Preparation of libraries of bispecific binders expressed in eukaryotic cells

Abstract

The preparation of a library of bispecific binders expressed in eukaryotic cells described herein is a method for producing a cloned library of eukaryotic cells containing DNA encoding a diverse library of bispecific binders, the present invention relates to bispecific conjugates, in particular bispecific conjugates comprising a first binding domain coupled to a first Fc domain and a second binding domain coupled to a second Fc domain.

Description

TECHNICAL FIELD

[0001] Aspects and embodiments described herein relate to the field of generation of eukaryotic cell clone libraries, in particular to eukaryotic cell clone libraries containing DNA encoding a diversified library of bispecific binders. BACKGROUND

[0002] Cancers and other complex diseases are often caused by a variety of factors and multiple signaling pathways. Targeting more than one molecule can be useful to circumvent the "escape" effect and avoid therapeutic resistance observed with monospecific antibodies. Bispecific antibodies have the potential to overcome these shortcomings, but their success is hampered by the complexity associated with producing the appropriate molecules for research and large-scale manufacturing purposes and the lack of suitable libraries of bispecific antibodies for functional screening. In fact: while more than 100 different bispecific antibody formats have been described, for most of them no reliable method for generating a diversified high-affinity antibody library has been shown. One of the main obstacles is related to the difficulties associated with successful expression of the correct format of bispecific antibodies. Depending on the bispecific antibody format, optimization is required to select two bispecific modules to achieve correct folding of both modules. In particular, the expressed bispecific antibody must still be fully functional, i.e. capable of simultaneously binding its two antigens. The functional activity of bispecifics, especially for bispecific antibodies that rely on simultaneous binding of 2 different targets, is difficult to predict from the functional activity of the single parent antibodies.

[0003] Currently available systems typically provide single antibody libraries against one target and do not allow to combine them in the final desired bispecific format. It is not obvious that when single specific antibody libraries against two different targets or two different epitopes on the same target are combined together in a particular antibody format, the corresponding bispecific molecule or biparatopic molecule shows simultaneous binding. The reason can be related to (1) steric hindrance of binding unit 1 to binding unit 2 to one or more targets or vice versa, or (2) the two antibody binding domains interfere with each other destroying the bispecific binding property. The method described in this application provides a solution to these problems.

[0004] The method described herein allows the use of mammalian display to display bispecific and multispecific antibodies in the final desired (therapeutic) format, thereby avoiding the need to investigate all individual members of such bispecific antibodies.

[0005] WO 2015 / 166272 describes a method for generating a eukaryotic clonal library containing DNA encoding a multi-binding library. WO 2015 / 166272 does not specifically address bispecific antibodies and does not disclose a method for generating a bispecific antibody library, particularly not a bispecific antibody comprising a first binding domain coupled to a first Fc domain and a second binding domain coupled to a second Fc domain. Summary of the Invention

[0006] Given the above, there remains a need for methods to generate eukaryotic cell clone libraries containing DNA that encodes a diverse library of bispecific binders in the ultimately desired (therapeutic) form.

[0007] Surprisingly, the method disclosed herein allows eukaryotic cells to successfully express and anchor heterodimeric bispecific antibodies to the cell surface, particularly for bispecific antibodies containing a first binding domain coupled to a first Fc domain and a second binding domain coupled to a second Fc domain. Specifically, as described elsewhere in this paper and in the experimental section, it has been found that bispecific antibodies, particularly those containing a first binding domain coupled to a first Fc domain and a second binding domain coupled to a second Fc domain, can be successfully displayed while retaining their full function of binding to both individual antigens (i.e., achieving proper expression and display on the cell surface, proper folding of the two binding domains, and avoiding steric hindrance or interference between the two variable domains). Therefore, the method disclosed herein enables the construction of bispecific mammalian display libraries.

