JP2010213725A - Yeast cell surface display of protein and uses thereof - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a genetic method for tethering polypeptides to a yeast cell wall in a form accessible to protein-protein binding. <P>SOLUTION: There is provided the method for selecting proteins on enhanced displayability exceeding a wild type protein on a yeast cell surface. The method includes the steps of: transforming yeast cells with a vector expressing a protein to be tested fused to a yeast cell wall protein, wherein mutagenesis is used to a generate a variegated population of mutants of the protein to be tested; bringing the yeast cell into contact with a label which binds to proteins that are displayed on the yeast cell surface and does not bind to proteins that are not displayed on the yeast cell surface; and the step of isolating the yeast cell to which the label binds, wherein the enhanced presence of the label bound to the protein to be tested indicates that the protein to be tested can be displayed on the yeast cell surface. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

(Background of the Invention)
(Field of Invention)
The present invention relates generally to the fields of immunology and protein chemistry. More specifically, the present invention relates to the display of peptides and proteins on the surface of yeast cells for the selection of sequences having the desired binding properties from combinatorial libraries.

(Description of related technical fields)
The structure of the site to which the antibody binds can be deduced with reasonable accuracy from peptide sequence data, but the ability to rationally manipulate improvements in binding affinity and binding specificity is more than despite some success. It has proven difficult to understand (eg Roberts et al., 87; Riechmann et al., 92). As a result, library mutagenesis and screening currently represent the most beneficial approach to antibody directed affinity maturation. Combinatorial library screening and selection methods have become a common tool for changing the recognition properties of proteins (Ellman et al., 1997, Physiky & Fields, 1995). In particular, the construction and screening of antibody immune repertoires in vitro promises improved control over the strength and specificity of antibody-antigen interactions.

  The most extensive technique is phage display, where the protein of interest is expressed as a polypeptide fusion to a bacteriophage coat protein and by binding to an immobilized or soluble biotinylated ligand. Subsequent screening (eg, Huse et al., 89; Clackson et al., 91; Marks et al., 92). Fusions are most commonly made against a minor coat protein (referred to as gene III protein (pIII)), which is present in 3-5 copies on the phage chip. Phages constructed in this way are considered compact gene “units” that have both a phenotype (the binding activity of the displayed antibody) and a genotype (the gene encoding that antibody) in one package. obtain. Phage display has been successfully applied to antibodies, DNA binding proteins, protease inhibitors, short peptides, and enzymes.

  Antibodies with the desired binding properties are selected by binding to immobilized antigen in a process called “panning”. The phage-bearing non-specific antibody is removed by washing, then the bound phage is eluted, and E. coli. Amplified by E. coli infection. This approach has been applied to generate antibodies against a number of antigens, including: hepatitis B surface antigen (Zebedee et al., 92); polysaccharide (Deng et al., 94), insulin-like growth factor 1 (Garard and Henner, 93), 2-phenyloxazol-5-one (Riechmann and Weill, 93), and 4-hydroxy-5-iodo-3-nitro-phenacetyl- (NIP) -caproic acid (Hawkins et al. 1992).

Nevertheless, phage display has several weaknesses. Panning an antibody phage display library is a powerful technique, but it has some inherent difficulties that limit its wide and successful application. For example, some eukaryotic secreted proteins and cell surface proteins require post-translational modifications such as glycosylation or extensive disulfide isomerization that are not available in bacterial cells. Furthermore, the nature of phage display makes quantitative and direct identification of ligand binding parameters impossible. For example, very high affinity antibodies (K _D ≦ 1 nM) are difficult to isolate by panning. Because the elution conditions required to disrupt this very strong antibody-antigen interaction are generally harsh enough to sufficiently denature the phage particles to make it sufficiently non-infectious. Some (eg, low pH, high salt). Furthermore, the need for physical immobilization of the antigen on the solid surface creates a number of artificial difficulties. For example, high antigen surface density introduces an avidity effect that hides true affinity. In addition, physical tethering reduces the translational and rotational entropy of the antigen, overestimating a smaller DS for antibody binding and the resulting binding affinity compared to that for soluble antigens. And the great effect from variability during mixing and washing procedures leads to reproducibility difficulties. Furthermore, the presence of only one to a few antibodies per phage particle induces a substantially stochastic variation and makes it impossible to distinguish between antibodies of similar affinity. For example, an affinity difference of more than 6 times is often required for efficient discrimination (Riechmann and Weil, 1993). Finally, the population can be overtaken by more rapidly growing wild type phage. In particular, since pIII is directly involved in the phage life cycle, the presence of some antibodies or bound antigen prevents or slows the growth of the associated phage.

  Several bacterial cell surface display methods have been developed. However, the use of prokaryotic expression systems sometimes induces an unpredictable expression bias and the bacterial capsular polysaccharide layer presents a diffusion barrier that limits such systems to small molecule ligands ( Roberts, 1996). E. E. coli has a lipopolysaccharide layer or capsule that can sterically hinder the polymer binding reaction. Indeed, the putative physiological function of the bacterial capsule is a limitation of the diffusion of macromolecules across the cell membrane to protect the cell from the immune system (DiRienzo et al., 78). E. The periplasm of E. coli has not evolved as a compartment for antibody fragment folding and assembly. Expression of antibodies in E. coli is typically very clone dependent (some clones are well expressed and others are not at all). Such variability is the result of E.I. It raises concerns about the equivalent presentation of all possible sequences in antibody libraries expressed on the surface of E. coli.

  The discovery of new therapeutic agents is facilitated by the development of a yeast selection system. The structural similarity between the B cell display antibody and the yeast cell display antibody provides a closer similarity to affinity maturation in vivo than is available with filamentous phage. Furthermore, the ease of growth culture and the ease of genetic manipulation available in yeast allow large populations to be rapidly mutagenized and screened. By contrasting with conditions in the mammalian body, the physiochemical conditions of binding and selection can be altered for yeast cultures within a wide range of pH, temperature, and ionic strength, providing additional freedom in antibody manipulation experiments. Can be provided. Development of yeast surface display systems for screening combinatorial antibody libraries and screening based on antibody-antigen dissociation rates is described herein.

A potential application to manipulate monoclonal antibodies for the diagnosis and treatment of human diseases is a long way off. In particular, applications for cancer therapy and tumor imaging are actively pursued. Antibody therapy for gram-negative sepsis is still promising despite the discouraging preliminary results. In vitro applications for immunohistochemistry, immunoassays, and immunoaffinity chromatography have already been well developed. For each of these applications, antibodies with high affinity (ie, K _D ≦ 10 nM) and high specificity are desirable. Case evidence suggests that the phage display system and bacterial display system do not appear to consistently produce antibodies with an affinity less than nM. To date, yeast display fills this gap, and such should be a key technology of enormous commercial and medical importance.

  The importance of T cell receptors for cell-mediated immunity has been known since the 1980s, but no method has been developed for manipulating higher affinity T cell receptors. Several groups have produced single chain T cell receptor constructs, but these expression systems have allowed biochemical analysis of T cell receptor binding, but change their binding properties in a directional manner. Did not allow a library method for. It was particularly difficult to produce this T cell receptor in a soluble form. In its endogenous form, it is the heterodimeric (αβ) membrane protein that non-covalently associates with the subunits of the CD3 complex on the surface of T lymphocytes. The extracellular α and β domains are composed of both constant regions (Cα and Cβ) and variable regions that function directly upon peptide / MHC antigen binding (Vα and Vβ). Several difficult methods for production have been developed: secretion of VαVβ from E. coli, production of chimeric molecules in myeloma cells consisting of VαCα VβCβ fused to the constant region of immunoglobulin κ light chain, VαCα VβCβ by cleavage from fusion with cell surface lipids in thymoma (Slanetz et al., 1991), cleavage of the rat basophilic leukemia cell line production VαCα VβCβ-z complex prior to the transmembrane region upon thrombin digestion, baculovirus-infected insects with or without leucine zipper association Production of VαCα VβCβ in cells, and secretion of immunoglobulin CH-gd TCR chimeras from COS cells (Eilat et al., 1992). A smaller single chain TCR fragment (scTCR), similar to a single chain antibody fragment containing the minimal binding subunit of a complete TCR, is described in E.C. constructed and produced in E. coli and either refolded from inclusion bodies at levels of 0.5-1.0 mg / L or folded in the periplasm.

  The prior art is deficient in that it lacks an effective means of displaying cell surface peptides and proteins for the selection of sequences with the desired binding properties. The prior art is also deficient in that it lacks an effective means of manipulating the T cell receptor for improved binding properties. More specifically, techniques for manipulating soluble T cell receptors to generate therapeutic intervention for cell-mediated immunity are not available. The present invention fulfills a longstanding need and desire in the art.

(Summary of the Invention)
The present invention provides the following.

(Item 1) A method for selecting a protein having enhanced phenotypic characteristics compared to the phenotypic characteristics of a wild-type protein, comprising:
  Transforming yeast cells with a vector expressing the protein to be tested, fused to the yeast cell wall protein, wherein mutagenesis is used to enrich for the variation of the protein to be tested Generating a population;
  Labeling said yeast cell with a first label, wherein said first label binds to yeast expressing said protein to be tested and does not bind to yeast not expressing said protein to be tested The process;
  Isolating the yeast cells to which the first label binds; and
  Analyzing the properties of the mutant protein expressed by yeast and comparing it with the properties of the wild type protein, wherein the yeast cells present the mutant protein having enhanced properties relative to the wild type protein Is selected, process,
Including a method.

2. The method of claim 1, wherein the phenotypic characteristic is selected from the group consisting of surface expression level, stability, association constant, and dissociation constant.

3. The method of claim 1, wherein the protein being tested is an antibody, Fab, Fv, or scFv antibody fragment.

4. The method of claim 1, wherein the protein being tested is a ligand binding domain of a cell surface receptor.

(Item 5) The method according to item 4, wherein the cell surface receptor is a T cell receptor.

6. The method of claim 1, wherein the protein to be tested is fused to the C-terminus of the yeast cell wall protein by its N-terminus.

(Item 7) The method according to item 1, wherein the yeast cell wall protein is agglutinin.

(Item 8) The method according to item 1, wherein the yeast strain is a genus selected from the group consisting of Saccharomyces, Pichia, Hansenula, Schizosaccharomyces, Kluyveromyces, Yarrowia, and Candida.

9. The method of claim 1, wherein the variant of the protein being tested is selected from the group consisting of a single variant and a plurality of variants.

10. The method of claim 1, wherein the first label is selected from the group consisting of magnetic particles and fluorescent labels attached to a ligand for the protein being tested.

11. The method of claim 1, wherein the selection of a mutated protein of interest having enhanced phenotypic characteristics comprises repeated cycles of the enrichment and labeling steps.

(Item 12) The method according to item 1, further comprising:
  Labeling said yeast cell with a second label, wherein said vector used to transform said yeast cell is fused to said protein to be tested to produce a fusion polypeptide. Means for expressing a polypeptide sequence, and wherein the second label binds to a yeast cell that expresses the fusion polypeptide and does not bind to a yeast cell that does not express the fusion polypeptide;
  Enriching the transformed yeast population by quantifying the second label, wherein the appearance of the second label is related to the amount of the fusion polypeptide expressed on the cell surface. Directly proportional to the process; and
  Comparing the quantification of the second label and the quantification of the first label to determine the surface expression level of the protein to be tested;
Including a method.

(Item 13) The method according to item 12, wherein an increase in the surface expression level of the mutant protein to be tested compared to the surface expression level of the wild type protein to be tested here is a desired value of the mutant protein. A method that can be used to select for phenotypic characteristics of

14. The method of claim 13, wherein the phenotypic characteristic is selected from the group consisting of intracellular expression level, stability, association constant, dissociation constant, secretion level, and solubility.

(Item 15) The method according to item 12, wherein the polypeptide portion of the fusion polypeptide recognized by the second label is an epitope tag.

(Item 16) The method according to item 12, wherein the first label and the second label are fluorescent labels.

(Item 17) The method according to item 1, further comprising the following steps:
  Cloning the gene encoding the selected mutant protein into a vector adapted for expression in a eukaryote; and
  Expressing the mutant protein in the eukaryote, wherein the enhanced property of the mutant protein compares the property of the wild-type protein with the property of the enhanced property of the mutant protein. Confirmed by the process,
Including a method.

(Item 18) The method according to item 17, wherein the eukaryote is selected from the group consisting of mammals, insects and yeasts.

(Item 19) The method according to item 1, further comprising:
  Cloning the gene encoding the selected mutant protein into a vector adapted for expression in prokaryotes; and
  Expressing the mutant protein in the prokaryote, wherein the enhanced property of the mutant protein compares the property of the wild-type protein with the property of the enhanced property of the mutant protein. Confirmed by the process,
Including a method.

(Item 20) A method for selecting a protein for displayability on the surface of a yeast cell, comprising:
  A process for transforming yeast cells with a vector expressing a protein to be tested fused to a yeast cell wall protein, wherein mutagenesis is used to enrich the variation of the protein to be tested. Generating a population, process;
  Labeling said yeast cells with a first label, wherein said first label binds to yeast expressing said protein to be tested and does not express said protein to be tested Not combined with the process;
  Isolating the yeast cells to which the first label binds by quantifying the first label, wherein the high appearance of the first label is that the protein being tested is desired Indicating that the first label has a display property and the low appearance of the first label indicates that the protein being tested does not have the desired display property;
Including a method.

21. The method of claim 20, wherein the protein being tested is an antibody, Fab, Fv, or scFv antibody fragment.

22. The method of claim 20, wherein the protein being tested is a ligand binding domain of a cell surface receptor.

(Item 23) The method according to item 22, wherein the cell surface receptor is a T cell receptor.

  In one embodiment of the invention, a genetic method is provided for tethering a polypeptide to the yeast cell wall in a form accessible to protein-protein binding. Combining this method with fluorescence activated cell sorting provides a means of selecting proteins with increased or decreased affinity for another molecule, altered specificity, or conditional binding.

In another embodiment of the invention, a method is provided for genetically fusing a polypeptide of interest to the C-terminus of a yeast Aga2p cell wall protein. Under conjugation conditions, the outer wall of each yeast cell contains about 10 ⁴ protein molecules called agglutinin. This agglutinin serves as a specific contact to mediate adhesion of the opposite mating type yeast cells during mating. In fact, yeast has evolved a platform for protein-protein binding without steric hindrance from cell wall components. By attaching the antibody to agglutinin, it can effectively mimic the cell surface display of the antibody by B cells in the immune system.

  In another embodiment, a method of fusing a 9 residue epitope (HA) tag to the C-terminus of an AGA2 protein is provided. This short peptide is accessible on the cell surface to antibodies in solution without cell fixation or digestion and can be detected by flow cytometry or fluorescence microscopy. Thus, yeast can be used to display peptides.

  In yet another embodiment, a method is provided for fusing a scFv fragment of a 4-4-20 monoclonal antibody to the C-terminus of AGA2 protein, the fragment being accessible on the cell surface and any immobilization of cells. Alternatively, fluorescent antigen can be bound without digestion and detected by flow cytometry or fluorescence microscopy. Thus, yeast can be used to display antibody fragments.

  In one aspect of the invention, a method is provided for selecting a protein having a desired binding characteristic comprising the following steps: a vector expressing the protein to be tested fused N-terminally to a yeast cell wall binding protein. Transforming the yeast cells with a step of labeling the yeast cells with a first label, wherein the first label binds to the yeast expressing the protein to be tested and is tested as described above. Not binding to yeast that does not express the protein; selecting for yeast cells that bind to the first label; and quantifying the first label, where the high appearance of the first label is tested A process indicating that the protein has the desired binding properties, wherein a low appearance of the first label indicates that the protein being tested does not have the desired binding properties.

  A preferred embodiment of the present invention further comprises the following steps: labeling the yeast cell with a second label, wherein the second label is fused to the protein encoded by this tested and vector. Binding to yeast expressing the tag and not binding to yeast not expressing the epitope tag encoded by the vector; quantifying the second label, wherein the appearance of the second label is the yeast cell surface A step indicating the expression copy number of the epitope-tagged protein tested above, as well as the quantification of the second label to determine the appearance of the first label normalized to the appearance of the second label; Comparing the quantification of the first label, wherein the appearance of the first label being high compared to the appearance of the second label indicates that the protein being tested has the desired binding properties; Degree.

  Another preferred embodiment of the present invention comprises the following steps: labeling yeast cells with a third label that competes with the first label for binding to the protein being tested; with the first label Labeling yeast cells; quantifying a first label; labeling yeast cells with a second label; quantifying a second label; and normalized to the appearance of the second label Comparing the quantification of the second label and the quantification of the first label to determine the appearance of the first label, wherein the appearance of the first label is low compared to the appearance of the second label Indicates that the protein being tested has the desired binding properties.

  In one embodiment of the invention, the first label is a fluorescent label attached to the ligand and the second label is a fluorescent label attached to the antibody. If the label is fluorescent, the quantification step is performed by flow cytometry or confocal fluorescence microscopy.

  Another aspect of the invention provides a vector for carrying out the method of the invention, the vector comprising a cell wall binding protein fused to the N-terminus of the protein of interest. Preferred embodiments of this aspect of the invention include means for expressing a polypeptide epitope tag fused to a protein of interest in yeast cells. A more preferred embodiment provides that the cell wall binding protein is a binding subunit of a yeast agglutinin protein, and even more preferably, the yeast agglutinin binding subunit is Aga2p.

  Another preferred embodiment of this aspect of the invention is that the epitope tag amino acid sequence is YPYDVPDYA (HA) (SEQ ID NO: 1), EQKLISEEDL (c-myc) (SEQ ID NO: 2), DTYRYI (SEQ ID NO: 3), TDFYLK ( Selected from the group of SEQ ID NO: 4), EEEEYMPME (SEQ ID NO: 5), KPPTPPPPEPET (SEQ ID NO: 6), HHHHHH (SEQ ID NO: 7), RYIRS (SEQ ID NO: 8), or DYKDDDDK (SEQ ID NO: 9), and purpose It is provided that the N-terminus of this protein is fused to the C-terminus of the cell wall binding protein.

