells and anti-PCSK- 9 (1D05) Fab displaying cells when mixed at 1 : 1 ratio. The figure shows that method can separate the two different cell populations.

In Figure 18B₅ Pichia pastoris cells engineered to display high affinity anti-PCSK-9 Fab fragments (1D05) and Pichia pastoris cells engineered to display low affinity anti-PCSK-9 Fab fragments (1H23) were each separately labeled with fluorophore-conjugated antigen and the flow cytometric profile for each was determined. Figure 18B shows an overlay of the FACS profiles for high and low affinity Fab fragments displaying cells. The panel shows that the method can be used to sort cells on basis of affinity for an antigen.

EXAMPLE 9 This example shows that the method can sort cells that display the antibody or Fab fragments of interest from a majority of cells that do not display the antibody of Fab of interest.

In a first experiment, Pichia pastoris cells engineered to display anti-PCSK-9 Fab fragments were mixed with Pichia pastoris cells engineered to display anti-CD20 Fab fragments. The cell populations were mixed at ratios of 1 : 1 ,000; 1 : 10,000; and 1 : 100,000, Each ratio of cells was then labeled sequentially with fluorophore-conjugated PCSK9 antigen (Alexa 647- conjugated) for one hour at room temperature and fluorophore-conjugated generic H+L antibody (Alexa FIuor488 -conjugated) for 30 minutes at room temperature. Afterwards, the cells were washed and the flow cytometric profile was determined.

The cells from the area corresponding to the highest 1% fluorescence (area expected for the anti-PCSK-9 Fab fragments) were isolated. The cells were plated out on selection media and incubated three to four days. The cells were then collected by washing the plate with BMGY media and re-induced with BMMY. The re-induced cells were labeled and subsequently sorted. This first round of sorting resulted in two distinct populations of cells (Figure 19A).

For the 1:10,000 and l;100,000 dilutions, the cells with the highest fluorescence were isolated and, as in the first round of sorting, grown, collected, induced, and labeled again. These cells were again analyzed using flow cytometry (Figure 19A). For reference, Panel A of Figure 18 shows the FACS profile of anti-CD20 Fab fragment displaying cells and anti-PCSK-9 (1D05) Fab fragment displaying cells when mixed at 1 : 1 ratio. The results show that after two rounds of sorting, cell population enriched for Fab fragments specific for PCSK-9 can be prepared.

In a second experiment, Pichia pastoris cells engineered to display high affinity anti- PCSK-9 Fab fragments (1D05) were mixed with Pichia pastoris cells engineered to display low affinity anti-PCSK-9 Fab fragments (1H23). The cell populations were mixed at ratios of 1 : 10,000 and 1 : 100,000. The cells were labeled with fluorophore-conjugated PCSK9 antigen (Alexa 647-conjugated) for one hour at room temperature. The cells were washed and the flow cytometric profile was determined. The cells from the area corresponding to the highest 1% fluorescence (area expected for high affinity 1D05 Fab fragments were isolated). The cells were plated out on selection media and incubated three to four days. The cells were then collected by washing the plate with BMGY media and re-induced with BMMY. The re-induced cells were labeled and subsequently sorted. This first round of sorting resulted in two distinct populations of cells (Figure 19B). For the 1 : 10,000 and 1 ; 100,000 dilutions, the cells with the highest fluorescence were isolated and, as in the first round of sorting, grown, collected, induced, and labeled again. These cells were again analyzed using flow cytometry (Figure 19B). For reference, Figure 18B shows an overlay of the FACS profiles for high and low affinity Fab fragment displaying cells. The results show that cells that display a high affinity Fab fragments can be separated from a vast excess of cells displaying low affinity Fab fragments.

These experiments in this example clearly demonstrate the versatility and power of a cell-sorting based approach to isolate and enrich for particular population of antibody or Fab fragments. The methods herein can be used to isolate and enrich for cells expressing particular populations of antibodies or Fab fragments. BRIEF DESCRIPTION OF THE SEQUENCES