[0008] In one aspect, a method is provided for generating a eukaryotic cell clone library containing DNA encoding a diverse library of bispecific binders, the method comprising:

[0009] - Provide eukaryotic cells;

[0010] - Provides a plurality of donor DNA molecules, each donor DNA molecule comprising a first nucleic acid sequence encoding a first binding domain coupled to a first Fc domain; and a second nucleic acid sequence encoding a second binding domain coupled to a second Fc domain;

[0011] - Introduce the donor DNA into these cells; and

[0012] - Culture recombinant cells to generate clones, thereby providing a eukaryotic cell clone library containing DNA encoding a bispecific binding library.

[0013] Preferably, the first and second Fc domains are engineered to promote heterodimerization. Therefore, a method is provided for generating a eukaryotic clonal library containing DNA encoding a diverse library of bispecific binders, the method comprising:

[0014] - Provide eukaryotic cells;

[0015] - Provides a plurality of donor DNA molecules, each donor DNA molecule comprising a first nucleic acid sequence encoding a first binding domain coupled to a first Fc domain; and a second nucleic acid sequence encoding a second binding domain coupled to a second Fc domain, wherein the first and second Fc domains are engineered to facilitate heterodimerization;

[0016] - Introduce the donor DNA into these cells; and

[0017] - Culture recombinant cells to generate clones, thereby providing a eukaryotic cell clone library containing DNA encoding a bispecific binding library.

[0018] In the context of this disclosure, the method according to this aspect may be referred to as "the method for generating a library as described herein" or "the method for generating a library," etc.

[0019] In some embodiments, the methods for generating the library as described herein cause the first and / or second binding domains to include a single antibody variable domain, preferably derived from a variable domain of a heavy chain antibody such as VHH or V. NAR And / or wherein the first and / or second binding domains comprise two antibody variable domains, preferably single-chain variable fragments (scFv). In some embodiments, the first binding domain comprises a single antibody variable domain, preferably derived from a variable domain of a heavy chain antibody such as VHH or V. NAR More preferably, VHH. In some embodiments, the second binding domain comprises two antibody variable domains, preferably single-chain variable fragments (scFv).

[0020] In some embodiments, the method for generating a library as described herein causes the first and second Fc domains to be engineered to facilitate heterodimerization, preferably wherein the first Fc domain contains a club-shaped mutation and the second Fc domain contains a mortar-shaped mutation, or wherein the first Fc domain contains a mortar-shaped mutation and the second Fc domain contains a club-shaped mutation.

[0021] The bispecific conjugates according to this disclosure are preferably multimeric conjugates comprising at least first and second subunits (i.e., individual polypeptide chains). More preferably, the bispecific conjugates according to this disclosure are dimer conjugates. Multimeric conjugates comprising dimer conjugates can be obtained by expression and assembly of individually encoded subunits. As mentioned above, multiple subunits can be encoded on the same molecule of donor DNA. However, it may also be desirable to integrate different subunits into separate loci, in which case the subunits can be provided on separate donor DNA molecules. These can be integrated in the same step of introducing donor DNA into the cell, or they can be integrated sequentially.

[0022] On one hand, methods for generating eukaryotic clonal libraries containing DNA encoding a diverse library of bispecific binding molecules (e.g., dimers) may include:

[0023] - A eukaryotic cell containing DNA encoding a first subunit of these bispecific binders, the DNA comprising a first nucleic acid sequence encoding a first binding domain coupled to a first Fc domain;

[0024] - Provide multiple donor DNA molecules encoding the second subunit of these bispecific binders, each donor DNA molecule containing a second nucleic acid sequence encoding a second binding domain coupled to the second Fc domain;

[0025] - Introducing the donor DNA into these cells to produce recombinant cells containing donor DNA integrated into the cell DNA; and

[0026] - Culture recombinant cells to generate clones containing DNA encoding these first and second subunits, thereby providing a eukaryotic clonal library containing DNA encoding a bispecific binding library.