  In yet another preferred embodiment of the present invention, a method is provided for selecting a protein having enhanced phenotypic characteristics compared to the phenotypic characteristics of the wild-type protein, the method comprising the following steps: A process for transforming yeast cells with a vector that expresses a protein to be tested fused to a yeast cell wall protein, wherein mutagenesis is used to enrich for the variation of the protein variant being tested. Labeling yeast cells with a first label, wherein the first label binds to a yeast that expresses the protein to be tested and does not express the protein to be tested Isolating yeast cells that bind to the first label; and a mutant tamper expressed by yeast cells having the properties of a wild-type protein A process of analyzing and comparing the characteristics of quality, yeast cells displaying mutant proteins with enhanced properties relative to the wild type protein, where is selected, step. As noted above, the second and / or third label may be used in this embodiment, and the selection of the mutant protein of interest having enhanced phenotypic characteristics includes an iterative cycle of the enrichment step and the labeling step. obtain.

  In a preferred embodiment, the yeast cell wall protein is agglutinin; the protein to be tested is fused by its N-terminus to the C-terminus of the binding subunit agglutinin; the yeast strain is And the protein to be tested is an antibody, Fab, Fv, or scFv antibody fragment, more preferably the protein to be tested is a ligand binding domain of a cell surface receptor Even more preferably, the cell surface receptor is a T cell receptor.

  An object of the present invention is to provide a method for selecting mutant proteins that exhibit enhanced phenotypic characteristics selected from the group consisting of surface expression, stability, association constant, dissociation constant, secretion level, and solubility. And the protein variants tested here may comprise a single mutation or multiple mutations.

  In yet another embodiment, a second label can be used to quantitatively determine cell surface expression levels, which can include other desired phenotypic characteristics (eg, intracellular expression, stability, binding constant, It can be used as an assay to select for (dissociation constant, level of secretion, and solubility).

  In another embodiment, the mutant protein selected by the method of the invention is obtained by cloning the gene encoding the selected mutant protein into a vector adapted for expression in eukaryotes; and It can be further characterized by expressing this mutant protein in its eukaryote. Here, the enhanced properties of the mutant protein are confirmed by comparing the enhanced phenotypic properties of the mutant protein with those of the wild-type protein. Preferably, the eukaryote is selected from the group consisting of mammals, insects or yeasts. Since the variant phenotype chosen is the genetic property of the “new” (ie variant) protein, this approach can also be applied to the expression of this variant in other non-eukaryotic expression systems. Is possible.

Yeast surface display and selection by flow cytometry have been used to isolate scFv variants specific for the Vb8 region of the T cell receptor. Selection was based on equilibrium binding by two fluorescently labeled probes (soluble Vb8 domain and an antibody against the c-myc epitope tag present at the carboxy terminus of the scFv). Mutants selected in this screen included scFv with a 3-fold increased affinity for Vb8, and scFv clones bound with reduced affinity by anti-c-myc antibody. The latter finding indicates that the yeast display system can be used for mapping conformational epitopes (which cannot be revealed by standard peptide screening). Equilibrium antigen binding constants can be evaluated in a surface display format, which allows for screening of isolated mutants without the need for subcloning and soluble expression. Only a relatively small library (3 × 10 ⁵ ) of yeast cells displaying randomly mutated scFv was screened to identify these variants. This indicated that this system provides a powerful means to manipulate the binding properties of eukaryotic secreted proteins and cell surface proteins.

Another preferred embodiment of this aspect of the invention provides a method of presenting a protein that is not presented as its normal ("wild type") sequence. In the example shown, the T cell receptor for the antigen was not expressed as its “wild type” sequence. However, this mutant receptor was expressed on the surface of yeast cells after random mutation, selection by flow cytometry using the appropriate configuration specific antibody. This strategy allows the discovery of new T cell receptors and provides a method for the presentation of virtually any polypeptide. Thus, the present invention also provides a method for selecting proteins for displayability on the yeast cell surface, which method comprises the following steps: fusing the yeast cell, the protein to be tested with the yeast cell wall protein Transforming with a vector to be expressed, wherein the mutation is used to generate a mottled population of mutants of the protein being tested; labeling the yeast cells with a first label The first label binds to the yeast that expresses the protein to be tested and does not bind to the yeast that does not express the protein to be tested; the yeast cell to which the first label binds is Isolating by quantifying the label, wherein a high presence of the first label indicates that the protein being tested has the desired display properties, and wherein There have encompasses protein to be tested indicates that no desired presentation characteristics, the process. Preferably, the protein to be tested is an antibody, Fab antibody fragment, Fv antibody fragment, or scFv antibody fragment, or a ligand binding domain of a cell surface receptor. A typical example of a cell surface receptor is the T cell receptor.
Other and further aspects, features and advantages of the present invention will be apparent from the following description of the presently preferred embodiments of the invention provided for the purpose of disclosure.

FIG. 1 is a schematic diagram showing in vitro affinity maturation by yeast display. FIG. 2 shows a schematic diagram of surface presentation on yeast. An epitope of a 9 amino acid peptide from the hemagglutinin antigen (HA) was fused to the C-terminus of the Aga2p subunit of a-agglutinin, followed by the 4-4-20 anti-fluorescein scFv sequence. An additional 10-residue epitope tag (c-myc) is fused at the C-terminus of the scFv, allowing quantification of fusion presentation independent of antigen binding by either the HA tag or the c-myc tag. became. The HA tag or c-myc tag can be used to normalize variations in the number of presented fusion proteins in dual-labeled flow cytometry. FIG. 3 shows a vector for yeast surface display. FIG. 3A shows the structure of vector pCT202. FIG. 3B shows specific restriction sites and transcriptional regulation by galactose, N-terminal HA tag and C-terminal c-myc epitope tag, and Factor XA protease cleavage site. FIG. 4 demonstrates that the presented fusion can be detected by fluorescence techniques and shows a flow cytometry histogram of yeast labeled with a-c-myc / a-mouse-PE. FIG. 5 shows that antigen binding by 4-4-20 scFv can be detected by fluorescence and shows a flow cytometry histogram of yeast labeled with FITC-dextran (2 × 10 ⁶ Da). FIG. 6 shows that 4-4-20 activity and c-myc can be detected simultaneously and demonstrate a 1: 1 correlation of fluorescence signal; thus, variation in intensity signal 1 (FITC) An intensity of 2 (PE) can normalize cell-to-cell variation in the expression of the protein of interest. FIG. 7 shows the sequence of the AGA2-HA-4-4-20-c-myc gene cassette. FIG. 8 shows a confocal microscopic image of yeast presenting scFv. Yeast containing plasmids directed against surface expression of HA peptides (FIG. 8A) or scFv fusions (FIG. 8B) were injected with mAb 9E10, followed by secondary anti-mouse IgG-R-phycoerythrin (PE) conjugate and FITC-dextran. Labeled. DIC (upper panel), red PE fluorescence (middle panel), and green FITC fluorescence (lower panel) images were collected. FIG. 9 shows a flow cytometric analysis of yeast presenting scFv. FIG. 9A shows that yeast strains displaying either an irrelevant peptide (FIG. 9A) or 4-4-20 scFv (FIG. 9B) were labeled with mAb 9E10 and FITC-dextran. Also, cells presenting scFv were treated with 5 mM DTT prior to labeling (FIG. 9C). (I) Univariate histogram of PE fluorescence combined with labeling with 9E10; (ii) Univariate histogram of FITC fluorescence; (iii) Bivariate histogram showing the correlation between PE and FITC fluorescence. FIG. 9 shows a flow cytometric analysis of yeast presenting scFv. FIG. 9A shows that yeast strains displaying either an irrelevant peptide (FIG. 9A) or 4-4-20 scFv (FIG. 9B) were labeled with mAb 9E10 and FITC-dextran. Also, cells presenting scFv were treated with 5 mM DTT prior to labeling (FIG. 9C). (I) Univariate histogram of PE fluorescence combined with labeling with 9E10; (ii) Univariate histogram of FITC fluorescence; (iii) Bivariate histogram showing the association between PE and FITC fluorescence. FIG. 9 shows a flow cytometric analysis of yeast presenting scFv. FIG. 9A shows that yeast strains displaying either an irrelevant peptide (FIG. 9A) or 4-4-20 scFv (FIG. 9B) were labeled with mAb 9E10 and FITC-dextran. Also, cells presenting scFv were treated with 5 mM DTT prior to labeling (FIG. 9C). (I) Univariate histogram of PE fluorescence combined with labeling with 9E10; (ii) Univariate histogram of FITC fluorescence; (iii) Bivariate histogram showing the association between PE and FITC fluorescence. FIG. 10 demonstrates the enrichment of yeast displaying improved scFv variants by kinetic mechanical selection and flow cytometry cell sorting. Fluorescent antigen with a 5-aminofluorescein yeast library expressing mutated 4-4-20 scFv (FIG. 10A), and a pool of yeast resulting from three kinetic selections and growth (FIG. 10B) The cells that were subjected to the competitive dissociation of the cells and presenting the mutants that bound most strongly at the highest ratio of FITC intensity / PE intensity remained. FIG. 10 demonstrates the enrichment of yeast displaying improved scFv variants by kinetic mechanical selection and flow cytometry cell sorting. Fluorescent antigen with 5-aminofluorescein, a mutated yeast library expressing 4-4-20 scFv (FIG. 10A), and a pool of yeast resulting from three kinetic selections and growth (FIG. 10B) The cells that were subjected to the competitive dissociation of the cells and presenting the mutants that bound most strongly at the highest ratio of FITC intensity / PE intensity remained. FIG. 11 shows the dissociation kinetics of the interaction between fluorescein and surface presented scFv. Yeast presenting 4-4-20 scFv (circle), mutant 4M1.1 (square) isolated from the library, and mutant 4M1.2 (triangle) are labeled with mAb 9E10 and FITC-dextran It was. 5-aminofluorescein was added as a competitor. The mean intensity of FITC fluorescence of the 9E10 positive population of cells followed a function of time. The slope of this straight line is equal to the kinetic mechanical dissociation rate, and the extrapolated value at time t = 0 is equal to the valence of the interaction. MFI _i = relative average fluorescence intensity of yeast at time t = i. FIG. 12 shows the expression levels and antigen binding properties of scFv-KJ16 (shaded) and control Aga2p / HA (unshaded) displayed on the yeast surface. Yeast EBY100 strain was transformed with scFv-KJ16 or pCT202 vector alone cloned into yeast display vector pCT202. Following induction in galactose medium overnight at 20 ° C., cells were stained with fluorescent antibody and analyzed by flow cytometry. (FIG. 12A) scFv-KJ16 / yeast or Aga2p / HA / yeast, mouse anti-c-myc Mab (9E10) stained with FITC-labeled goat anti-mouse IgG following mouse anti-HA Mab (12CA5) ScFv-KJ16 / yeast or Aga2p / HA / yeast, stained with FITC-labeled goat anti-mouse IgG, (FIG. 12C) stained with about 10 nM biotinylated scTCR followed by streptavidin-phycoerythrin conjugate. scFv-KJ16 / yeast or Aga2p / HA / yeast, and (Figure 12D) biotinylated scTCR followed by the presence (shaded) or absence (shaded) of 100 mg / ml intact IgG KJ16. Not dyed with streptavidin-phycoerythrin conjugate It has been scFv-KJ16 / yeast. FIG. 13 shows the equilibrium antigen binding isotherm of the cell wall presenting scFv-KJ16 as determined by flow cytometry. Yeast EBY100 strain displaying surface scFv-KJ16 was incubated with varying concentrations of biotinylated scTCR, labeled with streptavidin-phycoerythrin conjugate, and detected by flow cytometry. The data, as Scatchard diagram, or as titer (inset) was plotted, and determine the effective K _D of approximately 500 nM. MFU refers to the average fluorescence unit. FIG. 14 shows the selection window used to select the two-dimensional fluorescence histogram and scFv-KJ16 mutant. ScFv-KJ16 cloned into the display vector pCT202 is E. coli. The E. coli mutagenesis gene strain XL1-Red (Stratagene) was transformed and grown for a 6 night growth cycle. Mutant library plasmids were purified and used for LiAc transformation of EBY100 yeast (Gietz et al., 1995). After introduction at 30 ° C., the yeast was screened using a fluorescence activated cell sorter. (FIG. 14A) Representative histogram from the first round of cell sorting (screening window is shown) and (FIG. 14B) Representative histogram from the fourth (final) round of sorting (showing population enrichment in the sorting window) ). FIG. 14 shows the selection window used to select the two-dimensional fluorescence histogram and scFv-KJ16 mutant. ScFv-KJ16 cloned into the display vector pCT202 is E. coli. The E. coli mutagenesis gene strain XL1-Red (Stratagene) was transformed and grown for a 6 night growth cycle. Mutant library plasmids were purified and used for LiAc transformation of EBY100 yeast (Gietz et al., 1995). After introduction at 30 ° C., the yeast was screened using a fluorescence activated cell sorter. (FIG. 14A) Representative histogram from the first round of cell sorting (screening window is shown) and (FIG. 14B) Representative histogram from the fourth (final) round of sorting (showing population enrichment in the sorting window) ). FIG. 15 shows the average level of binding to anti-HA Mab, anti-c-myc Mab, or biotinylated scTCR for 10 randomly selected clones from the final selection shown in FIG. 14B. Ten mutants and wild type scFv-KJ16 / yeast were induced overnight at 30 ° C. in galactose medium. Cells were stained with mouse anti-HA Mab followed by FITC-labeled goat anti-mouse IgG (white bar), mouse anti-c-myc followed by FITC-labeled goat anti-mouse IgG (grey bar), or biotinylated scTCR After staining with streptavidin-phycoerythrin conjugate (black bar) following (about 40 nM), it was analyzed by flow cytometry. FIG. 16 shows the fluorescence label distribution for anti-c-myc binding or scTCR binding for the three selected mutants shown in FIG. Three classes of scFv-KJ16 / yeast mutants were double-stained with anti-c-myc and biotinylated scTCR followed by FITC-labeled goat anti-mouse IgG and streptavidin-phycoerythrin conjugate, then described in FIG. Were analyzed by flow cytometry. The fluorescence distribution for each scFv-KJ16 / yeast variant (shaded) and wild type scFv-KJ16 / yeast (unshaded) is shown. Figures 16A and 16B (mut4); Figures 16C and 16D (mut7); Figures 16E and 16F (mut10). FIG. 17 shows the equilibrium antigen binding isotherm for the three variants shown in FIG. Aga2p / HA / yeast, wild type scFv-KJ16 / yeast, and the three mutant scFv-KJ16 / yeast characterized in FIG. 16 stained with various dilutions of biotinylated scTCR followed by streptavidin-phycoerythrin conjugate It was done. After analysis by flow cytometry, the binding isotherm was graphed with MFU as a function of scTCR dilution. FIG. 18 shows the sequence analysis of wild type scFv-KJ16 / yeast, mut4 and mut7. Plasmids from the wild type scFv-KJ16 and the two mutants (mut4 and mut7) were recovered by plasmid rescue as described below, and E. coli to generate plasmids for sequencing. E. coli DH5α competent cells were transformed. Sequencing analysis was performed using primers adjacent to the scFv of the display vector. Mutations are shown in bold. FIG. 19 shows the flow cytometry profile of an antibody that binds to yeast transformed with a plasmid containing the T cell receptor single chain (VαVβ) gene. The normal or wild type (wt) sequence was compared to several variants (mTCR7, mTCR15, mTCR16) selected after random mutation of the scTCR plasmid. Selection included binding of antibody 1B2 (recognizing conformational epitopes on the T cell receptor) followed by several fluorescence activated cell selections. In the first panel, yeast cells were stained with an antibody against the HA tag (12CA5). In the second panel, yeast cells were stained with an antibody against the T cell receptor (1B2). The HA epitope is expressed on the surface in each case, but only cells expressing the mutant plasmid can express the native T cell receptor (1B2 positive). FIG. 20 shows the flow cytometry profile of antibodies that bind to yeast transformed with the double mutant from the selection shown in FIG. The cells were stained for flow cytometry as shown in FIG. The double mutant showed an increase in the level of T cell receptor (ie, 1B2 reactant). This result indicates that the level of cell surface expression of the T cell receptor can be enhanced by combining with a single mutation. FIG. 21 shows the sequence of mutations that lead to enhanced expression of the cell surface T cell receptor. These included residue 17 of Vβ, residue 43 of Vα, and residue 104 of Vα. FIG. 22 shows the relative secretion levels of scTCR. Soluble expression level of scTCR produced using a low copy yeast expression system (discretionary units). Triplicate cultures from independent clones were analyzed for 1B2 ELISA activity. FIG. 23A shows a flow histogram of yeast presenting scTCR. Represents cell-sorted populations for wild-type and representative single, double, and triple mutants. Average fluorescence units (FITC fluorescence: anti-HA, PE fluorescence: 1B2) are shown on each histogram. Anti-HA indicates the number of surface fusions, and 1B2 indicates the number of cells presenting properly folded scTCR. FIG. 23B shows the correlation between surface expression and soluble secretion. The 1B2 active surface scTCR determined by flow is compared to the 1B2 activity of the soluble secreted material determined by ELISA assay. A minimal, duplicate flow experiment was performed to determine the average fluorescence unit for a particular clone. FIG. 24A shows the temperature stability of the scTCR. Yeast supernatant samples containing scTCR were subjected to the indicated temperatures for 1 hour. Triplicate samples were analyzed for the 1B2 active fraction by ELISA. Fractions were individually normalized and unified by the highest intensity ELISA signal with no loss of activity. In both FIG. 24A and FIG. 24B, the comparison was made to TRX-TCR rather than wild-type scTCR. This is because wild type scTCR was not detected in the yeast supernatant. FIG. 24B shows the kinetics of scTCR temperature denaturation. The 1B2 ELISA activity of the supernatant samples was monitored as a function of the time each sample was incubated at 46 ° C. The observed kinetic mechanical thermal denaturation rate (k _obs ) at 46 ° C. is shown. In both FIG. 24A and FIG. 24B, comparisons were made to TRX-TCR over wild-type scTCR. This is because wild type scTCR was not detected in the yeast supernatant. FIG. 25 shows the correlation between thermostable expression and soluble expression. The scTCR 1B2 ELISA activity remaining after 1 hour incubation at 48 ° C. is compared to the relative secretion level (1B2 ELISA). Triplicate samples were analyzed for secretion levels and thermostability analysis.