TGACCTAGCCCAGCAGTACGGCGTGCGCGGCTATC

CCACCATCAAGTTCTTCAGGAATGGAGACACGGCTT

CCCCCAAGGAATATACAGCTGGCAGAGAGGCTGAT

GACATCGTGAACTGGCTGAAGAAGCGCACGGGCCC

GGCTGCCACCACCCTGCCTGACGGCGCAGCTGCAG

AGTCCTTGGTGGAGTCCAGCGAGGTGGCCGTCATC

GGCTTCTTCAAGGACGTGGAGTCGGACTCTGCCAA

GCAGTTTTTGCAGGCAGCAGAGGCCATCGATGACA

TACCATTTGGGATCACTTCCAACAGTGACGTGTTCT

CCAAATACCAGCTCGACAAAGATGGGGTTGTCCTCT

TTAAGAAGTTTGATGAAGGCCGGAACAACTTTGAA

GGGGAGGTCACCAAGGAGAACCTGCTGGACTTTAT

CAAACACAACCAGCTGCCCCTTGTCATCGAGTTCAC

CGAGCAGACAGCCCCGAAGATTTTTGGAGGTGAAA

TCAAGACTCACATCCTGCTGTTCTTGCCCAAGAGTG

TGTCTGACTATGACGGCAAACTGAGCAACTTCAAA

ACAGCAGCCGAGAGCTTCAAGGGCAAGATCCTGTT

CATCTTCATCGACAGCGACCACACCGACAACCAGC

GCATCCTCGAGTTCTTTGGCCTGAAGAAGGAAGAGT

GCCCGGCCGTGCGCCTCATCACCTTGGAGGAGGAG

ATGACCAAGTACAAGCCCGAATCGGAGGAGCTGAC

GGCAGAGAGGATCACAGAGTTCTGCCACCGCTTCC

TGGAGGGCAAAATCAAGCCCCACCTGATGAGCCAG

GAGCTGCCGGAGGACTGGGACAAGCAGCCTGTCAA

GGTGCTTGTTGGGAAGAACTTTGAAGACGTGGCTTT

TGATGAGAAAAAAAACGTCTTTGTGGAGTTCTATGC

CCCATGGTGTGGTCACTGCAAACAGTTGGCTCCCAT

TTGGGATAAACTGGGAGAGACGTACAAGGACCATG

AGAACATCGTCATCGCCAAGATGGACTCGACTGCC

AACGAGGTGGAGGCCGTCAAAGTGCACGGCTTCCC

CACACTCGGGTTCTTTCCTGCCAGTGCCGACAGGAC

GGTCATTGATTACAACGGGGAACGCACGCTGGATG

GTTTTAAGAAATTCCTAGAGAGCGGTGGCCAAGAT

GGGGCAGGGGATGTTGACGACCTCGAGGACCTCGA

AGAAGCAGAGGAGCCAGACATGGAGGAAGACGAT

GACCAGAAAGCTGTGAAAGATGAACTGTAA human PDI DAPEEEDHVLVLRKSNFAEALAAHKYPPVEFHAPWC

GTTGAAGAAGAGGAAGAAGAAAAGAAGCCAAAGA

CTAAGAAGGTTGAAAAGACTGTTTGGGACTGGGAG

CTTATGAACGACATCAAGCCAATTTGGCAGAGACC

ATCCAAAGAGGTTGAGGAGGACGAGTACAAGGCTT

TCTACAAGTCCTTCTCCAAAGAATCCGATGACCCAA

TGGCTTACATCCACTTCACTGCTGAGGGTGAAGTTA

CTTTCAAGTCCATCTTGTTCGTTCCAACTTCTGCTCC

AAGAGGATTGTTCGACGAGTACGGTTCTAAGAAGT

CCGACTACATCAAACTTTATGTTAGAAGAGTTTTCA

TCACTGACGACTTCCACGATATGATGCCAAAGTACT

TGAACTTCGTTAAGGGTGTTGTTGATTCCGATGACT

TGCCATTGAACGTTTCCAGAGAGACTTTGCAGCAGC

ACAAGTTGTTGAAGGTTATCAGAAAGAAACTTGTTA

GAAAGACTTTGGACATGATCAAGAAGATCGCTGAC

GACAAGTACAACGACACTTTCTGGAAAGAGTTCGG

AACTAACATCAAGTTGGGTGTTATTGAGGACCACTC

CAACAGAACTAGATTGGCTAAGTTGTTGAGATTCCA

GTCCTCTCATCACCCAACTGACATCACTTCCTTGGA

CCAGTACGTTGAGAGAATGAAAGAGAAGCAGGACA

AAATCTACTTCATGGCTGGTTCCTCTAGAAAAGAGG

CTGAATCCTCCCCATTCGTTGAGAGATTGTTGAAGA

AGGGTTACGAGGTTATCTACTTGACTGAGCCAGTTG

ACGAGTACTGTATCCAGGCTTTGCCAGAGTTTGACG

GAAAGAGATTCCAGAACGTTGCTAAAGAGGGTGTT

AAGTTCGACGAATCCGAAAAGACTAAAGAATCCAG

AGAGGCTGTTGAGAAAGAGTTCGAGCCATTGTTGA

ACTGGATGAAGGACAAGGCTTTGAAGGACAAGATC

GAGAAGGCTGTTGTTTCCCAGAGATTGACTGAATCC

CCATGTGCTTTGGTTGCTTCCCAATACGGATGGAGT

GGTAACATGGAAAGAATCATGAAGGCTCAGGCTTA

CCAAACTGGAAAGGACATCTCCACTAACTACTACG

CTTCCCAGAAGAAAACTTTCGAGATCAACCCAAGA

CACCCATTGATCAGAGACATGTTGAGAAGAATCAA

AGAGGACGAGGACGACAAGACTGTTTTGGATTTGG

CTGTTGTTTTGTTCGAGACTGCTACTTTGAGATCCG

GTTACTTGTTGCCAGACACTAAGGCTTACGGTGACA

GAATCGAGAGAATGTTGAGATTGTCCTTGAACATTG

pre-signal peptide

(protein)

Fab Anti-Her2 ATGAGATTCCCATCCATCTTCACTGCTGTTTTGTTCG HC-GRl CTGCTTCTTCTGCTTTGGCTGAGGTTCAGTTGGTTGA fusion with ATCTGGAGGAGGATTGGTTCAACCTGGTGGTTCTTT Pre-pro α- GAGATTGTCCTGTGCTGCTTCCGGTTTCAACATCAA mating factor GGACACTTACATCCACTGGGTTAGACAAGCTCCAG signal peptide GAAAGGGATTGGAGTGGGTTGCTAGAATCTACCCA (ScαMTprepr ACTAACGGTTACACAAGATACGCTGACTCCGTTAA o) (DNA) GGGAAGATTCACTATCTCTGCTGACACTTCCAAGAA

CACTGCTTACTTGCAGATGAACTCCTTGAGAGCTGA

GGATACTGCTGTTTACTACTGTTCCAGATGGGGTGG

TGATGGTTTCTACGCTATGGACTACTGGGGTCAAGG

AACTTTGGTTACTGTTTCCTCCGCTTCTACTAAGGG

ACCATCTGTTTTCCCATTGGCTCCATCTTCTAAGTCT

ACTTCCGGTGGTACTGCTGCTTTGGGATGTTTGGTT

AAAGACTACTTCCCAGAGCCAGTTACTGTTTCTTGG

AACTCCGGTGCTTTGACTTCTGGTGTTCACACTTTCC

CAGCTGTTTTGCAATCTTCCGGTTTGTACTCTTTGTC

CTCCGTTGTTACTGTTCCATCCTCTTCCTTGGGTACT

CAGACTTACATCTGTAACGTTAACCACAAGCCATCC

AACACTAAGGTTGACAAGAAGGTTGAGCCAAAGTC

CTGTGGTGGTGGTGGTAGTGGAGGTGGTGGAAGTG

GTGGCGGTGGTTCTGCGGCCGCTTATCCATATGATG

TTCCAGACTACGCTGGAGGTCATCATCATCACCACC

ATCACCATCATGGTGGTGAAGAGAAGTCCAGATTG

TTGGAGAAAGAGAACAGAGAGTTGGAGAAGATCAT

CGCTGAGAAAGAAGAGAGAGTTTCCGAGTTGAGAC

ACCAATTGCAATCCGTTGGTGGTTGTTAATAG

Anti-Her2 LC ATGAGATTCCCATCCATCTTCACTGCTGTTTTGTTCG with Pre-pro CTGCTTCTTCTGCTTTGGCTGACATCCAAATGACTC α- mating AATCCCCATCTTCTTTGTCTGCTTCCGTTGGTGACAG factor signal AGTTACTATCACTTGTAGAGCTTCCCAGGACGTTAA peptide TACTGCTGTTGCTTGGTATCAACAGAAGCCAGGAAA (ScαMTprepr GGCTCCAAAGTTGTTGATCTACTCCGCTTCCTTCTTG O) (DNA) TACTCTGGTGTTCCATCCAGATTCTCTGGTTCCAGA