[0027] In the above example, donor DNA encoding the second subunit is introduced into a cell that already contains DNA encoding the first subunit. An alternative approach is to integrate the first subunit in the first cycle of donor DNA introduction, followed by the introduction of the second subunit in the second cycle of donor DNA introduction.

[0028] Therefore, in one respect, methods for generating eukaryotic clonal libraries containing DNA encoding a diverse library of bispecific binding polymers (e.g., dimers) can include:

[0029] - Provide eukaryotic cells;

[0030] - Provide a plurality of first donor DNA molecules encoding first subunits of these bispecific binders, each donor DNA molecule containing a first nucleic acid sequence encoding a first binding domain coupled to a first Fc domain;

[0031] - Introduce the first donor DNA into these cells to produce a first set of recombinant cells containing the first donor DNA integrated into the cell DNA;

[0032] - Culturing the first set of recombinant cells to produce the first set of clones containing DNA encoding the first subunit;

[0033] - Provides a plurality of second donor DNA molecules encoding second subunits of these bispecific binders, each donor DNA molecule containing a second nucleic acid sequence encoding a second binding domain coupled to a second Fc domain;

[0034] - Introducing the second donor DNA into the cells of the first set of clones to produce a second set of recombinant cells containing the first and second donor DNA integrated into the cell DNA; and

[0035] - Culturing this second set of recombinant cells to produce a second set of clones containing DNA encoding these first and second subunits, thereby providing a eukaryotic clonal library containing DNA encoding a bispecific binding library.

[0036] Furthermore, on the other hand, methods for generating eukaryotic clonal libraries containing DNA encoding a diverse library of bispecific binding molecules for multimers (e.g., dimers) may include:

[0037] - Provide eukaryotic cells;

[0038] - Provide a plurality of first donor DNA molecules encoding first subunits of these bispecific binders, each donor DNA molecule containing a first nucleic acid sequence encoding a first binding domain coupled to a first Fc domain;

[0039] - Provides a plurality of second donor DNA molecules encoding second subunits of these bispecific binders, each donor DNA molecule containing a second nucleic acid sequence encoding a second binding domain coupled to a second Fc domain;

[0040] - Introducing the first and second donor DNAs into these cells to produce recombinant cells containing the first and second donor DNAs integrated into the cell DNA; and

[0041] - These recombinant cells are cultured to generate clones, thereby providing a eukaryotic clonal library containing DNA encoding a bispecific antibody library.

[0042] As used herein, “the methods described herein” and similar expressions refer to any of the above-described methods for generating a diverse library of eukaryotic clonal libraries encoding (multimers, including dimers) bispecific binders.

[0043] The methods described herein (e.g., methods for generating libraries) may involve providing a site-specific nuclease that cuts a recognition sequence in cellular DNA. More specifically, in the context of the methods described herein, in some embodiments, the step of introducing donor DNA into a cell includes providing a site-specific nuclease within the cell, wherein the nuclease cuts the recognition sequence in the cellular DNA to create an integration site at which the donor DNA integrates into the cellular DNA. Integration occurs through a DNA repair mechanism that is endogenous to the cell, resulting in a recombinant cell containing donor DNA integrated into the cellular DNA. In some embodiments, the flanking sides of the donor DNA molecule are homologous arms. This can improve integration efficiency. In some embodiments, the donor DNA molecule includes a first promoter operatively linked to a first nucleic acid sequence, and / or the donor DNA molecule includes a second promoter operatively linked to a second nucleic acid sequence. In some embodiments, the donor DNA molecule includes a bidirectional promoter operatively linked to both the first and second nucleic acid sequences. In some embodiments, the donor DNA molecule includes a bidirectional promoter operatively linked to both the first and second nucleic acid sequences. In some embodiments, the first and / or second nucleic acid sequences encode membrane anchors, such as transmembrane domains or membrane localization signals. In some embodiments, the eukaryotic cell is a genome size greater than 2 × 10⁻⁶. 7 Higher eukaryotic cells with 1 base pair, preferably mammalian, bird, insect or plant cells, more preferably mammalian cells.