(Detailed description of the invention)
As used herein, the term “affinity maturation” refers to a sequential mutation and selection process in which higher affinity antibodies are selected. As used herein, the term “agglutinin” refers to a yeast surface adhesion protein that binds two yeast cells together during conjugation. As used herein, the term “antibody” refers to a protein produced by the mammalian immune system that binds firmly and specifically to a specific molecule. As used herein, the term “ligand” refers to a molecule that is specifically bound by a particular protein. As used herein, the term “antigen” refers to a ligand that is specifically bound by an antibody. As used herein, the term “complementarity determining region” or “CDR” refers to the portion of an antibody that makes direct contact with the bound antigen. As used herein, the term “fluorescence activated cell sorting” or “flow cytometry” refers to a method of sorting a cell population based on a differential fluorescent label. As used herein, the term “hapten” refers to a small antigen that cannot stimulate an immune response without being bound to a carrier. As used herein, the term “single chain antibody” or “SCA” refers to a fusion of a portion of a heavy and light chain of an antibody that retains a single active binding site. The term scFv is used interchangeably and refers to a single chain antibody. As used herein, the term “epitope tag” refers to a sequence of amino acids that are specifically bound to an antibody when fused to another protein. As used herein, the term “HA” refers to the epitope tag sequence YPYDVPDYA (SEQ ID NO: 1). As used herein, the term “c-myc” refers to the epitope tag sequence EQKLISEEDL (SEQ ID NO: 2). As used herein, the term “scFv4-4-20” refers to an scFv that specifically binds to fluorescein conjugated to fluorescein and other molecules (eg, biotin or dextran). As used herein, the term “AGA2p” refers to the protein product of the yeast AGA2 mating type a-agglutinin gene. The term “displayability” is used to describe a combination of biophysical features that allow a protein to escape a secreted “quality maintenance” device that retains and degrades misfolded proteins. (Hammond and Helenius, 1995). Proteins displayed on the yeast cell surface must first successfully pass this quality maintenance step. Both protein kinetics and thermodynamic stability of protein folding are thought to determine the efficiency of avoiding quality maintenance equipment.

  In accordance with the present invention, conventional molecular biology, microbiology, and recombinant DNA techniques within the skill of the art can be used. Such techniques are explained fully in the literature. For example, Maniatis, Fritsh and Sambrook, “Molecular Cloning: A Laboratory Manual (1982);“ DNA Clonig: A Practical Approach ”, Volumes I and II (D. N. Glover, ed. (MJ Gait, 1984); “Nucleic Acid Hybridization” (BD Hames and SJ Higgins, (1985)); “Transscript and Translation” (B. D. Hames and S. J.). Higgins, (1984)); “Animal Cell Culture” (R.I.F. eshney ed., (1986)); "Immobilized Cells And Enzymes" (IRL Press, (1986)); B. See Perbal, “A Practical Guide To Molecular Cloning” (1984).

  A “vector” is a replicon (eg, a plasmid, phage, or cosmid) to which another DNA segment can be added such that replication of the added segment occurs.

  A DNA “coding sequence” is a double-stranded DNA sequence that is transcribed and translated into a polypeptide in vivo when placed under the control of appropriate regulatory sequences. The boundaries of the coding sequence are determined by a start codon at the 5 '(amino) terminus and a translation stop codon at the 3' (carboxyl) terminus. A coding sequence can include prokaryotic sequences, cDNA from eukaryotic mRNA, genomic DNA sequences from eukaryotic (eg, mammalian) DNA, and even synthetic DNA sequences. A polyadenylation signal and transcription termination sequence will usually be located 3 'to the coding sequence.

  Transcriptional control sequences and translational regulatory sequences are DNA regulatory sequences that provide for expression of a coding sequence in a host cell (eg, promoters, enhancers, polyadenylation signals, terminators, etc.).

  A “promoter sequence” is a DNA regulatory region capable of binding RNA polymerase in a cell and initiating transcription of a downstream (3 ′ direction) coding sequence. For purposes of defining the present invention, the promoter sequence has a transcription initiation site attached to its 3 ′ end and extends upstream (5 ′ direction) to initiate transcription at a level detectable in the above background. It contains the minimum number of bases or elements necessary to do. Within the promoter sequence are found the transcription initiation site (conveniently defined by mapping with nuclease SI) and the protein binding domain (consensus sequence) responsible for the binding of RNA polymerase. Eukaryotic promoters often (but not always) include a “TATA” box and a “CAT” box. Prokaryotic promoters contain Shine-Dalgarno sequences in addition to the -10 and -35 consensus sequences.

  An “expression control sequence” is a DNA sequence that controls and regulates the transcription and translation of another DNA sequence. A coding sequence is “under control” of transcriptional and translational control sequences in a cell when RNA polymerase transcribes the coding sequence into mRNA (which is then translated into the protein encoded by the coding sequence). It is.

  “Selection gene” refers to a gene that enables differentiation of cells presenting a required phenotype when performing certain conditions. For example, bacterial growth in a medium containing antibiotics to select bacterial cells containing antibiotic resistance genes.

  The term “primer” as used herein refers to an oligonucleotide (whether naturally occurring, such as with a purified restriction digest, or produced synthetically) that is primer extension. When placed under conditions that induce the synthesis of the product (which is complementary to the strand of the nucleic acid) (ie, in the presence of nucleotides and inducers (eg, DNA polymerase) and at the appropriate temperature and pH) The primer can be either single-stranded or double-stranded and is long enough to prime the synthesis of the desired extension product in the presence of the inducer. The exact length of a primer depends on many factors, including temperature, primer source, and method used, eg, diagnosis For application, depending on the complexity of the target sequence, the oligonucleotide primer typically including 15-25 or more nucleotides, which can comprise fewer nucleotides.

  The primers herein are selected to be “substantially” complementary to separate strands of a particular target DNA sequence. This means that these primers must be substantially complementary to hybridize with their respective strands. Thus, this primer sequence need not reflect the exact sequence of the template. For example, a non-complementary nucleotide fragment can be added to the 5 'end of the primer and the remaining primer sequence can be complementary to the strand. Alternatively, assuming that the primer sequence has sufficient complementarity with the sequence or hybridizes with the target sequence, thereby forming a template for synthesis of the extension product, non-complementary bases or longer sequences are , Can be interspersed in the primer.

  A cell has been “transformed” by exogenous or heterologous DNA when such DNA is introduced inside the cell. The DNA to be transformed may or may not be integrated (covalently linked) into the cell's genome. For example, in prokaryotic cells, yeast cells, and mammalian cells, the transforming DNA can be maintained on episomal elements (eg, plasmids). With respect to eukaryotic cells, a stably transformed cell is a cell in which the transforming DNA is integrated into the chromosome so that the transforming DNA is inherited by daughter cells through chromosome replication. This stability is demonstrated by the ability of eukaryotic cells to establish cell lines or clones consisting of a population of daughter cells containing transforming DNA. A “clone” is a population of cells derived from a single cell or common ancestor by mitosis. A “cell line” is a clone of a primary cell that can stably grow for many generations in vitro.

  A “heterologous” region of a DNA construct is an identifiable DNA segment in a larger DNA molecule that is not found in nature in association with the larger molecule. Thus, if the heterologous region encodes a mammalian gene, this gene is usually flanked by DNA that is not adjacent to the mammalian genomic DNA in the genome of the source organism. In another example, the coding sequence is a construct in which the coding sequence itself is not found in nature (eg, a cDNA in which the genomic coding sequence contains an intron, or a synthetic sequence that has different codons than the native gene). Allelic variants or naturally occurring mutational events do not result in a heterologous region of DNA as defined herein.

  Numerous polypeptide sequences that can be fused to proteins and specifically bound by antibodies are known and can be utilized as epitope tags. These include, for example, HA (SEQ ID NO: 1), c-myc (SEQ ID NO: 2), DTYRYI (SEQ ID NO: 3), TDFYLK (SEQ ID NO: 4), EEEEYMPME (SEQ ID NO: 5), KPPTPPPPEPET (SEQ ID NO: 6), HHHHHH (SEQ ID NO: 7), RYIRS (SEQ ID NO: 8) and DYKDDDDK (SEQ ID NO: 9).

  Antibodies are protein molecules produced by the human immune system that recognize foreign substances, bind to foreign substances, and mediate the elimination of foreign substances from the body. Techniques have been developed to take advantage of highly specific antibodies for cancer diagnosis and treatment. For example, tethering radioisotopes or toxins to antibodies that bind to tumor cells concentrated such cytotoxic agents against diseased tissue while leaving the surrounding tissue relatively intact. It is possible to deliver a medication. Antibodies are also important tools in biotechnology and are extensively used for analytical purposes (eg, for quantifying trace amounts of materials and isolates) and for purifying the desired biological product from complex mixtures. Used.

  In these applications, both the strength of binding of the antibody to its target (affinity) and the selectivity with which the antibody binds only to that particular target (specificity) are critical. For this reason, protein technicians seek to change and improve the binding properties of specific antibodies. A rational approach to the structural design of antibodies has met with limited success. And the methods available for random screening have serious limitations.

  The mammalian immune system approach to the problem of fine-tuning antibody affinity relies on a process called “affinity maturation”. Here, the cycle of mutation and evolutionary selection produces antibodies that bind more tightly to their targets. The present invention discloses a powerful new system for manipulating antibody affinity and antibody specificity by constructing microbial analogs of the B cell repertoire of the mammalian immune system. Antibodies were displayed on the surface of yeast cells by genetic fusion with cell wall proteins. After mutation, variants were selected based on improved binding properties with fluorescently labeled targets.

  Yeast antibody display methods were tested by studying model antibodies whose physical and chemical properties were already well characterized. These methods are then applied directly to the actual antibody of interest. The genetic adaptation of yeast, the ease of growth of this microorganism, and the ability to alter antibody binding conditions in vitro, combine to give unprecedented control over antibody affinity and antibody specificity manipulation.

An advantage of the library method of the present invention is that it is particularly suitable for proteins such as antibodies. The most widely used method currently consists of “panning” of antibodies displayed on the surface of bacteriophages. Yeast display has several advantages over phage display. First, antibody-antigen binding need not be broken to recover closely bound variants. The harsh conditions required to break this bond in prior art methods can reduce the infectivity of the phage. Second, increased library diversity due to reduced clonal deletion is an advantage. It is well known that many antibody structures cannot be accurately processed by bacterial secretion devices. Yeast cells are eukaryotic and have a secretion pathway very similar to mammalian cells. Third, the present invention provides a more accurate and precise determination of antibody-antigen affinity. The presence of 10 ⁴ molecules per cell eliminates the stochastic variation that occurs for only a few molecules per phage. Finally, quantification of fluorescence by flow cytometry is a continuous measurement of surface-bound antigens without the need for logic-based affinity knowledge by comparison with two binding / free bisection methods in phage panning. I will provide a. Bacteria also have a lipopolysaccharide layer that acts as a polymer permeable barrier that prevents antibodies or proteins from accessing the displayed molecules.

The present invention discloses a surface display system for in vitro expression and selection of peptide and protein libraries on yeast. A nine residue peptide epitope (HA) was fused to the binding subunit of the yeast cell wall protein (AGA2) and then to the 4-4-20 anti-fluorescein single chain Fv. Selection was performed by flow cytometry on a mixture of cells with and without the presented fusion. A 600-fold enrichment was achieved with a single pass of sort. This system of the invention illustrates a process for in vitro affinity maturation of antibodies and a process for directed evolution of other proteins and peptides, which has the following advantages: (i) Panning A dual-label flow cytometry selection scheme that allows finer affinity discrimination; (ii) as many as 10 ⁴ copies per cell that eliminates stochastic variation in selection, and (iii) E.E. Library expression in yeast with altered or potentially improved expression bias that can produce clones deleted from the library expressed in E. coli.

  One object of the present invention is the manipulation of antibodies for improved affinity and specificity. To this end, antibody-hapten binding was studied through mutagenesis and screening of antibodies expressed on the outer cell wall of the yeast Saccharomyces cerevisiae. As an experimentally easy and genetically flexible eukaryote, yeast exhibits significant advantages over filamentous phage display as a platform for antibody expression and manipulation. In essence, microbial analogs of the mammalian immune system B cell repertoire have been constructed in vitro, allowing affinity maturation of antibodies to be performed under tightly controlled conditions of mutagenesis and selection. . As a result, antibodies with significantly improved affinity and specificity could be achieved.

  One aspect of the invention provides a method for selecting a protein having a desired binding property comprising the steps of: expressing a protein to be tested fused at its N-terminus with a yeast cell wall binding protein Transforming the yeast cell with a vector to be labeled; labeling the yeast cell with a first label, wherein the first label is associated with and tested with the yeast expressing the protein to be tested. Non-associating with yeast that does not express the protein of interest; selecting yeast cells with which the first label is associated; quantifying the first label, wherein the frequent occurrence of the first label is: An indication that the protein to be tested has the desired binding properties, and the low appearance of the first label indicates that the protein to be tested does not have the desired binding properties. .

  A preferred embodiment of the present invention further comprises the following steps: labeling yeast cells with a second label, wherein the second label is to be tested and is encoded by a vector. Associated with yeast expressing an epitope tag fused to a protein and not associated with yeast not expressing the epitope tag encoded by the vector; quantifying the second label, the appearance of the second label Shows multiple expression copies of the epitope-tagged protein to be tested on the surface of the yeast cell; and comparing the quantification of the first label to the quantification of the second label and the appearance of the second label Determining the appearance of the first label normalized to where the appearance of the first label is high compared to the appearance of the second label, wherein the protein to be tested is desired It is shown to have a slip characteristic, process.

  Another preferred embodiment of the invention includes the following steps: labeling the yeast 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 Quantifying the first label; labeling the yeast cells with the second label; quantifying the second label; and comparing the quantification of the first label to the quantification of the second label; Determining the appearance of the first label normalized to the appearance of the two labels, wherein the appearance of the first label is low compared to the appearance of the second label, the protein to be tested is desired A process showing that it has a binding property of

  Another aspect of the invention provides a vector for carrying out the method of the invention, which includes a cell wall binding protein fused to the N-terminus of the protein of interest. Preferred embodiments of this aspect of the invention include means for expressing a polypeptide epitope tag in yeast cells. A more preferred embodiment provides that the cell wall binding protein is a yeast agglutinin protein binding subunit, and even more preferably, the yeast agglutinin protein is Aga2p.

  Another aspect of the present invention provides a method for selecting a protein having enhanced phenotypic characteristics compared to that of a wild type protein, which method comprises the following steps: fusion with a yeast cell wall protein Transforming yeast cells with a vector expressing the protein to be tested, wherein the process is used to generate various populations of variants of the protein to be tested for mutagenesis; Labeling yeast cells with one label, wherein the first label associates with a yeast that expresses the protein to be tested and does not associate with a yeast that does not express the protein to be tested; Isolating yeast cells associated with one label; as well as analyzing the properties of the mutant protein expressed by the yeast and comparing the properties and ratios of the wild-type protein Comprising the steps of, wherein yeast cells displaying mutant proteins with enhanced properties than the wild-type protein is selected, step. As noted above, a second label and / or a third label can be used in this embodiment, and selection of a mutant protein of interest having enhanced phenotypic characteristics can involve repeated cycles of the enrichment and labeling steps. Can be included.

  In a preferred embodiment, the yeast cell wall protein is agglutinin; the protein to be tested is fused by its N-terminus to the C-terminus of the binding subunit of agglutinin; the yeast strain is Saccharomyces, Pichia, Hansenula, Schizosaccharomyces, A yeast strain of the genus selected from the group consisting of Kluyveromyces, Yarrowia and Candida; the protein to be tested is an antibody, Fab, Fv, or scFv antibody fragment, more preferably the protein to be tested A ligand binding domain of a cell surface receptor, even more preferably, the cell surface receptor is a T cell receptor.

  The object of the present invention is to provide a method for selecting mutant proteins exhibiting enhanced phenotypic characteristics selected from the group consisting of surface expression, stability, association constant, dissociation constant, secretion level and solubility. Where the variant of the protein to be tested may comprise a single mutation or multiple mutations.

  In yet another aspect of the invention, the second label can be used to quantitatively determine cell surface expression levels. This can be used as an assay to select other desired phenotypic characteristics such as intracellular expression, stability, association constant, dissociation constant, secretion level and solubility.

  The mutant protein selected by the method of the invention is cloned into a vector adapted for expression in eukaryotes, and the gene encoding the selected mutant protein is cloned; and the mutant protein is expressed in eukaryotes Can be further characterized. Here the enhanced properties of the mutant protein are confirmed by comparing the enhanced phenotypic properties of the mutant protein with those of the wild-type protein. Preferably, the eukaryote is selected from the group consisting of mammals, insects or yeasts. Since the mutant phenotype selected is an inherent property of a “new” (ie, mutated) protein, this approach can also be used to express the mutant in other non-eukaryotic expression systems. Applicable.