TGATGGTTTCTACGCTATGGACTACTGGGGTCAAGG

AACTTTGGTTACTGTTTCCTCCGCTTCTACTAAGGG

ACCATCTGTTTTCCCATTGGCTCCATCTTCTAAGTCT

ACTTCCGGTGGTACTGCTGCTTTGGGATGTTTGGTT

AAAGACTACTTCCCAGAGCCAGTTACTGTTTCTTGG

AACTCCGGTGCTTTGACTTCTGGTGTTCACACTTTCC

CAGCTGTTTTGCAATCTTCCGGTTTGTACTCTTTGTC

CTCCGTTGTTACTGTTCCATCCTCTTCCTTGGGTACT

CAGACTTACATCTGTAACGTTAACCACAAGCCATCC

AACACTAAGGTTGACAAGAAGGTTGAGCCAAAGTC

CTGTGACAAGACTCATACTTGTCCACCATGTCCAGC

TCCAGAATTGTTGGGTGGTCCTTCCGTTTTTTTGTTC

CCACCAAAGCCAAAGGACACTTTGATGATCTCCAG

AACTCCAGAGGTTACATGTGTTGTTGTTGACGTTTC

TCACGAGGACCCAGAGGTTAAGTTCAACTGGTACG

TTGACGGTGTTGAAGTTCACAACGCTAAGACTAAGC

CAAGAGAGGAGCAGTACAACTCCACTTACAGAGTT

GTTTCCGTTTTGACTGTTTTGCACCAGGATTGGTTGA

ACGGAAAGGAGTACAAGTGTAAGGTTTCCAACAAG

GCTTTGCCAGCTCCAATCGAAAAGACTATCTCCAAG

GCTAAGGGTCAACCAAGAGAGCCACAGGTTTACAC

TTTGCCACCATCCAGAGATGAGTTGACTAAGAACC

AGGTTTCCTTGACTTGTTTGGTTAAGGGATTCTACC

CATCCGACATTGCTGTTGAATGGGAGTCTAACGGTC

AACCAGAGAACAACTACAAGACTACTCCACCTGTT

TTGGACTCTGACGGTTCCTTTTTCTTGTACTCCAAGT

TGACTGTTGACAAGTCCAGATGGCAACAGGGTAAC

GTTTTCTCCTGTTCCGTTATGCATGAGGCTTTGCACA

ACCACTACACTCAAAAGTCCTTGTCTTTGTCCCCTG

GTAAGGCGGCCGCTTATCCATATGATGTTCCAGACT

ACGCTGGAGGTCATCATCATCACCACCATCACCATC

ATGGTGGTGAAGAGAAGTCCAGATTGTTGGAGAAA

GAGAACAGAGAGTTGGAGAAGATCATCGCTGAGAA

AGAAGAGAGAGTTTCCGAGTTGAGACACCAATTGC

AATCCGTTGGTGGTTGTTAATAG

Anti-Her2 full ATGAGATTCCCATCCATCTTCACTGCTGTTTTGTTCG length HC CTGCTTCTTCTGCTTTGGCTGAGGTTCAGTTGGTTGA with single ATCTGGAGGAGGATTGGTTCAACCTGGTGGTTCTTT stop codon GAGATTGTCCTGTGCTGCTTCCGGTTTCAACATCAA between Ab GGACACTTACATCCACTGGGTTAGACAAGCTCCAG ORF and GRl GAAAGGGATTGGAGTGGGTTGCTAGAATCTACCCA ORF with ACTAACGGTTACACAAGATACGCTGACTCCGTTAA Pre-pro α- GGGAAGATTCACTATCTCTGCTGACACTTCCAAGAA mating factor CACTGCTTACTTGCAGATGAACTCCTTGAGAGCTGA signal peptide GGATACTGCTGTTTACTACTGTTCCAGATGGGGTGG (ScαMTprepr TGATGGTTTCTACGCTATGGACTACTGGGGTCAAGG o) (DNA) AACTTTGGTTACTGTTTCCTCCGCTTCTACTAAGGG