  Another preferred aspect of the invention provides methods for displaying proteins that are not displayed as their normal ("wild type") sequence. In the example shown, the T cell receptor for the antigen was not expressed as its “wild type” sequence. However, after random mutagenesis and selection by flow cytometry using appropriate configurationally specific antibodies, the mutant receptors were expressed on the yeast cell surface. This strategy allows the discovery of new T cell receptors and provides a method for presentation of virtually any polypeptide. Accordingly, the present invention also provides a method for selecting proteins for displayability on the surface of yeast cells comprising the following steps: a vector expressing a protein to be tested fused to a yeast cell wall protein. Transforming yeast cells, wherein mutagenesis is used to generate various populations of variants of the protein to be tested; in labeling yeast cells with a first label Wherein the first label associates with the yeast expressing the protein to be tested and does not associate with the yeast that does not express the protein to be tested; a yeast cell with which the first label is associated , By quantifying the first label, wherein the high frequency appearance of the first label indicates that the protein to be tested has the desired display properties It indicates that, and the appearance of a low frequency of the first label indicates the protein to be tested does not have the desired presentation characteristics, process. Preferably, the protein to be tested is an antibody, Fab, Fv, or scFv antibody fragment, or a ligand binding domain of a cell surface receptor. A typical example of a cell surface receptor is the T cell receptor.

  The following examples are given for the purpose of illustrating various embodiments of the invention and are not meant to limit the invention in any manner.

Example 1
(Medium / buffer)
In this specification, the following media / buffers were used.

(E. coli)
LB medium (1 ×): Bacto tryptone (Difco, Detroit, MI): 10.0 g; Bacto yeast extract (Difco): 5.0 g; NaCl: 10.0 g. Make to 1 L and autoclave. For plates, add 15 g / L agar and autoclave.

  Ampicillin stock: 25 mg / ml sodium salt of ampicillin in water. Filter sterilize and store in 4 ml aliquots at -20 ° C. [] = 35-50 mg / ml at work; 4 ml aliquot in 1 L → 100 mg / ml; 2 ml aliquot in 1 L → 50 mg / ml. The autoclaved LB is cooled to about 55 ° C. and then added to the LB.

SOC medium (100 mL): 2% Bacto tryptone: 2.0 g; 0.5% yeast extract (Difco): 0.5 g; 10 mM NaCl: 0.2 ml 5M; 10 mM MgCl ₂ : 1.0 ml 1M ; MgSO of 10mM _4: 1.0ml 1M; 20mM glucose: 0.36 g. Autoclave or filter sterilize.

(yeast)
(Minimum synthesis + casamino acid (SD-CAA) 500 mL)
Glucose (glucose) 10.00g
Amino acid-free enzyme nitrogen base (Difco) 3.35 g
Na ₂ HPO ₄ · 7H ₂ O 5.1 g
NaH ₂ PO ₄ .H ₂ O 4.28 g
Casamino acid (Trp-, Ura-) (Difco) 2.5g
Add distilled water to final volume. Filter sterilize and refrigerate. For plates, dissolve sodium phosphate and sorbitol to a final concentration of 1M in 400 ml distilled water. Add 7.5 g agar and autoclave. Glucose, N ₂ base and amino acid are dissolved in 100 ml distilled water and filter sterilized. After the autoclaved salt has cooled enough to touch, the filtered reagent is added.

(SG-CAA (induction medium) 500 mL)
Galactose 10.00g
Amino acid-free yeast nitrogen base (Difco) 3.35 g
Na ₂ HPO ₄ · 7H ₂ O 5.1 g
NaH ₂ PO ₄ .H ₂ O 4.28 g
Casamino acid (Trp-, Ura-) (Difco) 2.5g
Add distilled water to final volume. Filter sterilize and refrigerate.

  (Concentrated (YPD) (1000 mL); yeast extract: 10 g; peptone (Difco): 20 g; dextrose: 20 g. Add distilled water to 1 L. Autoclave.

TAE (Tris-acetic acid): Working solution: 0.04 M Tris-acetic acid and 0.001 M EDTA
Stock solution (50 ×) 242 g of Tris base in 1 L
Glacial acetic acid 57.1 ml
0.5M EDTA (pH 8.0) 100ml
TBE (Tris-borate): Working solution: 0.09 M Tris-borate and 0.001 M EDTA
Stock solution (5 ×) 54 g Tris base in 1 L
Boric acid 27.5g
0.5M EDTA (pH 8.0) 20.0ml
Stop buffer 10 × (restriction): 50% v / v glycerol; 0.1 M EDTA (pH 7.5); 1% w / v SDS; 0.1% w / v bromophenol blue. Combine all ingredients except the dye and bring the pH to 7.5 before adding the dye.

  Staining-ethidium bromide: 0.5 mg / ml in water; stock solution: 10 mg / ml. Stock solution dilution: 1/10 in TBE Add 100 ml dilution to 100 ml buffer.

  TBS working solution: 10 mM Tris-HCl, 140 mM NaCl and 5 mM EDTA. Sterilize by filtration.

(Example 2)
(Protocol: Replica plate method)
1. Select the appropriate material for the colony lift (make sure that the material is cleaned, dried and sterilized).
2. Mark the bottom of each fresh replica plate with an arrow and align the plates in a row. Also mark the starting plate.
3. Remove the lid of the starting plate, turn the plate over, and align the mark on the colony lift material with the bottom arrow. Lower the plate onto the surface of the material and gently press it across the plate. After the plate contacts the material, make sure it is not moving around. Remove the plate and place the lid again. A portion of the colony transferred to the material can be seen.
4). This procedure is repeated with one fresh replica plate. Make sure that the arrows are in line with the mark and that you are creating an exact replica. Put on light to see the small colonies transcribed.
5). Repeat the entire procedure for each replica plate to be made.
6). Incubate the replica plate at conditions appropriate for selective growth. Colonies usually grow in a day or so.

Example 3
(Protocol: Electrical transformation of yeast)
Cell preparation:
1. Inoculate 50 ml of YPD with overnight culture to an OD of 0.1.
2. Cells are grown at 30 ° C. with vigorous shaking to an OD ₆₀₀ of 1.3-1.5 (about 6 hours).
3. Recover with a rotor cooled at 3500 rpm for 5 minutes at 4 ° C. Discard the supernatant.
4). Cells are washed thoroughly by resuspending in 50 ml of cold sterile water. Centrifuge as above and discard the supernatant.
5. Repeat step 4 with 25 ml of cold water.
Resuspend in 6.2 ml ice cold sterile 1M sorbitol. Centrifuge as above and discard the supernatant.
Resuspend in 7.50 ml ice-cold 1M sorbitol. The final volume of cells is approximately 150 ml (sufficient for 3 transformations).

Electrical transformation:
1. Place a 0.2 cm cuvette and white slide chamber on ice.
2. In an Eppendorf tube, add 50 ml of yeast suspension and gently mix with <5 ml (0.1 mg) of plasmid DNA in TE. Make sure to add DNA to the yeast already in the Eppendorf tube. Place on ice for 5 minutes (this time frame is quite important).
3. Set GENE PULSER to 1.5 kV and 25 mF. Set the pulse controller to 200W. The time constant for this pulse should be 4.5-5.0 msec.
4. Transfer 40 ml of the cell / DNA mixture to a pre-cooled electroporation cuvette. Take the contents to the bottom and check that the sample is in contact with both aluminum sides of the cuvette. Place the cuvette in a cooled safety chamber slide. Push the slide into the chamber until the cuvette contacts the electrode at the base of the chamber.
5). Apply one pulse with the above settings.
6). Release the cuvette and slide and immediately add 1 ml of cold 1M sorbitol to the cuvette. Mix and return cuvette to ice. Spread 200 ml onto a selection plate containing 1 M sorbitol.

Example 4
(Protocol: E. coli transformation)
1. Thaw aliquots of competent subcloning efficient HB101 (-70 ° C storage) on ice and keep all reagents on ice. DH5α cells are used for lac z complementation; DM-1 is used for unmethylation.
2. 50 ml in the required number of Eppendorf tubes, 1 for each DNA sample, 1 for pBR322 (positive control) (pUC19 for DH5α and DM-1), 1 for no DNA (negative control) Distribute HB101.
3. Unused cells are aliquoted into 50 mL portions and refrozen in a dry ice / ethanol bath; stored at -70 ° C.

  4. Add 1 ml of DNA (1 mg) to Eppendorf (1 ml pBR322 or pUC19) and tap to mix. Then incubate on ice for 30 minutes. For ligation, use 1-2 ml of ligation mixture (excess amount makes transformation difficult).

  5. Heat shock for 20 seconds at 37 ° C (45 ° C for 42 seconds for DM-1 cells).

  6). Place on ice for 2 minutes, then add 0.95 ml of room temperature SOC medium. Incubate for 1 hour at 37 ° C in a bath (optionally shaking) or on a shaker at room temperature of 37 ° C. .

  7. Seed 100-200 mL of cells in LB (100 mg / mL ampicillin) and incubate overnight at 37 ° C.

(Example 5)
(Protocol: GELase (registered trademark) DNA purification)
Wizard ™ PCR prep (Promega; Madison, Wis.) Is an alternative protocol for DNA purification from gels. If the yield of Wizard (TM) prep is low, GELase (TM) is recommended. If the desired fragment is less than 200 kb or longer than 5 kb, the yield is low.

  1. DNA fragments are separated on a 1% low melting point agarose gel in fresh 1 × TAE buffer.

  2. The gel is stained in water containing ethidium bromide. Using a portable UV lamp, cut out the desired fragment with a new razor blade.

  3. This gel slice is placed in a pre-weighed Eppendorf tube. Both weights are measured again to determine the weight of the gel slice. If the gel slice weight exceeds 300 mg, after step 6, the sample is divided into two tubes.

  4). Add 2 ml of 50 × GELase® buffer per 100 mg of gel slice.

  5. The tube containing the gel slice is incubated at 70 ° C. until the gel is completely melted. This takes at least 20 minutes. A desirable technique is to hold for 30 minutes, raise and lower the mixture several times with a pipette, and then hold for another 10 minutes. Make sure the gel is completely melted.

  6). Equilibrate the melted gel at 45 ° C for at least 30 minutes.

  7). Add 1 U of GELase (R) per 150 mg of melted gel. Incubate for 4 hours. If more than 600 mg, add 2 U of GELase (R).

  8. Add 1 volume (1 volume = mg of gel slice) 5M ammonium acetate to this solution. If a gel heavier than 300 mg is used, a larger tube is required for ethanol precipitation.

  Add 9.2 volumes (1 volume = gel slice + mg of ammonium acetate) 100% ethanol at room temperature and invert several times.

  Pellet DNA by centrifuging at 10.19 RAL for at least 30 minutes at room temperature. If the DNA concentration is very low, add ethanol and hold for 30 minutes, then centrifuge for 30 minutes.

  11. Remove and discard the supernatant with a pipette.

  12 The pellet is washed with 70% alcohol at room temperature.

  13. Dissolve DNA in water or TE. The DNA can then be stored at -20 ° C.

(Example 6)
(Protocol: Concatenation)
Materials: T4 DNA ligase and 2 × T4 DNA ligase buffer phosphatase treatment: Calf intestinal phosphatase (CIP) and buffer (if required). Blunt end treatment: dNTP mixture (0.5 mM). E. Klenow fragment of E. coli DNA polymerase I or T4 DNA polymerase. Ligation treatment: oligonucleotide linker-0.2 mM DTT.

  1. In an 20 ml reaction solution, cleave individual DNA components with the appropriate restriction endonuclease. After the reaction is complete, the enzyme is inactivated by heating to 65 ° C. for 15 minutes. If no further enzymatic treatment is required, go to step 6.

  2. If one 5 'phosphate of DNA is to be removed, add 2 ml of 10x CIP buffer and 1 U of CIP and incubate at 37 ° C for 30-60 minutes. After the reaction is complete, the CIP is inactivated by heating to 75 ° C. for 15 minutes. If no further enzymatic treatment is required, go to step 6.

  3. For blunt end treatment, add 1 ml of solution (containing all 4 dNTPs (0.5 mM each) and appropriate amount of E. coli DNA polymerase I or K4 fragment of T4 DNA polymerase) Perform trimming reaction. After completion, the enzyme is inactivated by heating to 75 ° C. for 15 minutes. If an oligonucleotide linker is to be added, proceed to step 4. If a DNA fragment containing only one blunt end is desired, the reaction product is cleaved with an appropriate restriction endonuclease. If no further enzymatic treatment is required, go to step 6.

  4. Add 0.1-1.0 mg of the appropriate oligonucleotide linker, 1 ml of 10 mM ATP, 1 ml of 0.2 M DTT, and 20-100 sticky end units of T4 DNA ligase and incubate at 15 ° C. overnight. Inactivate the ligase by heating to 75 ° C for 15 minutes.

  5. Cleave the product of step 4 with a restriction enzyme that recognizes the oligonucleotide linker (adjust buffer conditions if necessary). If only one of the two ends is to contain a linker, the product is cleaved with additional restriction enzymes.

  6). If necessary, the desired DNA segment is isolated by gel electrophoresis. It is then purified (GeneClean II ™ or GElase ™).

  7). Ligation: 9 ml DNA component (0.1-5 mg), 20 ml water, 4 ml 5 × ligase buffer, 1 ml (sticky end) T4 DNA ligase (BRL: 1 unit = 300 sticky end units, 20 ~ 500 sticky end units required). Incubate at 16 ° C. for 1-24 hours.

  8.1 ml of the ligated product was transformed into E. coli. E. coli cells are selected and transformants are selected. A miniprep and restriction enzyme mapping is then performed and screened for the desired product.

(Example 7)
(Cloning :)
All transformants were E. coli according to the manufacturer's protocol below. in E. coli strain DH5α.

PCR
Ampliwax PCR-gem mediated hot start PCR (Perkin Elmer Cetus, Norwalk, CT)-manufacturer protocol for thin wall tubes GeneAmp PCR center reagent (Perkin-Elmer-Cetus)
DNA thermal cycler 480 (Perkin-Elmer-Cetus)
AGA2
Cloned by PCR.
The template for PCR is CEN BANK S. cerevisiae genomic library (American Type Culture Collection, Rockville, MD).
Primer: 5′-ATTAGATCTCACTACTCATACATTTTCAA-3 ′ (SEQ ID NO: 10)
And 5'-ATTACTCGAGCTATTTACTGCAGagcgttagctggaacgtcgttatgggtaAAAAAACATACTGTGTGTTTATGGGG-3 '(SEQ ID NO: 11).
・ Thermal profile:
Denaturation 1 minute at 94 ° C. Annealing 2 minutes at 41 ° C. (first 5 cycles), 2 minutes at 45 ° C. (an additional 25 cycles)
Extension 25 seconds at 72 ° C. Last polishing step 72 ° C. 10 seconds The PCR product was transferred to plasmid pCR-Script using the pCR-Script SK (+) cloning kit (Stratagene, La Jolla, Calif.) According to the manufacturer's protocol. Cloned. The 342 bp AGA2 fragment was excised with EcoRI and XhoI, purified on a 1% agarose gel (protocol 6.1.7.2) and subcloned into pCR-Script containing the CUP1 promoter as a KpnI / EcoRI fragment.

HA peptide The HA peptide was inserted by cassette mutagenesis. A complementary oligonucleotide chain encoding the factor Xa recognition sequence and the HA epitope was synthesized with an attachment overhang that allowed ligation to the 3′XhoI site of the AGA2 clone while simultaneously destroying this site; pCR-Script An internal downstream SacI site was annealed and ligated to the CUP1-AGA2 construct in pCR-Script. This insertion contains a new XhoI site at the 3 ′ end of the HA sequence. CUP1-AGA2-HA was excised as a KpnI / XhoI fragment, purified on a 1% agarose gel and subcloned into the yeast shuttle vector pRS314 (1) already containing the α-factor termination sequence, and the surface display vector pCT101 was Formed. Oligo sequence:
5'-TCGACGATTGAAGGTAGATACCCCATACGACGTTCCAGACTACGCCTCTGCAGTAATAGAATTATCCCTCGAGCT-3 '(SEQ ID NO: 12) and 5'-CGAGGATAATCTATTACTGCAGAGCGTAGTCTGTACTG-3GTCGTCATCTGGGTCATCCT-3

  The GAL promoter was excised from the vector YCplac22-GAL. A 12 bp palindromic linker with appropriate attachment overhangs was first cloned into this vector, changing the restriction sites at both ends: EcoRI → KpnI (E / KLINK) and BamHI → EcoRI (B / ELINK). The resulting KpnI / EcoRI fragment was cloned into pCT101 to form vector pCT201. Oligonucleotide sequence: E / LLINK 5'-AATTGGTACC-3 '(SEQ ID NO: 14); B / ELLINK 5'-GATCGAATTC-3' (SEQ ID NO: 15).

4-4-20 scFv was amplified by PCR as described above:
Template: 4-4-20 in GeneX vector (obtained from D. Kranz (UIUC Dept. of Biochem.))
Primers: 5′-ggttggccagactagcGACGTCGTTATGACTCAA-3 ′ (SEQ ID NO: 16) and 5′-ggccggccactacgagttattacattacttttctagagaagattttttgttGTGAGGCTGATGGATGGATGGATGGATGGATGGATGGATGGATGGATGGATGGATGGATGGAG
・ Thermal profile:
Denaturation 1 minute at 94 ° C. Annealing 2 minutes at 40 ° C. (first 5 cycles), 2 minutes at 48 ° C. (an additional 30 cycles)
Extension 50 seconds at 72 ° C. Last polishing step 10 minutes at 72 ° C.

The PCR product was cloned into pCR-Script and subcloned into pCT201 as a NheI / XhoI fragment using the method described above to create vector pCT202. The vector pCT302 was generated in-frame between AGA2 and the 4-4-20 open reading frame of pCT202 by inserting a synthetic oligonucleotide (UIUC Biotechnology Center) encoding a (Gly ₄ -Ser) ₃ linker. .

AGA1
Amplified by PCR as above:
・ Template: CEN BANK
Primers: 5′-ATTAGATTCAGCTCAAAAAAACCAAAAAAT-3 ′ (SEQ ID NO: 18) and 5′-ATTACTCGAGcttaTTAACTGAAAATTACATTGC-3 ′ (SEQ ID NO: 19)
・ Thermal profile:
Denaturation 1 minute at 94 ° C. Annealing 2 minutes at 41 ° C. (first 5 cycles), 2 minutes at 45 ° C. (an additional 25 cycles)
Extension 2 min 20 sec at 72 ° C. Last polish step 10 min at 72 ° C. The PCR product was gel purified using the GELase® kit. The KpnI / SstI fragment of pCT201 was cloned into the vector pRS316. This was then digested with EcoRI and XhoI, the excised AGA2 fragment was removed by gel purification of the remaining vector fragment and ligated to the purified AGA1 PCR product and digested with EcoRI and XhoI. The resulting clone was pCT211. The KpnI / SstI fragment of pCT211 was substantially cloned into vector YIplac211 to form vector pIU211.

(Example 8)
(Expression in yeast)
Yeast strain, S. cerevisiae BJ5465 (ura3-52 trp1 leu2D1 his3D200 pep4: HIS2 prb1D1.6R can1 GAL). This strain is a pep4 and prb1 mutant, creating a mutant lacking a protease. Three nutritional markers were removed and could be used for plasmid selection: URA3, TRP1, and LEU2. HIS3 was removed, but PEP4 removal is supplemented with a HIS marker.

Transformation:
Vector pIU211 was cut with BsiWI, which is uniquely present in the AGA1 sequence. Approximately 100 ng of this linearized vector and 200 ng of pCT202 were simultaneously transformed into yeast strain BJ5465 by electroporation (protocol 2.6.1). Transformants were selected on SD-CAA plates.

Experimental induction conditions:
A single colony of yeast transformed with pIU211 and pCT202 was inoculated into 3 ml of SD-CAA and grown at 30 ° C. for about 24 hours. At this point, the cell density was about 10 ⁷ to 10 ⁸ cells per meter (ie, OD ₆₀₀ was about 1 to 3). A sufficient amount of cells was collected by centrifugation and inoculated into 3 ml SG-CAA so that the starting OD ₆₀₀ was ˜0.5. The culture was grown at 30 ° C. for about 20 hours.

Example 9
(Fluorescent labeling of yeast cells)
Fluorescence labeling of yeast cells was performed using the following method:
1. After culturing for 20 hours in SG-CAA, 0.2 OD-ml (600 nm) cells are harvested by centrifugation at 14,000 g for about 10-30 seconds.

  2. The cell pellet is washed by resuspending in TBS and centrifuging.

  3. The pellet is resuspended in an appropriate volume of TBS to bring the final volume to 100 ml for incubation. Add the following quantities of labeling reagents as appropriate: 1 ml 25 mg / ml FITC-dextran (MW 2,000,000) (Sigma); 1 ml 9E10 Mab ascites (Babco) or 10 ml at 100 mg / ml 9E10 (Santa Cruz Biotechnology); 10 ml / ml 100 ml 12CA5 Mab (Boehringer-Mannheim) in TBS. Mix by vortexing or pipetting up and down.

  4). Incubate for 1 hour at room temperature and mix cells approximately every 20 minutes by flicking the tube with your finger, vortexing, or pipetting.

  5. The cells are spun down and resuspended in an appropriate amount of TBS to a final volume of 100 ml. The following amounts of second reagent are added as appropriate amounts: 4 ml mouse-PE (Sigma); 2 ml mouse-FITC (Sigma); 1 ml FITC-dextran.

  6). Incubate for 30 minutes at room temperature.

  7). Centrifuge and wash as in step 2.

8). The pellet is resuspended in about 100 ml of 10 mM Tris base, pH 8.3 (for any labeled with FITC) or TBS (for microscopy). For flow cytometry, suspended in 10 mM Tris of 500 ml (final cell density of about 10 ⁶ / ml or more). Samples should be in 0.5 ml microcentrifuge tubes for flow cytometry. For experiments using biotin-fluorescein, cells are grown, induced, harvested, and labeled with 10 mM biotin-fluorescein as the primary label as a substitute for FITC-dextran as described, and As a secondary labeling reagent, it was labeled with a mixture of 3 mg streptavidin-PE and 1 mg RED613-conjugated goat anti-mouse F (ab ′) ₂ (Life Technologies, Grand Island, NY).

(Example 10)
(Confocal fluorescence microscope)
Yeast containing plasmid-directed surface expression of HA peptide (pCT201) or scFV fusion (pCT202) is grown for 20 hours in medium containing 2% galactose as the sole carbon source and labeled with mAb 9E10 substantially It was then labeled with secondary anti-mouse IgG-R-phycoerythrin (PE) conjugate and FITC-dextran as described. Labeled cells were mounted on polylysine-coated slides in 90% glycerol mounting media containing 1 mg / ml p-phenylenediamine as an anti-fade reagent and laser confocal fluorescence microscopy (UIUC Beckman Institute Microscopy Suite) And analyzed at a magnification of 63 × objective lens at a rate of 8 seconds. Images from DIC, red PE fluorescence, and green FITC fluorescence were collected.

(Example 11)
(Flow cytometry analysis and classification)
The labeled yeast cell suspension was analyzed on a Coulter Epis XL flow cytometer at the flow cytometry center of UIUC Biotechnology Center. The event rate was maintained at approximately 500 cells / second. The population was gated by light scattering, avoiding examination of aggregated cells, and collecting data on 100,000 events. For initial cell sorting experiments, the yeast carrying the pCT202 vector was mixed with the untransformed parent strain BJ5465 and based on the FITC signal, the Coulter 753 cell sorting bench (UIUC flow cytometry center) modified with CICERO sorting electronics. ). Samples before and after classification were seeded on non-selective medium, and replicas were then seeded on medium selective for the pCT202 vector. Purity was determined as the non-selective colony fraction that could survive on selective plates.

Example 12
(Quantification of surface antibody expression level)
Cells carrying vector pCT202 and Quantum Simply cell beads (Sigma, St. Louis, MO) are labeled with FITC-conjugated mAb 12CA5 (Boehringer Mannheim, Indianapolis, IN) (10 g / ml in TBS) as described. And analyzed on a Coulter Epis XL flow cytometer. Comparison of the fluorescence intensity of yeast samples with standard beads allowed determination of the antibody binding capacity of the displayed yeast cells by linear regression using Quantum Simply Cellular (Sigma) QuickCal.

(Example 13)
(Kinetic analysis of antigen dissociation from cells presenting scFv)
Yeast cells harboring plasmid pCT202 were grown and labeled with anti-c-myc mAb 9E10 and FITC-dextran or biotin-fluorescein as described. A fraction of the labeled population was analyzed by flow cytometry to determine the initial level of fluorescence. A non-fluorescent competitor (5-aminofluorescein) was added at a final concentration of about 10 mM (about 1000-fold excess) and FITC or PE fluorescence of the c-myc positive cell population was room temperature (21-23 ° C.) with Coulter Epis XL. And tracked as a function of time. Data were fitted as exponential decay. The possibility that a multivalent antigen binds to a cell as a function of time is P = 1− (1-e ^−kt ) ^N (N is the valence, K is the kinetic rate constant for dissociation, and t is the time). For long time t, this decreases to P = Ne- ^kt . Thus, the estimation of long t to 0 hour data yields a fluorescence intensity of P = N or F _ext = NF ₀ . Where F _ext is the fluorescence estimated upon addition of the competitor and F ₀ is the actual initial fluorescence. Therefore, the valence of the interaction between the surface-presented scFv 4-4-20 and the multivalent FITC-dextran was determined as the y-intercept of the curve in FIG.

Binding to soluble fluorescein (FDS) was assayed by observing fluorescence quenching in the vicinity of whole cells presenting scFv. Cells are suspended in TBS + 0.1% BSA at 2 × 10 ⁷ cells / ml in a quartz cuvette thermostatted at 23 ° C. and titrated with FDS ranging from 0 to 7.5 nM. Fluorescence at 520 nm was observed with an SLM Aminco SPF-500 spectrofluorometer using an excitation wavelength of 488 nm. Control cells presenting an inappropriate scFv were titrated to obtain a slope for a two parameter fit to the equilibrium binding data model to obtain an equilibrium constant and effective scFV concentration. Following equilibrium titration, 5-aminofluorescein was added to 1 mM and the change in fluorescence of the sample was followed over time to determine the k _off for FDS.

(Example 14)
(Mutagenicity of scFv gene)
About 100 ng of pCT302 was prepared in duplicate according to the manufacturer's protocol. E. coli strain XL-1Red (Stratagene, La Jolla, Calif.). After 1 hour induction in SOC medium, two groups of transformants are pooled and 1/2000 of the pool is seeded on LB medium containing 100 mg / ml ampicillin to determine transformation efficiency. did. 5 ml of liquid LB medium (LB-AMP50-CARB100) containing 50 mg / ml ampicillin and 100 mg / ml carbenicillin was inoculated with the remaining transformants and grown overnight at 37 ° C. (OD ₆₀₀ approximately 1 0.0). A sufficient amount of the culture is collected, inoculated into 50 ml of LB-AMP50-CARB100 until OD ₆₀₀ = 0.01 in a baffled shake flask, and OD _{600 of} about 1.0-1 1. Grown at 37 ° C. until 1. Cells were harvested by centrifugation and used to inoculate 200 ml of LB-AMP50-CARB100 to OD ₆₀₀ = 0.001, and the culture was grown to 37 ° C. to an OD _{600 of} about 1.0. . Plasmid DNA was isolated by the QIAGEN ™ Maxiprep kit (QIAGEN ™, Santa Clarita, CA). The recovered DNA was again transformed into XL1-Red and the growth cycle was repeated three times and the resulting final product was subjected to approximately 90 generations of growth in the mutagenized gene strain.

(Example 15)
(Library expression and kinetic screening)
50 mg of the mutagenized pCT302 DNA was transformed into yeast strain EBY100 by the method of Gietz and Schiestl in 10 separate reactions. The product was pooled and a total of 1/2000 was plated on selective media to determine the total number of transformants. The residue was inoculated in 50 ml selective glucose medium, grown overnight at 30 ° C., passaged to OD ₆₀₀ = 0.1 and expanded 10-fold. Selective galactose medium (5 ml) was inoculated to OD ₆₀₀ = 0.5 and grown overnight at 30 ° C. to OD ₆₀₀ = 1.0-2.0. A sample of 10 ⁷ cells (1 OD ₆₀₀ ml) was labeled with FITC-dextran as described. Following labeling, cells are resuspended in 10 mM 5-aminofluorescein and 9E10 mAb for 20 minutes at room temperature, while simultaneously rinsing the samples with ice-cold buffer to stop the competitive dissociation of FITC-dextran and described Labeled with anti-mouse PE secondary antibody. Samples were sorted on a Coulter 753 bench with a sorting window as shown in FIG. 4 and an event rate of 4000 / sec. 6 × 10 ⁷ cells were tested during the first round of sorting and the window was set to collect 0.2% of the population. Harvested cells were regrown in glucose medium and switched to galactose as described before repeated competition and sorting. A total of 4 selections and amplifications were performed. 4 × 10 ⁷ cells were tested in the second round and 2 × 10 ⁷ cells were tested in each third and fourth round. The first and second rounds are performed in an enrichment mode to provide a high recovery of all positive clones, and the third and fourth rounds reject simultaneously occurring negative cells and enrich more Performed in a purification mode to achieve the factors. The fourth round product was plated to isolate individual colonies.

(Example 16)
(Establishment of fusion display system)
A gene encoding a peptide epitope tag fusion with a yeast cell wall protein was constructed and the surface expression of that epitope was altered. This cell wall protein, a-agglutinin, is involved in cell-cell adhesion during yeast conjugation. Thus, it is exposed on the external cell surface as a receptor for agglutinin on opposing mating cells. Pilot mixing and sorting experiments were performed to determine the single-pass and double-pass purification yields and purity for affinity screening by flow cytometry.

(Example 17)
(Yeast-conjugated agglutinin)
In the yeast life cycle, haploid cells occur as one of two possible mating types (a or α). When a-haploid cells and α-haploid cells are in physical contact, they adhere to each other through strong and specific interactions between cell surface adhesion proteins called “agglutinins”. Once bound in this manner, the cells fuse to form diploid cells. As a platform for antibody presentation on the yeast cell wall, a fusion of polypeptides to a-agglutinin subunits was constructed. Because the physiological role of agglutinin is to present protein binding sites outside the cell for specific, high affinity interactions with other proteins, artificial steric hindrance from yeast cell wall components Was minimal.

  As a eukaryote, yeast possesses a secretion device that is very homologous to that of antibody-secreting B lymphocytes. As a result, artificial clone-dependent inefficient secretion should be minimized by this host. Numerous studies have revealed significant homology between yeast and mammalian secretion pathways, such that certain molecules can be converted both in vitro and in vivo without significant loss of function. Mouse BiP expression is functionally replaced with yeast BiP, whose expression is essential for growth (Normington, 1989). Mammalian PDI expression is functionally replaced with yeast PDI, another essential yeast ER luminal protein (Gunther et al., 1993). Given the wide similarity between yeast secretion and mammalian secretion, clonal variability in antibody expression due to misfolding is substantially reduced in yeast compared to bacterial hosts. Should be done. In fact, active antibody folding, assembly, and secretion in yeast has been previously demonstrated (Wood et al., '85; Horwitz et al., '88).

Yeast a-agglutinin is synthesized as the following two subunits; Aga1p (which carries phosphatidyl-inositol-glycantail for fixation to the cell wall) and Aga2p (which has high affinity (K _D Binds to α-agglutinin at ≈1 nM) (Lipke and Kurjan, 1992). Aga1p and Aga2p are linked by an intersubunit disulfide bond, and Aga2p is released into the growth medium after incubation with a reducing agent such as DTT. While phosphatidyl-inositol-glycantale is generally localized to the membrane, substantial evidence has accumulated that Aga1p is linked to fibrous glucans in the cell wall by glycosyl transfer (de Nobel and Lipke, 1994).

(Example 18)
(Fusion structure)
In order to establish the possibility of using an agglutinin fusion to display the polypeptide on the yeast cell surface, an “epitope tag” peptide was first genetically fused to Aga2p. Extending this approach to antibody fusion is straightforward. Passage of scFv molecules through the yeast secretory pathway is efficient. Because folding, assembly, and secretion of activated IgG and Fab has been demonstrated in yeast (Horwitz et al., 1988; Wood et al., 1985).

  The AGA2 gene was cloned from a yeast genomic library by PCR and subcloned into an expression vector containing a strong copper-inducible CUP1 promoter that allowed for a 25-fold modification in expression level. The coding sequence for the influenza HA epitope tag (Tyr-Pro-Tyr-Asp-Val-Pro-Asp-Tyr-Ala) (SEQ ID NO: 1) was fused to the 3 ′ end of the AGA2 open reading frame, before There is a site-specific protease cleavage site (Ile-Glu-Gly-Arg ′) (SEQ ID NO: 26) of factor Xa. The DNA sequence of this construct is shown in FIG. Convenient restriction sites are included for in-frame fusions of single chain antibody genes.

  The Aga2p-HA fusion is simply an intermediate of the construct to the final Aga2p-HA-antibody fusion, which means that the fusion peptide is immobilized and accessible on the cell surface in this system. Provides a means to confirm. Immobilization of the HA peptide to the outer cell wall by fusion was confirmed by immunofluorescence staining of all unfixed cells using 12CA5 mAb (Boehringer Mannheim, Indianapolis, IN) and detected by flow cytometry and fluorescence microscopy (FIG. 9). Since the entire 12CA5 antibody molecule binds to the HA epitope without any destructive cell wall biochemical processing, Aga2p is accessible outside the cell for macromolecular recognition.

  As described above, the Aga2p binding subunit is attached to the cell wall through a disulfide bond to Aga1p, which is covalently immobilized with other cell wall components. Treatment with DTT negated the labeling with 12CA5. This indicates that the Aga2p-HA fusion is attached to the cell surface by a disulfide bond. The AGA1 gene was cloned by PCR and subcloned downstream of the GAL1 promoter. AGA1 expression was induced by switching to galactose growth medium.

The HA epitope tag is included in the antibody fusion to allow dual fluorescent labeling for both surface antibody levels and fluorescently labeled antigen binding. This approach decouples cell-to-cell variations in antibody expression levels from single-cell measurements of antigen affinity. For example, indirect immunofluorescence with phycoerythrin-labeled secondary IgG against a-HA monoclonal antibody provides a measure of surface antibody count, whereas fluorescein-labeled antigen that binds to the antibody has binding affinity. Provide measurement. Since approximately 10 ⁴ copies of a-agglutinin are presented per cell, the stochastic effect on binding measurements is minimal. Since a commercial flow cytometer can detect under 10 ³ fluorophore molecules, the signal to noise ratio should not be a problem. The ratio of green fluorescence (fluorescein, ie antigen binding) to red fluorescence (phycoerythrin, ie number of antibodies) is proportional to the fraction of antibody bound by antigen.

  To test possible separation factors by this method, cells expressing the HA epitope tag were mixed with wild-type cells at a defined ratio. This mixture of cells was labeled with fluorescein labeled a-HA IgG and the highest fluorescent subpopulation was selected by flow cytometry. Sorted fractions were re-cultured and the fraction of cells carrying the Aga2p-HA fusion was determined by replica plating for genetic markers associated with the expression vector. From this information, the single pass purification possible by this method was estimated.

(Example 19)
(Fusion of single-chain anti-fluorescein antibody to a-agglutinin)
Anti-fluorescein single chain antibodies based on the monoclonal antibody 4-4-20 have been constructed and characterized (Bird, '88; David et al., 1991). This single chain antibody is known to fold stably and retain affinity comparable to Fab fragments. The gene for monoclonal antibody 4-4-20 was fused to an existing AGA2-HA fusion gene for expression on the yeast cell surface.