ACCATCTGTTTTCCCATTGGCTCCATCTTCTAAGTCT

ACTTCCGGTGGTACTGCTGCTTTGGGATGTTTGGTT

AAAGACTACTTCCCAGAGCCAGTTACTGTTTCTTGG

AACTCCGGTGCTTTGACTTCTGGTGTTCACACTTTCC

CAGCTGTTTTGCAATCTTCCGGTTTGTACTCTTTGTC

CTCCGTTGTTACTGTTCCATCCTCTTCCTTGGGTACT

CAGACTTACATCTGTAACGTTAACCACAAGCCATCC

AACACTAAGGTTGACAAGAAGGTTGAGCCAAAGTC

CTGTGACAAGACTCATACTTGTCCACCATGTCCAGC

TCCAGAATTGTTGGGTGGTCCTTCCGTTTTTTTGTTC

CCACCAAAGCCAAAGGACACTTTGATGATCTCCAG

AACTCCAGAGGTTACATGTGTTGTTGTTGACGTTTC

TCACGAGGACCCAGAGGTTAAGTTCAACTGGTACG

TTGACGGTGTTGAAGTTCACAACGCTAAGACTAAGC

CAAGAGAGGAGCAGTACAACTCCACTTACAGAGTT

GTTTCCGTTTTGACTGTTTTGCACCAGGATTGGTTGA

ACGGAAAGGAGTACAAGTGTAAGGTTTCCAACAAG

GCTTTGCCAGCTCCAATCGAAAAGACTATCTCCAAG

GCTAAGGGTCAACCAAGAGAGCCACAGGTTTACAC

TTTGCCACCATCCAGAGATGAGTTGACTAAGAACC

AGGTTTCCTTGACTTGTTTGGTTAAGGGATTCTACC

CATCCGACATTGCTGTTGAATGGGAGTCTAACGGTC

AACCAGAGAACAACTACAAGACTACTCCACCTGTT

TTGGACTCTGACGGTTCCTTTTTCTTGTACTCCAAGT

TGACTGTTGACAAGTCCAGATGGCAACAGGGTAAC

GTTTTCTCCTGTTCCGTTATGCATGAGGCTTTGCACA

ACCACTACACTCAAAAGTCCTTGTCTTTGTCCCCTG GTAAGTAGGCGGCCGCTTATCCATATGATGTTCCAG ACTACGCTGGAGGTCATCATCATCACCACCATCACC

ATCATGGTGGTGAAGAGAAGTCCAGATTGTTGGAG

AAAGAGAACAGAGAGTTGGAGAAGATCATCGCTGA

GAAAGAAGAGAGAGTTTCCGAGTTGAGACACCAAT

TGCAATCCGTTGGTGGTTGTTAATAGGGCCGGCCAT

TTAA

Anti-CD20 ATGGTTGCTTGGTGGTCTTTGTTCTTGTACGGATTGC C2B8 foil AAGTTGCTGCTCCAGCTTTGGCTcaagttcagctgcaacaacca length HC ggtgctgaattggttaagcctggtgcttctgltaagatgtcttgtaaggcttctggttacac with stop tttcactlcctacaacatgcactgggttaagcaaactccaggtagaggattggaatggat codon tggtgctatctacccaggtaacggtgacacttcttataaccaaaagttcaagggaaagg between Ab ctactttgactgctgacaaatcttcttctactgcttacatgcaattgtcctccltgacltctga ORF and GRl agattctgctgtttactactgtgctagatccacttactacggtggtgactggtactttaatgt ORF with ttggggtgctggtactactgttactgtctcgagtgcttctactaagggaccatctgttttcc Alpha cattggctccatctlctaagtctacttccggtggtacCGCTGCTTTGGGAT amylase GTTTGGTTAAAGACTACTTCCCAGAGCCAGTTACTG signal peptide TTTCTTGGAACTCCGGTGCTTTGACTTCTGGTGTTCA (from CACTTTCCCAGCTGTTTTGCAATCTTCCGGTTTGTAC

Aspergillus TCTTTGTCCTCCGTTGTTACTGTTCCATCCTCTTCCT niger α- TGGGTACTCAGACTTACATCTGTAACGTTAACCACA amylase) AGCCATCCAACACTAAGGTTGACAAGAAGGTTGAG (DNA) CCAAAGTCCTGTGACAAGACTCATACTTGTCCACCA

TGTCCAGCTCCAGAATTGTTGGGTGGTCCTTCCGTT

TTTTTGTTCCCACCAAAGCCAAAGGACACTTTGATG

ATCTCCAGAACTCCAGAGGTTACATGTGTTGTTGTT

GACGTTTCTCACGAGGACCCAGAGGTTAAGTTCAA

CTGGTACGTTGACGGTGTTGAAGTTCACAACGCTAA

GACTAAGCCAAGAGAGGAGCAGTACAACTCCACTT

ACAGAGTTGTTTCCGTTTTGACTGTTTTGCACCAGG

ATTGGTTGAACGGAAAGGAGTACAAGTGTAAGGTT

TCCAACAAGGCTTTGCCAGCTCCAATCGAAAAGAC

TATCTCCAAGGCTAAGGGTCAACCAAGAGAGCCAC

AGGTTTACACTTTGCCACCATCCAGAGATGAGTTGA

CTAAGAACCAGGTTTCCTTGACTTGTTTGGTTAAGG

GATTCTACCCATCCGACATTGCTGTTGAATGGGAGT

CTAACGGTCAACCAGAGAACAACTACAAGACTACT CCACCTGTTTTGGACTCTGACGGTTCCTTTTTCTTGT ACTCCAAGTTGACTGTTGACAAGTCCAGATGGCAA

CAGGGTAACGTTTTCTCCTGTTCCGTTATGCATGAG

GCTTTGCACAACCACTACACTCAAAAGTCCTTGTCT

TTGTCCCCTGGTAAGTAGGCGGCCGCTTATCCATAT

GATGTTCCAGACTACGCTGGAGGTCATCATCATCAC

CACCATCACCATCATGGTGGTGAAGAGAAGTCCAG

ATTGTTGGAGAAAGAGAACAGAGAGTTGGAGAAGA

TCATCGCTGAGAAAGAAGAGAGAGTTTCCGAGTTG

AGACACCAATTGCAATCCGTTGGTGGTTGTTAATAG

Anti-CD20 ATGGTTGCTTGGTGGTCCTTGTTCTTGTACGGATTGC Genmab full AAGTTGCTGCTCCAGCTTTGGCTgctgttcagctggttgaatctg length HC gtggtggattggttcaacctggtagatccttgagattgtcctgtgctgcttccggttttact with single ttcggtgactacactatgcactgggttagacaagctccaggaaagggattggaatggg stop codon tttccggtatttcttggaactccggttccattggttacgctgattccgttaagggaagattc between Ab actatctccagagacaacgctaagaactccttgtacttgcagatgaactccttgagagct ORF and GRl gaggatactgctttgtactactgtactaaggacaaccaatacggttctggttccacttac ORF with ggattgggagtttggggacagggaactttggttactgtctcgagtgcttctactaaggg Alpha accatccgtttttccattggctccatcctctaagtctacttccggtggtacCGCTGC amylase TTTGGGATGTTTGGTTAAAGACTACTTCCCAGAGCC signal peptide AGTTACTGTTTCTTGGAACTCCGGTGCTTTGACTTCT (from GGTGTTCACACTTTCCCAGCTGTTTTGCAATCTTCCG

Aspergillus GTTTGTACTCTTTGTCCTCCGTTGTTACTGTTCCATC niger α- CTCTTCCTTGGGTACTCAGACTTACATCTGTAACGT amylase) TAACCACAAGCCATCCAACACTAAGGTTGACAAGA (DNA) AGGTTGAGCCAAAGTCCTGTGACAAGACTCATACTT

GTCCACCATGTCCAGCTCCAGAATTGTTGGGTGGTC

CTTCCGTTTTTTTGTTCCCACCAAAGCCAAAGGACA

CTTTGATGATCTCCAGAACTCCAGAGGTTACATGTG

TTGTTGTTGACGTTTCTCACGAGGACCCAGAGGTTA

AGTTCAACTGGTACGTTGACGGTGTTGAAGTTCACA

ACGCTAAGACTAAGCCAAGAGAGGAGCAGTACAAC

TCCACTTACAGAGTTGTTTCCGTTTTGACTGTTTTGC

ACCAGGATTGGTTGAACGGAAAGGAGTACAAGTGT

AAGGTTTCCAACAAGGCTTTGCCAGCTCCAATCGAA

AAGACTATCTCCAAGGCTAAGGGTCAACCAAGAGA

GCCACAGGTTTACACTTTGCCACCATCCAGAGATGA

ORF with ccgaggacactgctgtttactactgtgctagatccacttactacggtggtgactggtactt Alpha taatgtttggggacagggaactttggttactgtctcgagtgcttctactaagggaccatc amylase cgtttttccattggctccatcctctaagtctacttccggtggtacCGCTGCTTTG signal peptide GGATGTTTGGTTAAAGACTACTTCCCAGAGCCAGTT (from ACTGTTTCTTGGAACTCCGGTGCTTTGACTTCTGGT

Aspergillus GTTCACACTTTCCCAGCTGTTTTGCAATCTTCCGGTT niger α- TGTACTCTTTGTCCTCCGTTGTTACTGTTCCATCCTC amylase) TTCCTTGGGTACTCAGACTTACATCTGTAACGTTAA (DNA) CCACAAGCCATCCAACACTAAGGTTGACAAGAAGG

TTGAGCCAAAGTCCTGTGACAAGACTCATACTTGTC

CACCATGTCCAGCTCCAGAATTGTTGGGTGGTCCTT

CCGTTTTTTTGTTCCCACCAAAGCCAAAGGACACTT

TGATGATCTCCAGAACTCCAGAGGTTACATGTGTTG

TTGTTGACGTTTCTCACGAGGACCCAGAGGTTAAGT

TCAACTGGTACGTTGACGGTGTTGAAGTTCACAACG

CTAAGACTAAGCCAAGAGAGGAGCAGTACAACTCC

ACTTACAGAGTTGTTTCCGTTTTGACTGTTTTGCACC

AGGATTGGTTGAACGGAAAGGAGTACAAGTGTAAG

GTTTCCAACAAGGCTTTGCCAGCTCCAATCGAAAAG

ACTATCTCCAAGGCTAAGGGTCAACCAAGAGAGCC

ACAGGTTTACACTTTGCCACCATCCAGAGATGAGTT

GACTAAGAACCAGGTTTCCTTGACTTGTTTGGTTAA

GGGATTCTACCCATCCGACATTGCTGTTGAATGGGA

GTCTAACGGTCAACCAGAGAACAACTACAAGACTA

CTCCACCTGTTTTGGACTCTGACGGTTCCTTTTTCTT

GTACTCCAAGTTGACTGTTGACAAGTCCAGATGGCA

ACAGGGTAACGTTTTCTCCTGTTCCGTTATGCATGA

GGCTTTGCACAACCACTACACTCAAAAGTCCTTGTC

TTTGTCCCCTGGTAAGTAGGCGGCCGCTTATCCATA

TGATGTTCCAGACTACGCTGGAGGTCATCATCATCA

CCACCATCACCATCATGGTGGTGAAGAGAAGTCCA

GATTGTTGGAGAAAGAGAACAGAGAGTTGGAGAAG

ATCATCGCTGAGAAAGAAGAGAGAGTTTCCGAGTT

GAGACACCAATTGCAATCCGTTGGTGGTTGTTAATA

G

Anti-CD20 ATGGTTGCTTGGTGGTCCTTGTTCTTGTACGGATTGC LC with AAGTTGCTGCTCCAGCTTTGGCTgagatcgttttgacacagtccc Alpha cagctactttgtctttgtccccaggtgaaagagctacattgtcctgtagagcttcctcttcc

Saccharomyce TFGQGTKVEIKRTVAAPSVFIFPPSDEQLKSGTASVVC s cerevisiae LLNNFYPREAKVQWKVDNALQSGNSQESVTEQDSKD mating factor STYSLSSTLTLSKADYEKHKVYACEVTHQGLSSPVTKS pre-signal FNRGEC peptide

DNA ATGGTTGCTTGGTGGTCCTTGTTCTTGTACGGATTGC sequence of AAGTTGCTGCTCCAGCTTTGGCTCAAGTTCAGTTGG 1H23 heavy TTGAATCCGGTGGTGGATTGGTTCAACCTGGTGGTT chain with CTTTGAGATTGTCCTGTGCTGCTTCCGGTTTTACTTT Aspergillus CTCCGACTACTACATGCACTGGGTTAGACAAGCACC amylase TGGAAAGGGATTGGAATGGGTTTCCAACATTTCTGG signal TTCCGGTTCCACTACTTACTACGCTGATTCCGTTAA sequence, GGGAAGATTCACTATCTCCAGAGACAACTCCAAGA linker and ACACTTTGTACTTGCAGATGAACTCCTTGAGAGCTG GRl AGGATACTGCTGTTTACTACTGTGCTAGAGGAATGT