  Monoclonal antibodies against fluorescein are a useful model system for physicochemical studies on antibody-hapten interactions. Antibody-antigen binding kinetics (Kranz et al., 1982), thermodynamic analysis of complexation (Herron et al., 1986), role of electrostatic interactions (Omelyanenko et al., 1993), and site-directed mutagenesis of antibody binding sites (Denzin et al., 1993) have been studied using this system.

  The functionality of the proposed 4-4-20 fusion was determined by binding to a variety of fluorescein-labeled dextrans (Sigma). Since antibody binding quenches fluorescein luminescence, the detected fluorescein represents a moiety attached to dextran that is bound to yeast through 4-4-20 conjugated fluorescein.

(Example 20)
(Development of presentation skeleton)
Yeast possesses two related cell surface receptors known as a-agglutinin and α-agglutinin. These two agglutinins function to mediate cell-cell adhesion between a haploid cells and alpha haploid cells prior to fusion to form diploids (Lu, 1995). ). α-Agglutinin has been shown to be covalently linked to the cell wall glucan at the C-terminus (Lu, 1995; Schreuder, 1993) and a-agglutinin is thought to be immobilized by a similar bond ( Lu, 1995). Fusion of a-agglutinin to the C-terminal portion has been used previously to immobilize enzyme and viral antigens on the yeast surface (Schreuder, 1993).

  As a model system for the development of yeast surface display library screening methods, we have used yeast by fusion to a-agglutinin (which is a glycoprotein of two subunits unlike α-agglutinin). Functional anti-fluorescein scFv and c-myc epitope tags are presented on the cell wall (Fig. 2). A 725-residue Aga1p subunit anchors the assembly to the cell wall through a β-glucan covalent bond (Lu, 1995) (Roy, 1991); a 69 amino acid binding subunit Aga2p is linked to Aga1p by two disulfide bonds (Cappellaro, 1994). Natural a-agglutinin binding activity is localized at the c-terminus of Aga2p (Cappellaro, 1994); thus, it is a molecular domain with accessibility to extracellular macromolecules, and a tethered protein for presentation Represents a useful site. A vector for the displayed protein was constructed as a C-terminal fusion to Aga2p (FIG. 3).

(Example 21)
(Confirmation of scFv expression and surface localization)
Expression of the Aga2p-scFv fusion is directed by the inducible GAL1 promoter (Johnston, 1984). Yeast growth in glucose medium allows essentially complete repression of transcripts from the GAL1 promoter and allows important considerations to avoid counterselection against sequences that negatively affect host growth. Switching cells to media containing galactose induces the generation of Aga1p and Aga2p fusion gene products that are relevant in the secretory pathway and exposed to the cell surface. The surface localization of the Aga2p-scFv fusion has been confirmed by confocal fluorescence microscopy and flow cytometry. Cells simultaneously labeled with anti-c-myc mAb and fluorescein-conjugated dextran (FITC-dextran) were examined by laser scanning confocal microscopy (FIG. 8). Control cells carrying vectors directed to the presentation of unrelated peptides (ie hemagglutinin (HA) epitope tag alone) are not labeled with c-myc epitopes or mAbs specific for FITC-dextran (FIG. 8A). .

In contrast, cells carrying the surface display vector pCT202 expressing the Aga2p-scFv-c-myc fusion were co-labeled with both anti-c-myc antibody and FITC-dextran (FIG. 8B) and the antigen binding site was Demonstrate that it is accessible to very large polymers. Both of these strains stained positively with mAb 12CA5 directed against the HA epitope tag. The accessibility of the fusion for binding to both intact IgG (150 kDa) and 2 × 10 ⁶ Da dextran polymer indicates that there is no significant steric hindrance from cell wall components. An important advantage in connection with the E. coli surface presenting proteins embedded within the lipopolysaccharide layer that forms a barrier to macromolecular diffusion.

  Two-color flow cytometric analysis of these yeast strains similarly demonstrates the scFv that was readily accessible on the cell surface. Negative control and scFv presenting strains were simultaneously labeled with anti-c-myc mAb 9E10 and FITC-dextran. The bivariate histogram demonstrates a linear relationship between the intensity of phycoerythrin fluorescence (level of mAb 9E10 binding) and FITC fluorescence (antigen binding) for the cell population carrying the 4-4-20 presenting plasmid. On the other hand, the control population shows background fluorescence (FIGS. 9A and B). The distribution of fluorescence intensity within the positive fraction illustrates the importance of correcting the antigen binding signal for cell-to-cell variability in the number of displayed fusions, as determined by epitope tag labeling.

Quantification of presentation efficiency by comparison of scFv-presenting cell populations with calibration standards for known antibody binding capacity yields an average value greater than 3 × 10 ⁴ fusions per cell. Treatment of cells displaying Aga2p-scFv fusion with dithiothreitol before labeling eliminated staining of the cell surface with both FITC-dextran and mAb 9E10 (FIG. 9C), which is expressed as a recombinant Aga2p subunit. This is consistent with the attachment of the fusion protein to the cell surface by specific disulfide bond interactions with Aga1p. This feature illustrates another important feature of the yeast display system, namely that the protein can simply be released from the cell surface by reduction for further characterization.

To further test the specificity of the 4-4-20 / fluorescein interaction, a competitive dissociation assay was performed using a non-fluorescent analog of fluorescein (5-aminofluorescein). Analysis of these data shows that for FITC-dextran, a monovalent dissociation rate constant (k _off ) of 3.7 × 10 ⁻³ / sec at 21 ° C. and 3.9 for fluorescein-biotin. A monovalent dissociation rate constant of × 10 ⁻³ / sec is obtained. The exponential fit extrapolation for t = 0 seconds shows that the average valence of FITC-dextran molecule interaction with scFv is less than 1.5. Similar results were obtained using fluoresceinated inulin, fluorescein-conjugated bovine serum albumin, and fluorescein-biotin as competitors, indicating that labeling of cells with FITC-dextran or fluorescein-biotin It was shown that due to the specific interaction between the presented fusion and the fluorescein moiety. Furthermore, the dissociation kinetics of surface-presented 4-4-20 scFv-derived fluorescein disodium salt (FDS) is derived from the soluble 4-4-20 scFv-derived yeast produced as observed by spectrofluorimeter. Consistent with the dissociation kinetics of FDS.

(Example 22)
(Concentration of presentation cells by cell sorting in flow cytometry)
To test the effectiveness of flow cytometry selection using yeast surface display, a yeast mixture lacking the selectable marker associated with the yeast carrying the surface display vector was selected and the purity was determined independently by replica plate method. Decided. Important enrichment factors (up to 600 times) were obtained (Table I). Thus, rare clones can be selected from a yeast display library by first enriching positive cells with moderate stringency and high yield to provide a smaller population, and then rare clones Can be subjected to a more stringent sort of several passes to isolate.

(Example 23)
(Isolation of scFv variants with lower k _off from yeast-displayed mutagenesis libraries)
E. The selection of scFv genes randomly mutagenized by growth in a “mutagenizing gene” strain of E. coli has been described (Low, 1996). A library of approximately 5 × 10 ⁵ 4-4-20 scFv variants, generated by propagation of yeast surface display vectors in such strains, was expressed in yeast. The pool of cells displaying the scFv library was subjected to kinetic selection by competition of FITC-dextran labeled cells with 5-aminofluorescein. C-myc positive cells showing the highest proportion of FITC relative to PE fluorescein were collected by flow cytometry sorting (FIG. 10A) and amplified by regrowth under fusion-inhibited conditions (glucose carbon source) The fusion presentation was induced and rescreened. Cells demonstrating a substantial increase in the duration of labeling with FITC-dextran were dramatically enriched after three rounds of sorting and amplification (FIG. 10).

The FITC-dextran dissociation kinetics for two individual clones selected from the scFv library were 2.9-fold different compared to the wild type 4-4-20 scFv (FIG. 11). The rate constant of the mutant was 1.9 × 10 ⁻³ / sec (mutant 4M1.1) and 2.0 × 10 ⁻ at 23 ° C. compared to 5.6 × 10 ⁻³ / sec for the wild type. ³ / sec (4M1.2); from similar experiments, 5.0 × 10 ⁻³ / sec, 2.4 × 110 ⁻³ / sec, and 2.8 × 10 ⁻³ / sec fluorescein, respectively. The k _off value for biotin was obtained. Furthermore, the soluble fluorescein dissociation kinetics determined by spectrofluorimeter showed a 2.2-fold improvement for both mutants compared to the wild type, and the initial balanced fluorescence quenching experiment showed that the affinity of the binding reaction A similar improvement in sex constant is suggested. Isolation of clones with an off-rate reduced by only a factor of 3 demonstrates the ability of this screening method to achieve accurate quantitative discrimination.

Of the 26 selected clones analyzed individually, 2 are equally improved at k _off (4M1.1 and 4M1.2 above); 2 are c- distorting the linear expression level / activity relationship Demonstrate wild type k _off with a decrease in myc label; one demonstrates wild type k _off and c-myc label; and 21 are monovalent with an apparent k _off about 10 times lower than wild type It bound only to the multivalent 2 × 10 ⁶ Da FITC-dextran, but not to FITC-dextran or fluorescein-biotin. Enrichment of clones with increased avidity results from the use of multivalent antigens (approximately 90 fluorescein per dextran); avidity effects are efficiently avoided by appropriate design of screening conditions that ensure monovalent antigen binding. obtain. In addition, the selection of epitope tag variants can be achieved by alternately detecting expression levels with c-myc and HA tag tags in successive rounds of selection, or alternatively targeting changes only to the scFv gene. Can be eliminated by a common mutagenesis strategy.

  These results indicate that scFv fragments can be displayed on the surface of yeast in a manner accessible to macromolecular recognition and in a manner that is amenable to combinatorial library construction and screening. The presented scFv specifically binds to the antigen—the first demonstration of a functional antibody fragment presented on the yeast cell surface. Application of this display system to library methods for in vitro antibody affinity maturation, and library methods for display of other mammalian proteins includes phage display, bacterial surface display, and yeast two-hybrid method Is an important complementary alternative to existing technologies such as. Indeed, the success of the very first attempt of the yeast display system in recovering improved fluorescein-binding scFv variants from a relatively small library under non-optimized screening conditions is the robustness of this technology. Clearly demonstrate. Proven highly quantitative kinetic analysis of surface-bound scFv and good identification of clones with similar binding properties further demonstrate the important potential of yeast display for protein combinatorial optimization .

(Example 24)
(Presentation of antibody to T cell receptor in yeast display system)
Herein, a scFv (KL16) specific for the Vb8 region of the T cell receptor (Roehm et al., 1985) was expressed in the yeast display system. This scFv-KJ16 inhibits T cell activity by competitively blocking the recognition of TCR ligands (eg, the superantigen staphylococcal enterotoxin B (Cho et al., 1995)). Since this affinity variant of scFV can show enhanced T cell inhibition, the use of this yeast display system in manipulating the higher affinity form of scFv-KJ16 was tested.

A screen based on equilibrium binding of cell surface scFv to a fluorescently labeled antigen, Vb8 single chain TCR (Schlueter et al., 1996) was developed. Using two-channel flow sorting, selection was also based on the binding of a fluorescently labeled anti-c-myc antibody to a 10-residue c-myc tag at the carboxy terminus of the scFv. A scFv variant with a higher affinity for the TCR or a scFv variant with a lower affinity for the anti-c-myc antibody was isolated. As expected, the former has a mutation in CDR (V _L CDR1), and the latter had a mutation in c-myc epitope. Thus, these discoveries can be used either for the yeast display approach to either isolate higher affinity scFvs or to identify epitopes of display proteins that are recognized by specific Mabs. Demonstrate that you get.

(Plasmids and strains)
The scFv-KJ16 V _L and scFv-KJ16 V _H genes linked by a modified 205 linker (Cho et al., 1995) are transferred into the vector pCR-Script (Stratagene, La Jolla, Calif.) According to the manufacturer's protocol. Subcloned by PCR. A c-myc epitope tag was included at the carboxy terminus of the scFv. An approximately 800 bp NheI / XhoI fragment containing scFv was excised from pCR-Script and ligated into the AGA2 open reading frame downstream of the yeast surface display vector pCT202 containing the 9-residue epitope tag (HA) and the inducible GAL1 promoter. did. The resulting construct was transformed into S. cerevisiae containing AGA1 integrated into the chromosome, controlled by the GAL1 promoter. cerevisiae strain BJ5465 (ura3-52 trp1 leu2D1 his3D200 pep4 :: HIS2 prbD1.6 can1 GAL; Yeast Genetic Stock Center, Berzley, Calif.) (Strain EBY100).

(Example 25)
(Induction and detection of scFv-KJ16 on the yeast surface)
Yeast cells transformed with pCT202 / scFv-KJ16 were shaken in 3 ml of selective glucose medium SD-CAA (glucose 2 wt%, Difco yeast nitrogen base 0.67 wt%, casamino acid 0.5 wt%) Grow overnight at 30 ° C. After about 18-20 hours, recombinant AGA1 + AGA2-scFv expression was shaken in 5 ml selective galactose medium (SG-CAA, where 2% galactose replaces glucose in SD-CAA) Induction at 20 ° C. Cultures, after about 20-24 hours (1-2培増), collected by centrifugation, 0.1% bovine serum albumin and 0.05% including azide _{PBS (10mM NaPO 4, 150mM NaCl} , pH7. 3) and 25 ml of 10 mg / ml anti-HA Mab 12CA5 (Boehringer Mannheim, Indianapolis, Ind.), Anti-c-myc Mab 9E10 (1: 100 dilution of ascites; Berkeley Antibody Co., Richon) Or E. Incubated with biotinylated sc-TCR (Schodin et al., 1996) [about 360 nM] prepared from inclusion bodies expressed in E. coli for 45 minutes on ice. Cells were washed with PBS and FITC-labeled F (ab ′) ₂ goat anti-mouse IgG (1:50; Kirkegaard and Perry Labs, Inc., Gaithersburg, MD) or streptavidin-phycoerythrin (SA-PE) conjugate (1: 100; PharMingen, San Diego, Calif.) And incubated for 30 minutes on ice. Labeled yeast cells were analyzed on a Coulter Epis XL flow cytometer at the Flow Cytometry Center of the UIUC Biotechnology Center. The event rate was approximately 250 cells / second. Data for 10,000 events were collected and the population was gated according to light scatter (size) to prevent cell mass analysis. These conditions were also used to produce equilibrated antigens that bound isothermally after incubation of scFv-KJ16 yeast with various dilutions of scTCR. Perform scattering analysis, using the average fluorescence units obtained directly from the estimated concentration and flow data of biotinylated scTCR, to determine K _D values.

(Example 26)
(Generation of scFv-KJ16 random mutation library)
Approximately 50 ng of pCT202 / scFv-KJ16 was prepared according to the manufacturer's protocol. E. coli XL1-Red cells (Stratagene, La Jolla, Calif.) were transformed. Following 1 hour induction in SOC medium, the harvest was centrifuged at 2000 rpm for 5 minutes and 500 ml liquid LB medium (containing 100 mg / ml ampicillin and 50 ml / ml carbenicillin) (LB− AMP100-CARB50). This resuspension was added to 15 ml LB-AMP100-CARB50 in a 50 ml Erlenmeyer flask and grown at 37 ° C. with shaking. The culture was supplemented with fresh 15 ml of LB-AMP100-CARB50 in the mid-log phase (OD ₆₀₀ ˜0.2-0.4) and then grown to saturation (OD ₆₀₀ approximately 1.0-1.1; this was considered one “cycle” or one time of mutation). A small fraction (0.75 ml) of this culture was added to the next cycle (15 ml LB-AMP100-CARB50). After 6 cycles of growth, preparation of Wizard Miniprep (Promega, Madison, Wis.) DNA plasmid was performed in 15 ml cultures. Approximately 4.5 ml of pCT202 / scFv-KJ16 DNA from cycle 6 was transformed into each of the three tubes of yeast strain EBY100 using the LiAc method (Gietz et al., 1995). After the three reactions were pooled and resuspended in 1 ml ddH ₂ O, 1/2000 pools were placed on selection plates to determine transformation efficiency. 50 ml of SD-CAA is seeded with the residue of this culture, grown overnight at 30 ° C. with shaking, passaged to OD ₆₀₀ = 0.05 and at 30 ° C. to OD ₆₀₀ > 1.0. Grow overnight. 5 ml of SG-CAA was then seeded at an OD _{600 of} about 0.5 and grown overnight at 30 ° C. with shaking to OD ₆₀₀ = 1.0-2.0.

(Example 27)
(Selection of scFv-KJ16 mutation library by FACS)
Cells were double labeled with anti-c-myc Mab and biotinylated scTCR (used at a concentration of about 10 nM) as described above. The reaction volume was adjusted to maintain an approximately 10-fold molar excess of antigen (scTCR) relative to the surface scFv. Samples were separated on a Coulter 753 bench equipped with a sort window as shown in FIG. 3 and at an event rate of 4,000 cells / second. A total of 8 × 10 ⁷ cells were examined during the first separation round with 0.1-0.4% of the collected population. Harvested cells were regrown in SD-CAA at 30 ° C. and switched to SG-CAA before the next separation. The total amount of 4 separations is performed in the enrichment mode (high recovery of all positive clones) in the first 2 separations and in the purification mode (rejection of concurrent negative cells). ) Immediately after the final separation, the collected cells were re-separated and placed on a selection plate to isolate individual clones.

(Example 28)
(Rescue and sequencing of mutant scFv-KJ16 gene)
A plasmid (wild type and two mutants) from scFv-KJ16 yeast was used except that the cells were destroyed using a bead beater (BioSpec Products, Inc., Bartlesville, OK) for 2 minutes instead of vortex. Rescue according to the protocol described in Ward, 1991). The cells were centrifuged for 1 minute and the upper phase (aqueous phase) was collected. Plasmid DNA was prepared using the Wizard® miniprep kit (Promega, Madison, WI) and E. coli. E. coli DH5α competent cells (GibroBRL, Gaitherburg, MD) were transformed with 1 ml DNA preparation using the CaCl ₂ method. Transformants were plated on LB-AMP50. Sequencing of wild-type scFv-KJ16 and two mutants (mut4 and mut7) was performed using primers flanking the display vector scFv and a fluorescence automated sequencing instrument (Genetic Engineering Facility of the UIUC Biotechnology Center). .