TTGACTTCTGGGGACAGGGAACTTTGGTTACTGTCT

CGAGTGCTTCTACTAAGGGGCCCTCTGTTTTTCCAT

TGGCTCCATGTTCTAGATCCACTTCCGAATCCACTG

CTGCTTTGGGATGTTTGGTTAAGGACTACTTCCCAG

AGCCAGTTACTGTTTCTTGGAACTCCGGTGCTTTGA

CTTCTGGTGTTCACACTTTCCCAGCTGTTTTGCAATC

TTCCGGTTTGTACTCCTTGTCCTCCGTTGTTACTGTT

ACTTCCTCCAACTTCGGTACTCAGACTTACACTTGT

AACGTTGACCACAAGCCATCCAACACTAAGGTTGA

CAAGACTGTTGAGAGAAAGGGTGGTGGTGGTAGTG

GAGGTGGTGGAAGTGGTGGCGGTGGTTCTGCGGCC

GCTTATCCATATGATGTTCCAGACTACGCTGGAGGT

CATCATCATCACCACCATCACCATCATGGTGGTGAA

GAGAAGTCCAGATTGTTGGAGAAAGAGAACAGAGA

GTTGGAGAAGATCATCGCTGAGAAAGAAGAGAGAG

TTTCCGAGTTGAGACACCAATTGCAATCCGTTGGTG

GTTGTTAATAG

Amino acid MVAWWSLFLYGLOVAAPALAOVOLVESGGGLVQPG sequence of GSLRLSCAASGFTFSDYYMHWVRQAPGKGLEWVSNI 1H23 HC SGSGSTTYYADSVKGRFTISRDNSKNTLYLQMNSLRA with EDTAVYYCARGMFDFWGQGTLVTVSSASTKGPSVFP

While the present invention is described herein with reference to illustrated embodiments, it should be understood that the invention is not limited hereto. Those having ordinary skill in the art and access to the teachings herein will recognize additional modifications and embodiments within the scope thereof. Therefore, the present invention is limited only by the claims attached herein.

Claims

WHAT IS CLAMED:

1. A method for selecting proteins for displayability on a lower eukaryote host cell surface, comprising: (a) providing a host cell that expresses a capture moiety comprising a cell surface anchoring protein fused to a first binding moiety;

(b) transforming the host cell with a nucleic acid encoding a protein fused to a second binding moiety that is capable of specifically interacting with the first binding moiety fused to the cell surface anchoring protein, wherein mutagenesis is used to generate a plurality of host cells encoding a variegated population of mutants of the proteins;

(c) contacting the plurality of host cells with a detection means that specifically binds to proteins that are displayed on the surface of the host cell and does not bind to proteins that are not displayed on the surface of the host cell; and

(d) isolating the host cells with which the detection means is bound, wherein the presence of the detection means bound to a protein on the surface of the host cells indicates the protein is displayable on the lower eukaryote cell surface.

2. The method of Claim 1 , wherein the first binding moiety is a first adapter peptide and the second binding moiety is a second adapter peptide wherein the first and second adapter peptides are capable of a specific pairwise interaction.

3. The method of Claim 2, wherein the first and second adapter peptides are coiled coil peptides that capable of the specific pairwise interaction.

4. The method of Claim 3, wherein the coiled coil peptides are GABAB-Rl and GABAB-R2 subunits that are capable of the specific pairwise interaction.

5. The method of Claim 1, wherein the cell surface anchoring protein is a GPI protein.

6. The method of Claim 5, wherein the cell surface anchoring protein is selected from the group consisting of α-agglutinin, Cwplp, Cwp2p, Gas Ip, Yap3p, Flo Ip, Crh2p, Pirlp, Pir4p, Sedlp, Tiplp, Wpip, Hpwplp, Als3p, and Rbt5p.

7. The method of Claim 1 , wherein the cell surface anchoring protein is

Sedlp.

8. The method of Claim 1 , wherein the lower eukaryote is a yeast.

9. The method of Claim 8, wherein the yeast is Pichia pastoris.

10. The method of Claim 1 , wherein the proteins are antibodies.

11. The method of Claim 1 , wherein expression of the nucleic acids encoding the capture moiety and the mutants of the proteins is constitutive.

12. The method of Claim 1, wherein expression of the nucleic acids encoding the capture moiety and the mutants of the proteins are induced simultaneously.

13. The method of Claim 1 , wherein expression of the nucleic acids encoding the capture moiety and the mutants of the proteins are induced sequentially.

14. A method for selecting a recombinant lower eukaryote host cell that displays a desired protein on the surface of the host cell, comprising:

(a) providing host cells that expresses a capture moiety comprising a cell surface anchoring protein fused to a first binding moiety; (b) transforming the host cells with nucleic acids encoding proteins, each fused to a second binding moiety that is capable of specifically interacting with the first binding moiety fused to the cell surface anchoring protein, to produce a plurality of host cells wherein at least one host cell is suspected of displaying the desired protein on the cell surface;

(c) contacting the transformed host cells with a detection means that specifically binds to the desired proteins that are displayed on the cell surface; and

(d) isolating the host cells with which the detection means is bound to select the host cell that displays the desired protein.

15. The method of Claim 14 wherein the first binding moiety is a first adapter peptide and the second binding moiety is a second adapter peptide wherein the first and second adapter peptides are capable of a specific pairwise interaction.

16. The method of Claim 15 wherein the first and second adapter peptides are coiled coil peptides that capable of the specific pairwise interaction.

17. The method of Claim 16 wherein the coiled coil peptides are GABAB-Rl and GABAB-R2 subunits that are capable of the specific pairwise interaction.

18, The method of Claim 14 wherein the cell surface anchoring protein is a GPI protein.

19. The method of Claim 14 wherein the cell surface anchoring protein is selected from the group consisting of α-agglutinin, Cwplp, Cwp2p, Gas Ip, Yap3p, FIoIp₁ Crh2p, Pirlp, Pir4p, Sedlp, Tiplp, Wpip, Hpwplp, Als3p, and Rbt5p.

20. The method of Claim 14 wherein the cell surface anchoring protein is Sedlp.

21. The method of Claim 14 wherein the lower eukaryote is a yeast.

22. The method of Claim 21 wherein the yeast is Pichiapastoris.

23. The method of Claim 14 wherein the desired protein is an antibody or fragment thereof.

24. The method of Claim 14, wherein expression of the nucleic acids encoding the capture moiety and the mutants of the proteins is constitutive.