(Example 29)
(TCR binding by yeast cell surface scFv)
Monoclonal anti-TCR antibody KJ16 recognizes a conformational epitope on the Vb8 chain of TCR (Brodnicki et al., 1996). KJ16 has been used in a number of in vivo studies in mice, including attempts to target and delete the Tb Vb8 population (Born et al., 1987, McDuffie et al., 1986, Roehm et al., 1985). To assess possible effects of altering antibody affinity in mediating these effects, the use of a yeast display system to identify KJ16 variants with increased affinity for TCR was tested. ScFv gene from the anti-TCR antibody KJ16 is cloned previously and the scFv protein exhibited almost the same affinity as KJ16Fab fragment (K _D approximately 120 nM) (Cho et al., 1995).

  The scFv-KJ16 coding sequence was subcloned to be expressed as a fusion polypeptide with an Aga2p agglutinin subunit expressed on the yeast cell surface. This fusion polypeptide contains a hemagglutinin (HA) epitope tag at the N-terminus of the scFv and a c-myc epitope tag at the carboxy terminus. The inclusion of these epitopes indicates that monoclonal anti-HA (12CA5) and anti-c-myc (9E10) antibodies can be used in flow cytometry to quantify surface expression of full-length scFv, regardless of antigen binding activity. enable. Such normalization helps reveal cell-to-cell variability effects at the surface level of the fusion polypeptide. As discussed below, the utility of two independent epitope tags can also be controlled for the selection of individual epitope variants that may not be desired to be screened for ligands that bind to the variant. To assess the binding properties of cell surface scFv, soluble single chain Vb8-Va3 TCR (Schodin et al., 1996) was biotinylated and bound ligand was detected using a phycoerythrin-streptavidin conjugate.

  FIG. 12 shows that yeast transformed with scFv-KJ16 / Aga2 plasmid expressed HA epitope (FIG. 12A) and c-myc epitope (FIG. 12B). Control yeast transfected with the Aga2p / HA expression vector alone was positive for anti-HA Mab but not for anti-c-myc antibody. The fraction of cells in the non-fluorescent population was found to depend on plasmid stability and culture growth phase, but the physiological processes involved are unknown. Nevertheless, the induction temperature is reduced to 20 ° C. and the induction time is reduced to less than the time that the two cultures double, producing a population with more than 75% of the cells displaying scFv-KJ16. To do. scFv-4-4-20 was displayed using this system with approximately the same percentage of positive cells.

  Binding of biotinylated scTCR to cell surface scFv was also detected by flow cytometry (FIG. 12C). The fraction of cells in which the expressed active scFv was similar to that detected using anti-HA and c-myc antibodies was consistent with full-length expression, correctly folded scFv. In addition, the two-color histogram demonstrated a strong correlation of scTCR binding with both HA and c-myc epitope displays. Biotinylated scTCR binding was specific for yeast displaying scFv-KJ16 and was completely inhibited by excess soluble KJ16 IgG (FIG. 12D).

The approximate affinity of the scFv-KJ16 displayed on the surface was determined in situ on the cell wall by titration of whole cells with altered concentrations of biotinylated scTCR. Equilibrium binding was measured by analyzing scTCR bound to cells by flow cytometry. Scatchard analysis of the binding data (Figure 13) to give a K _D of 500 nM, which was within 5 times a K _D observed for soluble scFv-KJ16. _The K _D, likely be overestimated, because 100% of the scTCR was calculated under the assumption that were active, such matching is reasonable (i.e., only 20% correctly folded If the surface scFv has a K _D of approximately 100 nM). Previously, solubilized E. coli. It was found that a substantial fraction of scTCR purified from E. coli inclusion bodies was incorrectly folded (Schodin et al., 1996).

(Example 30)
(Selection of mutated scFv-KJ16 / yeast by fluorescence activated cell sorting)
E. E. coli mutants were used to mutagenize scFv for affinity maturation by phage display (Low et al., 1996). This approach was successful in identifying scFv-4-4-20 variants with high affinity for fluorescein using yeast display. The advantage of this mutagenesis approach is its simplicity, ie, E.I. only need E. coli transformation and cell growth. Furthermore, this E.I. The E. coli mutagenesis gene line introduces mutations through the expression plasmid, and therefore does not change biased towards a portion of the scFv, which may be important for determining binding characteristics. Whether this aspect of mutagenesis of a mutagenic gene line is convenient depends on the ability to identify key residues that can affect antigen binding based on available structural information. Tests of published affinity mutation studies suggest that the position of such residues (generally at non-contact residues) is not yet a priori predictable (Hawkins et al., 1993, Patten et al., 1996, Schier et al., 1996, Thompson et al., 1996, Yang et al., 1995, Yelton et al., 1995).

To apply this strategy for scFv-KJ16, scFv-KJ16 / Aga2 plasmids were transformed into E. coli for 6 cycles of growth. E. coli mutagenesis gene line XL1-Red (Stratagene). This procedure was expected to introduce an average of 2-3 mutations in the scFv coding sequence based on the mutation rate per cycle at 2000 bp. The resulting plasmid preparation was transformed into yeast producing a library size of approximately 3 × 10 ⁵ transformants. In another study, a larger library ^(107), depending on further optimization of transformation procedures, and were obtained by independent transformants are pooled. This number does not represent an upper size limit for the library construct, but scale-up can be applied directly as a further optimization attempt.

  The mutagenized yeast library was subjected to 4 successful separation cycles and amplification using double staining for anti-c-myc antibody binding (FITC) and biotinylated scTCR binding (PE). Biotinylated TCR was used at a 1: 5000 dilution (approximately 10 nM) produced slightly below the detectable threshold for binding by wild type scFv-KJ16 / yeast (FIG. 13). Shown are two channel fluorescence profiles of mutated scFv-KJ16 samples after one separation cycle (FIG. 14A) and after four separation cycles (FIG. 14B). Cells showing fluorescence on the diagonal window shown in FIG. 14 were collected for repopulation. The principle of this diagonal window was based on antigen binding per fusion of polypeptides for which the separation criteria were displayed at any given time. For example, selection based solely on higher PE fluorescence levels (ie, scTCR binding) includes not only those variants with higher affinity scFv, but also variants that display a higher density of scFv per yeast cell. It is. The latter mutant is in principle removed by inducing anti-c-myc antibody as one of two parameters to normalize for surface expression variability. To perform the first two separations in an enriched manner, to isolate about 0.5% of the cell population with the highest fluorescence and to avoid coincidence (two cells in droplets separated simultaneously) No separation software was set up. This last two separations were performed for purification to avoid high agreement. After 4 cycles, the cells were immediately re-separated and plated. Ten colonies (mut 1-10) were selected for further analysis.

(Example 31)
(Characterization of mutant scFv-yeast)
Each of the 10 selected mutants was labeled with anti-HA antibody, anti-c-myc antibody, and biotinylated TCR and analyzed by flow cytometry (FIG. 15). As can be expected, one clone (mut6) appeared phenotypically similar to wild type scFv-KJ16 / yeast. Another clone (mut7) was found to show a higher TCR binding level and the results were confirmed by several independent titrations. Finally, many mutants (mut1-5, 8, 9) consistently showed reduced binding to anti-HA antibody or anti-c-myc antibody compared to biotinylated scTCR binding. The presence of this class of variants can be explained by the details of the diagonal separation window, as shown in FIG. Cells either increase the scTCR (PE) binding with a constant c-myc (FITC) signal, or alternatively decrease the c-myc (FITC) binding with a constant scTCR (PE) signal. Depending on what can be moved to the separation window. The selection of these variants can be easily avoided by using both HA and c-myc, which are epitope tags in the fusion. Thus, by substituting the label for each of these epitope tags in each round of separation, the reduced binding to one of the epitope tags is not enriched in successive rounds of separation as in this case.

  The fluorescence histograms of the putative c-myc epitope mutant (mut4), scTCR binding mutant (mut7) and another mutant (mut10) were compared to the wild type scFv (FIG. 16). mut4 (FIGS. 16A and 16B) showed a decrease in anti-c-myc labeling, mut7 showed enhanced scTCR binding (FIGS. 16C and 16D), and mut10 showed no change in either but positive The fraction of cells that was was higher relative to the wild type scFv (FIGS. 16E and 16F). As shown in FIGS. 16E and 16F, nearly 100% of mut10 cells were positive for each drug tested. This contrasts with each other variant similar to wild-type scFv-KJ16 yeast that shows two distinct cell populations that reduce the level of cell surface scFv (see, eg, mut4 and mut7). . mut10 enhanced plasmid stability and mut10 to E. coli. Repeated failures to rescue the expression plasmid into E. coli suggest that chromosomal integration has occurred with this mutant plasmid. Thus, the altered surface expression characterization of mut10 appears to be the result of integration of the expression plasmid.

The binding affinity for scTCR was assessed for the variants shown in FIG. 16 by titration with soluble biotinylated scTCR (FIG. 17). Non-linear curve fitting of the data, while indicating K _D unchanged for mut4 and mut10, indicating that the affinity is increased 3-fold for mut7. The increase in average fluorescence of mut10 is due to the absence of non-fluorescent ends in this distribution, rather than increased scTCR binding, as is apparent in FIGS. 16E and 16F.

(Example 32)
(Sequencing of mutant scFv)
The nucleotide sequences of wt-scFv-KJ16 cloned into the yeast display plasmid and mut4 and mut7 following rescue of the yeast-derived plasmid were determined (FIG. 18). Wild-type scFv-KJ16 contained two silent changes from the originally published scFv sequence (Cho et al., 1995). These could be introduced by PCR prior to scFv cloning into the yeast display plasmid. The mut4 sequence contained one mutation and mut7 contained two mutations. The only mutation in mut4 was present in the c-myc epitope (Lys to Glu), consistent with decreased binding by anti-c-myc antibody as described above. mut7 contained a change from Arg to Lys in the framework region of the _VL region and a change from Ser to Arg in CDR1 of the _VL chain. The latter mutation was consistent with the higher binding affinity observed for mut7.

  Phage display was used for selection of scFv with higher antigen binding affinity as well as isolation of new scFv from untreated libraries (Hoogenboom, 1997). However, in part, due to toxicity, codon bias, or folding problems, E. coli. There were differences in the expression of several mammalian proteins in E. coli (eg, Knappik and Pluckthum, 1995, Ulrich et al., 1995, Walker and Gilbert, 1994). Yeast expression potentially has some of these by proposing the advantage that proteins can be expressed with eukaryotic post-translational modifications such as glycosylation and efficient disulfide isomerization. The problem can be prevented in advance. Furthermore, phage display generally does not have quantitative accuracy to distinguish between variants with different binding affinities less than 5 times (Kretzschmar et al., 1995). In contrast, fluorescent labeling and separation allowed the isolation of 4-4-20 scFv clones with an affinity increased only 3 fold. Identify minor improvements in affinity since the maximum change in antigen binding affinity is caused by a directed combination of point mutations each having a smaller effect (Hawkins et al., 1993, Schier et al., 1996, Yang et al., 1995) The ability to do can be a significant value. In view of these advantages, the use of a yeast display system for affinity maturation of the anti-T cell receptor scFv has been developed.

  A scFv that is specific for the Vb8 region of mouse TCR was used to generate anti-TCR reagents that could ultimately enhance T cell targeting properties in vivo (Cho et al., 1997, Cho et al., 1995). . The active scFv was expressed as an Aga2p fusion protein on the surface of yeast with an affinity similar to native scFv (approximately 500 nM compared to 120 nM for scFv). In order to select a higher affinity scFv, Random mutagenesis using E. coli DNA repair deficient strains mutated at a frequency of about 2-3 per 1000 base pairs after 6 growth cycles. Flow cytometry using fluorescent labeled scTCR and anti-c-myc antibody was used to separate cell display scFv with increased scTCR affinity. The anti-c-myc antibody controls for variants with increased TCR binding, not for higher affinity but for higher cell surface expression of the scFv-c-myc fusion. Included as 2 criteria. After multiple selections, three mutant phenotype classes were observed: 1) decreased binding to c-myc antibody but did not change scTCR binding (mut1-5, 8, 9); 2) c-myc labeling Enhanced binding to scTCR without changing (mut7); and 3) more efficient surface expression by chromosomal vector integration (mut10).

  Isolation of the class of mutants represented by mut4 and mut7 can be predicted from the selection criteria illustrated in FIG. That is, either an increase in scTCR (PE) signal or a decrease in c-myc (FITC) signal places the cell in the separation window, so any mutant cell identified on the diagonal separation window boundary is It can be explained by any of the properties described for mut4 and mut7. This is due to the availability of two independent epitope tags, but does not show substantial problems with this approach. By utilizing HA and c-myc tags in changing separation cycles, progressive enrichment for the reduced labeling of one epitope tag should not occur.

  Isolation of epitope tag variants highlights additional applications for yeast surface display: mapping of epitopes recognized by monoclonal antibodies. Although an alternative strategy using a peptide library is successful for linear epitopes in this regard, the approaches described herein can be extended against conformational epitopes. Thus, correctly folded proteins can be displayed on yeast cells, and direct random mutagenesis as described herein can be used to identify epitope residues from non-contiguous polypeptide sequences. Can be applied. Since unfolded proteins are retained and degraded by the eukaryotic secretion regulator and various expression levels are identified by HA or c-myc labeling, misidentification of epitope residues can be achieved by this procedure. Should be minimized. This described approach is substantially easier than alanine scan mutagenesis.

  The reason why mut10 was enriched in this screen is not clear since the average single chain T cell receptor label of mut10 per c-myc label was not changed. It is possible to bias this clone for enrichment because of the higher fraction of clearly labeled cells due to a random excess to the separation window. In any case, neither scTCR tag or c-myc tag is different for this clone, and the structural rearrangement of the expression plasmid indicates that the expression plasmid has chromosomal integration.

The identification of the only unique CDR mutation in mut7 is consistent with the finding that this mutant scFv enhanced binding to the T cell receptor. Future attempts to obtain only scFv with higher affinity for the T cell receptor (and not the c-myc variant) are anti-HA antibodies and anti-c-myc to control for cell surface levels of scFv. Includes alternative selection using antibodies. This strategy, in combination with DNA shuffling techniques between selected variants (Stemmer, 1994), than the wild-type scFv (K _D approximately 120 nM), enables the isolation of scFv-KJ16 with significantly higher affinity Should be. Such mutant KJ16 scFvs can be used to test kinetic phenomena of T cell signaling as well as targeting T cell mediated killing via biantigen specific antibodies (Cho et al., 1997). .

The present invention demonstrates the intentional isolation of affinity matured antibodies via cell surface display. As described above, off-rate selection was used to identify mutants with a reduced dissociation rate, and thus balanced antigen binding was used in the expression of scFv-KJ16. These two approaches are complementary and depend on the affinity of the starting scFv. The K _D of greater than 1 nM, so to allow an effective identification of the variation of the reaction rate dissociation rate is too fast, it is reasonable to carry out strategies equilibration with soluble labeled antigen. Furthermore, at these lower affinities, large amounts of soluble antigen are not substantially depleted from the labeling reaction mixture if the displayed scFv is present at an effective concentration of about 1-10 nM. In contrast, strongly bound antibodies (eg, 4-4-20 (K _D = 0.4 nM)) allow soluble labeled antigen to be used at concentrations below K _D unless a large inconvenient labeling volume is used. Deficient. However, the dissociation kinetics for such tightly bound antibodies are sufficiently slow to allow quenching, sorting, and analysis via manual mixing procedures. Therefore, to obtain using a strategy whereby scFv is until K _D of about 1 nM, via the screening and repetition of the mutagenesis based on equilibration, followed by repetition of the off-rate screening and mutagenesis, affinity maturation And still further improvements can be obtained.

Cell surface display and flow cytometry screening kinetic binding parameters (e.g., K _D and the dissociation rate constant (k _diss)) allows the selection of the based library derived clones. The binding parameters of the selected mutants can then be quantitatively estimated in situ in a display format without the need for subcloning or soluble expression, as shown in FIG. In contrast, selection of phage-displayed antibodies often involves incremental stringent wash and elution conditions, even to the extent of pH 2 and 8M GuHCl. Such stringency selection has poor quantitative accuracy and is not always directly related to binding parameters (eg, K _D or k _diss ) under environmental or physiological conditions.

  A bacterial cell surface display system has been described for the manipulation of antibodies and other proteins (Gunnerison et al., 1996). Although these systems possess some of the advantages of the yeast display system of the present invention, they do not provide the post-translational processing capabilities of the eukaryotic secretory pathway. Macromolecular access to proteins displayed on bacteria can also be limited by the diffusion barrier presented by the lipopolysaccharide layer (Roberts, 1996). For this reason, binding to soluble protein antigens or epitope tags labeled with monoclonal antibodies is not possible. Surface display systems in cultured mammalian cells are also available, but the construction and screening of combinatorial libraries for these systems is not as rapid or versatile as for yeast.

A fairly small library (3 × 10 ⁵ ) was screened to isolate the variants described herein. This does not represent an upper limit on the yeast library size. Yeast libraries with 10 ⁷ clones have been constructed and further increases in library size can be achieved if necessary. The present invention shows that yeast surface display can be used to isolate mutant scFv with increased affinity and that mutants with altered mAb epitopes can be enriched or eliminated as desired. Further, K _D, without requiring subcloning and soluble expression can be estimated in situ in a display format. Quantitative optimization of the screening conditions allows further improvements in the method. The application of yeast surface display extends beyond antibody affinity maturation to isolation of binding domains from cDNA expression libraries, or isolation of mutant receptors or ligands based directly on kinetics and equilibrium binding parameters.