25. The method of Claim 14, wherein expression of the nucleic acids encoding the capture moiety and the mutants of the proteins are induced simultaneously.

26. The method of Claim 14, wherein expression of the nucleic acids encoding the capture moiety and the mutants of the proteins are induced sequentially.

27. A method for producing an antibody comprising:

(a) providing a host cell that expresses a capture moiety comprising a cell surface anchoring protein fused to a first binding moiety;

(b) transforming the host cell with a nucleic acid encoding the heavy and light chains of an antibody wherein the heavy chain is fused to a second binding moiety that is capable of specifically interacting with the first binding moiety fused to the cell surface anchoring protein, wherein mutagenesis is used to generate a plurality of host cells encoding a variegated population of mutants of the antibodies; (c) contacting the plurality of host cells with a detection means that specifically binds to antibodies that are displayed on the surface of the host cell and does not bind to antibodies that are not displayed on the surface of the host cell; and

(d) isolating the host cells with which the detection means is bound, wherein the presence of the detection means bound to an antibody on the surface of the host cell indicates the host cell produces the antibody.

28, The method of Claim 27 wherein the first binding moiety is a first adapter peptide and the second binding moiety is a second adapter peptide wherein the first and second adapter peptides are capable of a specific pairwise interaction.

29. The method of Claim 28 wherein the first and second adapter peptides are coiled coil peptides that capable of the specific pairwise interaction.

30. The method of Claim 29 wherein the coiled coil peptides are GABAB-Rl and GABAB-R2 subunits that are capable of the specific pairwise interaction.

31. The method of Claim 27 wherein the cell surface anchoring protein is a GPI protein.

32. The method of Claim 27 wherein the cell surface anchoring protein is selected from the group consisting of α-agglutinin, Cwplp, Cwp2p, Gas Ip, Yap3p, Flo Ip, Crh2p, Pirlp, Pir4p, Sedlp₅ Tiplp. Wpip, Hpwplp, Als3p, and Rbt5p.

33. The method of Claim 27 wherein the cell surface anchoring protein is

Sedlp.

34. The method of Claim 27 wherein the lower eukaryote is a yeast.

35. The method of Claim 37 wherein the yeast is Pichia pastoris.

36. The method of Claim 27, wherein expression of the nucleic acids encoding the capture moiety and the mutants of the proteins is constitutive.

37. The method of Claim 27_? wherein expression of the nucleic acids encoding the capture moiety and the mutants of the proteins are induced simultaneously.

38. The method of Claim 27, wherein expression of the nucleic acids encoding the capture moiety and the mutants of the proteins are induced sequentially.

39. A method for selecting a recombinant lower eukaryote host cell that displays a desired antibody on the surface of the host cell, comprising:

(a) providing host cells that expresses a capture moiety comprising a cell surface anchoring protein fused to a first binding moiety;

(b) transforming the host cell with nucleic acids encoding the heavy and light chains of antibodies wherein the heavy chains are fused to a second binding moiety that is capable of specifically interacting with the first binding moiety fused to the cell surface anchoring protein, to produce a plurality of host cells wherein at least one host cell is suspected of displaying the desired antibody on the cell surface;

(c) contacting the transformed host cells with a detection means that specifically binds to the desired antibody that is displayed on the cell surface; and (d) isolating the host cell with which the detection means is bound to select the host cell that displays the desired antibody.

40. The method of Claim 39 wherein the first binding moiety is a first adapter peptide and the second binding moiety is a second adapter peptide wherein the first and second adapter peptides are capable of a specific pairwise interaction.

41. The method of Claim 40 wherein the first and second adapter peptides are coiled coil peptides that capable of the specific pairwise interaction.

42. The method of Claim 41 wherein the coiled coil peptides are GABAB-Rl and GABAB-R2 subunits that are capable of the specific pairwise interaction.

43. The method of Claim 39 wherein the cell surface anchoring protein is a GPI protein.

44. The method of Claim 39 wherein the cell surface anchoring protein is selected from the group consisting of α-agglutinin, Cwplp, Cwp2p, Gaslp, Yap3p, Flolp, Crh2p, Pirlp, Pir4p, Sedlp, Tiplp, Wpip, Hpwplp, Als3p, and Rbt5p.

45. The method of Claim 39 wherein the cell surface anchoring protein is

Sedlp.

46. The method of Claim 39 wherein the lower eukaryote is a yeast.

47. The method of Claim 46 wherein the yeast is Pichia pastoris .

48. The method of Claim 39, wherein expression of the nucleic acids encoding the capture moiety and the mutants of the proteins is constitutive.

49. The method of Claim39, wherein expression of the nucleic acids encoding the capture moiety and the mutants of the proteins are induced simultaneously.

50. The method of Claim 39, wherein expression of the nucleic acids encoding the capture moiety and the mutants of the proteins are induced sequentially.

51. A method of producing a member of a specific binding pair, wherein the specific binding pair member is an antibody or antibody fragment, comprising an antibody VH domain and an antibody VL domain, and having an antigen binding site with binding specificity for an antigen of interest, the method comprising:

(a) providing a library of lower eukaryote host cells displaying on their surface a specific binding pair member, which specific binding pair member is an antibody or antibody fragment comprising a synthetic human antibody VH domain and a human antibody VL domain, wherein the library is created by:

(i) providing lower eukaryote host cells that express a capture moiety comprising a cell surface anchoring protein fused to a first binding moiety;

(ii) providing a library of nucleic acid sequences encoding a genetically diverse population of the specific binding pair member, wherein the VH domains of the genetically diverse population of the specific binding pair member are biased for one or more

VH gene families and wherein the specific binding pair member includes a second binding moiety that is capable of specifically interacting with the first binding moiety fused to the cell surface anchoring protein; (iii) expressing the library of nucleic acid sequences in the lower eukaryote host cells, whereby each specific binding pair member is displayed at the surface of a lower eukaryote host cell;

(b) selecting one or more specific binding pair members having a binding specificity for the antigen of interest, by binding the one or more specific binding pair members with the antigen of interest, each thus selected specific binding pair member being displayed on the lower eukaryote host cell.