(Example 33: Displayability and expression of T cell receptor in yeast display system)
The invention also relates to a novel process for manipulating T cell receptors, for example with respect to improved binding properties to peptide-MHC complexes or superantigens. The present invention establishes a method for presenting T cell receptors in a yeast surface display library format. This method can be used for: 1) generally to express polypeptides that are not normally expressed on the surface or in yeast, and 2) more specifically, for higher affinity T cell receptors for selected ligands. For operating.

  Protein manipulation has not reached the level of development that allows rational and directed manipulation of increased affinity binding. As a result, approaches have been developed to identify improved mutants from large mutant populations. The most widely used approach is “phage display”, which has been used to manipulate antibodies, particularly in the form of linked “single chain” antibodies. However, phage display methodologies were unable to successfully display single chain T cell receptors (scTCRs). This may be due to the fact that the folding of the isolated single chain T cell receptor is very inefficient in the absence of other components of the CD3 complex and the protein folding mechanism of the eukaryotic endoplasmic reticulum. The bacterial periplasm cannot effectively fold these fragments.

  The establishment of the T cell receptor presented on the yeast surface is illustrated in FIGS. Significant improvements have been made to isolate mutant T cell receptors that can be presented in this system. The wild type T cell receptor is not functionally presented as shown by the absence of binding by an antibody (1B2) that is specific for the natural conformation of the T cell receptor (FIG. 19). Mutating the T cell receptor and screening the library for 1B2 binding identified mutant single chain T cell receptors presented in yeast. This establishes a system that can now be used to isolate mutant single chain T cell receptors with improved binding properties.

  The present invention provides a successful yeast cell-surface display system for expressing T cell receptors. Second, full-length T cell receptor expression can only be achieved after randomly mutagenizing the T cell receptor gene and then selecting by flow cytometry for surface expression. This method utilized an evolutionary approach to “correct” the expression defect in the T cell receptor.

  This same approach can be applied to any polypeptide that is not efficiently presented in its wild-type form. Selection for “displayability” reduced the practice for the T cell receptor, as described in Examples 33-37. Once mutant versions of the polypeptide that can be presented are obtained, these versions can then be subjected to a screening process for improved binding properties as described in Examples 1-32.

  Improved T cell receptor molecules are useful in the treatment of cancer, sepsis, and autoimmune diseases (eg, arthritis, diabetes, or multiple sclerosis). For example, soluble forms of high affinity T cell receptors act as antagonists of deleterious T cell mediated autoimmune diseases, thereby providing a potential treatment for these diseases. A similar strategy was used successfully for soluble tumor necrosis factor receptor (TNF-R). And this receptor form is in clinical trials for septic shock and rheumatoid arthritis (Moosmayer et al., 1995).

  In the methods of the present invention, yeast surface display allows single chain T cell receptors to be engineered to bind with high affinity to MHC peptide complexes or superantigens. Such molecules find a variety of medical uses. Examples include but are not limited to: 1) preventing inappropriate T cell attack on healthy tissue in autoimmune diseases (eg, arthritis, diabetes, and multiple sclerosis); 2) T cells Interfering with septic shock due to bacterial superantigens that interact with the bulk of the inflammatory response; and 3) using high-affinity T cell receptors with anti-CD3 bispecific agents to repopulate T cells Destruction of tumor cells carrying a T cell receptor ligand (eg, a specific tumor peptide / MHC complex) by attacking cancerous cells.

(Plasmids and strains)
A single chain TCR gene (V linked by a modified 205 linker (8.2-linker-V (3.1) gene (Cho et al., 1995)) was obtained by PCR according to the manufacturer's protocol by vector pCR-Script. (Stratagene, La Jolla, Calif.) A 6-His epitope tag was included at the carboxy terminus of the scTCR for purification purposes. And ligated to the yeast surface display vector pCT202 containing an AGA2 open reading frame downstream of the 9-residue epitope tag (HA) and the inducible GAL1 promoter. S. cerevisiae BJ5465 strain (aura3-52 trpl leu2D1 his3D200 containing AGA1 integrated into the chromosome controlled by the GAL1 promoter (strain EBY100) by the lithium acetate (LiAc) transformation method of stl (Gietz et al., 1995). pep4 :: HIS2 prbD1.6 canal GAL; Yeast Genetic Stock Center, Berkeley, CA).

(Example 34: Generation of scTCR random mutant library)
About 50 ng of pCT202 / scTCR E. coli XL1-Red cells (Stratagene, La Jolla, Calif.) were transformed according to the manufacturer's protocol. After 1 hour induction in SOC medium, the recovery was centrifuged at 2000 rpm for 5 minutes and resuspended in 500 ml liquid LB medium (LB-AMP100-CARB50) containing 100 mg / ml ampicillin and 50 mg / ml carbenicillin. It became cloudy. The resuspension was added to 15-ml LB-AMP100-CARB50 in a 50-ml Erlenmeyer flask and grown at 37 ° C with shaking. Cultures were supplemented with fresh 15-ml LB-AMP100-CARB50 at mid-log phase (OD ₆₀₀ (0.2-0.4)) and then grown to saturation (OD ₆₀₀ approximately 1.. 0-1.1; this was considered one “cycle” or one mutation). A small portion of this culture (0.75 ml) was added to the next cycle (15-ml LB-AMP100-CARB50). After 6 cycles of growth, Wizard® miniprep (Promega, Madison, WI) DNA plasmid preparation was performed on 15-ml cultures. Approximately 10 mg of pCT202 / scTCR DNA from the sixth cycle was transformed into each of 10 tubes of yeast EBY100 strain using the LiAc method. Ten reactions were pooled after resuspension in 1-ml ddH ₂ O / tube and 1 / 10,000 pools were plated on selection plates to determine transformation efficiency. The library size was about 7 × 10 ⁶ . A 50 ml volume of SD-CAA (glucose 2 wt%, Difco yeast nitrogen base 0.67 wt%, casamino acid 0.5 wt%) is inoculated with the remaining culture, cultured overnight at 30 ° C. with shaking, and OD _It was passaged to ₆₀₀ = 0.05 and grown overnight at 30 ° C. so that OD ₆₀₀ > 1.0. Then 5 ml of selective galactose medium SG-CAA (where 2% galactose replaces glucose in SD-CAA) is inoculated to an OD ₆₀₀ = 0.5 and about 20 ° C. Grow overnight with shaking for 20-24 hours (1-2 doublings).

(Example 35: Selection of scTCR mutant library by fluorescence activated cell sorting)
Cells were labeled with 25 mL Mab 1B2 (anti-Vb8.2Va3.1; prepared from ascites fluid and conjugated with biotin) at a concentration of 20 mg / ml. Samples were sorted on a Coulter 753 bench (Flow Cytometry Center, UIUC Biotechnology Center) at an event rate of approximately 4,000 cells / second. A total of 6 × 10 ⁷ cells were examined during the first round of sorting and approximately 5% of this population was collected. Harvested cells were regrown during selection at 30 ° C. in 4 ml selective glucose medium SD-CAA. After approximately 18-20 hours, recombinant AGA1 + AGA2-scFv expression was induced at 20 ° C. with shaking in 5 ml SG-CAA. A total of 3 sorts were performed, with the first sort being an enrichment mode (all positive clones with high recovery) and the last two rounds being a purification mode (rejecting simultaneous negative cells). Immediately after the final sorting, the collected cells were again sorted and collected as two separate populations (“high expression” and “low expression”) and plated on selection plates to isolate individual clones. Twenty clones were examined by flow cytometry.

Example 36: Induction and detection of mutant scTCR on yeast surface
Individual clones from the pCT202 / scTCR library selection were grown overnight in 3 ml SD-CAA with shaking at 30 ° C. and subsequently induced as described above in SG-CAA. Cultures were collected by centrifugation after 20-24 hours (1-2 doublings) and PBS containing 0.1% bovine serum albumin and 0.05% azide (10 mM NaPO ₄ , 150 mM NaCI, pH 7. 3) and incubated for 45 minutes on ice with 25 L of 10 mg / ml anti-HA Mab 12CA5 (Boehringer Mannheim, Indianapolis, Ind.) Or biotinylated 1B2 Mab (20 mg / ml) prepared from ascites did. Cells are washed with PBS and FITC-labeled F (ab ′) ₂ goat anti-mouse IgG (1:50; Kirkegaard and Perry Labs, Inc., Gaithersburg, MD) or streptavidin-phycoerythrin (SA-PE) One of the conjugates (1: 100; PharMingen, San Diego, Calif.) Was used for 30 minutes incubation on ice. Labeled yeast cells were analyzed on a Coulter Epis XL flow cytometer. The event rate was approximately 250 cells / second. Data for 10,000 events was collected and the population was gated according to light scatter (size) to prevent analysis of cell aggregation. Results from wild type (wt) TCR and several representative TCR variants are shown in FIG. Double mutants containing combined mutations from several of these isolates were also constructed and the flow cytometry results are shown in FIG.

Example 37: Rescue and sequencing of mutated scTCR gene
The plasmid from scTCR yeast (wt and 20 variants) was Ward (Ward, 1991) except that instead of vortexing the cells, a bead beater (BioSpec Products, Inc., Bartlesville, OK) was used for 2 minutes. Rescue according to the protocol described by The cells were centrifuged for 1 minute and the upper layer (aqueous layer) was collected. Plasmid DNA was prepared using the Wizard® DNA Clean-Up kit (Promega, Madison, WI) and E. coli. E. coli DH5α ElectroMAX® competent cells (GibcoBRL, Gaithersburg, MD) were transformed by electroporation with 1 ml DNA preparation. Transformants were plated on LB-AMP50. Sequencing of wt scTCR and 20 variants (mTCR1-mTCR20) was performed using primers adjacent to the display vector scTCR and fluorescence engineering sequencing of the UIUC Biotechnology Center. Single mutations were found in the TCR for each of the isolates shown in (Figure 21).

Example 38: Surface display and soluble expression of single TCR variants, double TCR variants, and triple TCR variants
The single site variant contained the following residues: VaL43P (mTCR7), VaL104P (mTCR16), and VbG17E (mTCR15). The combination of these mutations in double or triple mutants resulted in surface display of scTCR even at higher levels.

  To examine the mutant scTCRs selected by yeast surface display in more detail, single mutants (mTCR7, mTCR15, mTCR16), double mutants (mTCR7 / 15, mTCR7 / 16, mTCR15 / 16), and triple mutants (MTCR7 / 15/16) was cloned into a yeast secretion plasmid under the control of the inducible GAL 1-10 promoter, along with a synthetic prepro region based on the consensus signal sequence (Clements et al., 1991). The construct was expressed in yeast and the resulting supernatant was monitored in a quantitative 1B2 binding assay as a measure of correctly folded TCR protein. The expression level of the 1B2 active scTCR variant varied from an undetectable level for the wild type scTCR to the triple mutant that expressed the highest level (FIG. 22). The order of secretion levels was as follows: triple mutant> double mutant> single mutant> wild type. Selected scTCR double mutants were affinity purified using an anti-TCR antibody column (KJ16-Affigel) and the absolute secretion level using a low copy expression system was found to be about 100 mg / L. .

(Example 39: Comparison between TCR secretion level and surface display level)
In order to directly compare the level of scTCR secretion in this system with the level of TCR expressed in the surface display system, the same mutant was examined as an Aga-2 fusion in yeast using flow cytometry. The scTCR displayed on the surface of the yeast was labeled with an anti-HA antibody, followed by a fluorescein labeled secondary antibody and biotinylated 1B2, followed by streptavidin-phycoerythrin. The resulting fluorescence intensity of the yeast population was monitored by flow cytometry and determined as the average fluorescence unit (FIG. 23A). The level of fluorescence varied between single mutant, double mutant, and triple mutant, but the relative level was exactly the same as in the yeast secretion system (FIG. 23B) (ie triple mutant> Double mutant> single mutant> wild type). Thus, this yeast display system could identify these TCR variants produced at high levels in secretable expression systems by simply selecting those expressed at higher levels on the yeast surface.

(Example 40: Temperature stability of soluble TCR)
In order to explore the properties of the protein that could dominate the secretion (and presentation) level for the mutant scTCR, the stability of the mutant scTCR was investigated by performing heat denaturation. Yeast scTCR supernatants were incubated for 1 hour at various temperatures and protein activity was monitored as the percent of remaining 1B2 activity. Since wild type scTCR is not expressed in yeast systems, Wild-type scTCR refolded from E. coli inclusion bodies was used for comparison. This scTCR was used as a thioredoxin fusion protein (TRX-TCR). This fusion protein is expected to increase the stability and solubility of the protein. The results showed that the single mutant had a higher heat denaturation temperature than the TRX-TCR (FIG. 24A). Furthermore, the double and triple mutants were more stable and had a higher denaturation temperature than either the wild type TCR or the single mutant (FIG. 24A).

  The heat denaturation kinetics for mutant scTCRs were determined at 46 ° C. to compare the denaturation rates for each mutant (FIG. 24B). In addition, yeast scTCR supernatants and their 1B2 binding activity were monitored for TCR activity remaining after incubation for various times. The heat denaturation rate was highest for TRX-TCR and lowest for triple mutants, with single and double mutants in the middle range (FIG. 24B). Affinity purified mTCR15 / 16 had similar heat denaturation kinetics determined for mTCR15 / 16 measured directly in yeast culture supernatant. This indicates that the presence of endogenous yeast protein in the supernatant does not affect the measured denaturation kinetics.

  A direct correlation was observed when the amount of native TCR remaining after 1 hour at 48 ° C. was plotted against the secretion level for the mutated scTCR (FIG. 4 from Shusta et al. = FIG. 25). I need). Thus, in the case of TCR, the intrinsic property of stability was a reliable predictor of secretion efficiency and correlated directly with the yeast cell surface level of TCR. This correlation allows this display system to be used by selecting for higher surface level proteins to isolate mutant proteins that exhibit increased stability.

Example 41 Post-translational modification and yeast display
The method of the present invention has identified variants of 2C scTCR that are readily expressed in yeast and display significantly enhanced thermal stability. In fact, the identified single-site mutants have a scTCR that is normally unstable as stable as a covalently disulfide “stabilized” or chemically cross-linked single chain antibody fragment (scFv). It was converted to scTCR which seems to be. This is particularly important for TCRs, which have potential as antagonists in the treatment of autoimmune diseases. Heterologously expressed wild-type TCR has previously been very unstable and pharmacokinetics at physiological temperatures are important.

  It has been shown that glycosylation can result in increased secretion and thermal stability of the protein. For example, glucoamylase is E. coli. In the non-glycosylated form produced in E. coli, yeast S. It was less stable than the glycosyl form produced in cerevisiae. Therefore, the stability of scTCRs produced in yeast with one N-linked glycosylation site in the Vα region is It was possible to be greater than the stability of the non-glycosylated TRX-TCR produced in E. coli. However, this does not appear to be the only explanation as double and triple mutants have increased stability over single mutants and both species are glycosylated. Furthermore, although wild type scTCR was also expected to be glycosylated, expression of wild type scTCR in yeast was not detected.

  The expression level of various mutant scTCRs in the secretory system was completely correlated with the level of TCR on the surface in the yeast display system. This relationship can arise from the fact that both systems require the same secretory device. Thus, mutations that affect protein folding and stability can affect the same step of protein export to the outside (ie, cell surface or secreted). This is a clear advantage (as compared to bacteriophage or bacteria) of expressing eukaryotic proteins in yeast. This is because there is a qualitative control mechanism that ensures that only correctly folded proteins cross the entire exocytosis pathway.

  Mutant scTCR expression levels also correlated with the rate of heat denaturation. This effect is similar to that seen for bovine pancreatic trypsin inhibitor. In this effect, mutations that resulted in increased stability also led to increased secretion levels in yeast. This correlation indicates that proteins with increased stability in the folded form are more likely to be properly packaged and transported from the cell than are retained and degraded by eukaryotic quality control mechanisms. Support the theory that is higher. Of course, many other factors, such as forward folding rate, disulfide bond formation, and association with endoplasmic reticulum protein folding aids also play an important role. However, as the present invention demonstrates, thermostability appears to be a good measure of secretory capacity for molecules that were previously unstable and rarely secreted (eg, TCR).

  Based on the findings presented herein, a methodology has been developed that allows rarely expressed and / or unstable proteins to be systematically evolved to be expressed at higher levels and / or as more stable molecules. I can imagine. The gene encoding this protein can be randomly mutated and the library can be expressed on the surface of yeast. Once presented, this population can be screened for clones that exhibit high surface concentrations (high average fluorescence due to specific probes). As demonstrated for scTCR, variants with improved surface expression have the potential to be more stable molecules that are expressed at high levels. Thus, yeast surface display can be a powerful tool to increase protein stability for crystallization, industrial applications, and medical applications.

  The following references were cited herein:

  Any patents or publications mentioned in this specification are indicative of the level of ordinary skill in the art to which this invention pertains. These references are incorporated by reference to the same extent as each individual publication is specifically and individually incorporated by reference.

  Those skilled in the art will readily appreciate that the present invention is well adapted to carry out the objects and obtain the ends and advantages mentioned, as well as those inherent therein. The present examples with methods, procedures, molecules, and specific compounds described herein are presently representative of preferred embodiments, are exemplary, and are within the scope of the present invention. It is not a limitation to. Those skilled in the art will envision these modifications and other uses that fall within the spirit of the invention as defined by the claims.

  In order that the subject matter in which the above-cited features, advantages and objectives of the present invention are achieved will be understood in detail, by referring to specific embodiments thereof illustrated in the accompanying drawings. A particular description can be understood. These drawings form part of the present specification. It should be noted that the accompanying drawings illustrate preferred embodiments of the present invention and, therefore, should not be considered limiting of the scope of those embodiments.

Claims (1)

Invention described in the specification.

Images