52. The method of Claim 51 , wherein the specific binding pair member comprises a synthetic human antibody VH domain and a synthetic human antibody VL domain and wherein the synthetic human antibody VH domain and the synthetic human antibody VL domain comprise framework regions and hypervariable loops, wherein the framework regions and first two hypervariable loops of both the VH domain and VL domain are essentially human germ line, and wherein the VH domain and VL domain have altered CDR3 loops.

53. The method of Claim 52, wherein in addition to having altered CDR3 loops the human synthetic antibody VH and VL domains contain mutations in other CDR loops.

54. The method of Claim 51 , wherein each human synthetic antibody VH domain CDR loop is of random sequence.

55. The method of Claim 51 , wherein human synthetic antibody VH domain CDR loops are of known canonical structures and incorporate random sequence elements.

56. The method of Claim 51 , wherein the displayed specific binding pair member comprises a single-chain Fv antibody fragment.

57. The method of Claim 51 , wherein the displayed specific binding pair member comprises an antibody.

58. The method of Claim 51 wherein the first binding moiety is a first adapter peptide and the second binding moiety is a second adapter peptide wherein the first and second adapter peptides are capable of a specific pairwise interaction.

59. The method of Claim 58, wherein the first and second adapter peptides are coiled coil peptides that capable of the specific pairwise interaction.

60. The method of Claim 59, wherein the coiled coil peptides are GABAB-Rl and GABAB-R2 subunits that are capable of the specific pairwise interaction.

61. The method of Claim 51 , wherein the cell surface anchoring protein is a GPI protein.

62. The method of Claim 61 , wherein the cell surface anchoring protein is selected from the group consisting of α-aggtutinin, Cwplp, Cwp2p_f Gaslp, Yap3p, Flolp, Crh2p, Pirlp, Pir4p, Sedlp, Tiplp, Wpip, Hpwplp, Als3p, and Rbt5p.

63. The method of Claim 51 , wherein the cell surface anchoring protein is

Sedlp.

64. The method of Claim 51 _> wherein the lower eukaryote is a yeast.

65. The method of Claim 64, wherein the yeast is Pichia pastor is.

66. A method of producing an antibody or antibody fragment, comprising an antibody VH domain and an antibody VL domain, and having an antigen binding site with binding specificity for an antigen of interest, the method comprising: (a) providing a library of lower eukaryote host cells displaying on their surface an antibody or antibody fragment comprising a synthetic human antibody VH domain and a human antibody VL domain, wherein the library is created by:

(i) providing lower eukaryote host cells that express a capture moiety comprising a cell surface anchoring protein fused to a first binding moiety; (ϋ) providing a library of nucleic acid sequences encoding a genetically diverse population of the antibody or antibody fragment, wherein the VH domains of the genetically diverse population of the antibody or antibody fragment are biased for one or more VH gene families and wherein the antibody or antibody fragment includes a second binding moiety that is capable of specifically interacting with the first binding moiety fused to the cell surface anchoring protein;

(iii) expressing the library of nucleic acid sequences in the lower eukaryote host cells, whereby each antibody or antibody fragment is displayed at the surface of a lower eukaryote host cell;

(b) selecting one or more antibodies or antibody fragments having a binding specificity for the antigen of interest, by binding the one or more antibodies or antibody fragments with the antigen of interest, each thus selected antibody or antibody fragment being displayed on the lower eukaryote host cell.

67. The method of Claim 66, wherein the antibody or antibody fragment comprises a synthetic human antibody VH domain and a synthetic human antibody VL domain and wherein the synthetic human antibody VH domain and the synthetic human antibody VL domain comprise framework regions and hypervariable loops, wherein the framework regions and first two hypervariable loops of both the VH domain and VL domain are essentially human germ line, and wherein the VH domain and VL domain have altered CDR3 loops.

68. The method of Claim 67, wherein in addition to having altered CDR3 loops the human synthetic antibody VH and VL domains contain mutations in other CDR loops.

69. The method of Claim 66, wherein each human synthetic antibody VH domain CDR loop is of random sequence.

70. The method of Claim 66, wherein human synthetic antibody VH domain

CDR loops are of known canonical structures and incorporate random sequence elements.

71. The method of Claim 66, wherein the first binding moiety is a first adapter peptide and the second binding moiety is a second adapter peptide and wherein the first and second adapter peptides are capable of a specific pairwise interaction.

72. The method of Claim 71 , wherein the first and second adapter peptides are coiled coil peptides that capable of the specific pairwise interaction.

73. The method of Claim 72, wherein the coiled coil peptides are GABAB-Rl and GABAB-R2 subunϊts that are capable of the specific pairwise interaction.

74. The method of Claim 66, wherein the cell surface anchoring protein is a GPI protein.

75. The method of Claim 74, wherein the cell surface anchoring protein is selected from the group consisting of α-agglutinin, Cwplp, Cwp2p, Gaslp, Yap3p, Flolp, Crh2ρ, Pirlp, PMp, Sedlp, Tiplp, Wpip, Hpwplp, Als3ρ, and Rbt5ρ.

76. The method of Claim 66, wherein the cell surface anchoring protein is

Sedlp.

77. The method of Claim 66, wherein the lower eukaryote is a yeast.

78. The method of Claim 77, wherein the yeast is Pichia pastoris.

79. A vector comprising a nucleic acid encoding a fusion protein comprising a desired protein fused at its C-terminus to the N-terminus of a polypeptide that includes a binding moiety wherein the nucleic acid includes a single stop codon between the nucleotide sequence encoding the desired protein and the nucleotide sequence encoding the polypeptide.

80. The vector of Claim 79, wherein the desired protein is the heavy chain of an antibody.

81. An antibody produced by the method of Claim 27.

82. An antibody produced by the method of Claim 39.

83. An antibody produced by the method of Claim 51.

84. An antibody produced by the method of Claim 66.

85. A lower eukaryote host cell comprising a nucleic acid encoding a capture moiety comprising a cell surface anchoring protein fused to a first binding moiety and nucleic acids encoding the heavy and light chains of antibodies wherein the heavy chains are fused to a second binding moiety that is capable of specifically interacting with the first binding moiety fused to the cell surface anchoring protein.