[0044] Donor DNA is integrated into cellular DNA to generate recombinant cells, which can be cultured to produce clones. Thus, a single recombinant cell in which donor DNA has been integrated is replicated to produce a population of clonal cells—“clones”—each derived from an original recombinant cell. Therefore, this method produces a number of clones corresponding to the number of cells in which donor DNA has been successfully integrated. The collection of clones forms a library encoding a bispecific binding subunit (or, in the intermediate stage where bispecific binding subunits are integrated in individual rounds, a clone may encode a set of bispecific binding subunits). Therefore, the method described herein can provide a eukaryotic clonal library containing DNA encoding a bispecific binding library.

[0045] Therefore, in one aspect, eukaryotic cell clonal libraries containing DNA encoding a diverse library of bispecific binders are provided, wherein the library is obtained via the method described herein for generating the library. In the context of this application, such a library according to this aspect may be referred to as "the library described herein," etc.

[0046] The method described herein can generate eukaryotic clonal libraries expressing diverse libraries of bispecific binders, each cell containing recombinant DNA in which donor DNA encoding the bispecific binder or a subunit of the bispecific binder is integrated. The donor DNA may be integrated at a fixed locus in the cellular DNA, or optionally at multiple fixed loci. "Fixed" means that the loci are identical across cells. Therefore, the cells used to generate the library may contain nuclease recognition sequences at fixed loci representing universal landing sites in the cellular DNA where the donor DNA can be integrated. Site-specific nuclease recognition sequences may be present at one or more locations in the cellular DNA. Therefore, in one respect, eukaryotic cell clone in vitro libraries are provided, which are diverse libraries expressing at least 10^3, 10^4, 10^5, 10^6, 10^7, 10^8, or 10^9 different bispecific binders, each cell containing recombinant DNA, wherein donor DNA encoding the bispecific binder or a subunit of the bispecific binder is integrated into at least a first and / or a second locus in the cell DNA; optionally, wherein the locus is a fixed locus. As used herein, "library" and the like also refer to such eukaryotic cell clone in vitro libraries.

[0047] The libraries generated according to the methods described herein can be used in a variety of ways. The libraries can be cultured to express the bispecific binders, thereby generating a diverse library of bispecific binders. The libraries can be used to screen cells exhibiting a desired phenotype, wherein the phenotype is generated by the cells expressing the bispecific binders. Therefore, in one aspect, a method is provided for screening cells with a desired phenotype, wherein the phenotype is generated by the cells expressing the bispecific binders, the method comprising:

[0048] The library is provided via the methods for generating the library described herein, or as described herein;

[0049] These library cells were cultured to express these bispecific binders; and

[0050] The test determines whether the desired phenotype is exhibited.

[0051] This method may be referred to as "the method described herein for screening cells with the desired phenotype," etc. As used herein, "the method described herein" also refers to the aforementioned method for screening cells with the desired phenotype.

[0052] Phenotypic screening is possible, in which library cells are cultured to express a bispecific binder, and then the desired phenotype is detected in clones of the library. Cell readings can be based on alterations in cell behavior, such as changes in the expression of endogenous or exogenous reporter genes, differentiation state, proliferation, survival, cell size, metabolism, or interactions with other cells. When the desired phenotype is detected, clones expressing the desired phenotype can be recovered. Optionally, DNA encoding the bispecific binder is then isolated from the recovered clones, thus providing DNA encoding the bispecific binder that produces the desired phenotype when expressed in cells. Optionally, the DNA encoding the bispecific binder that produces the desired phenotype can be sequenced.

[0053] The library can also be used for screening to identify bispecific binders that target a target (optionally two target targets). Therefore, in one aspect, a method is provided for screening to identify bispecific binders targeting a target (optionally two target targets), the method comprising:

[0054] The library is provided via the methods for generating the library described herein, or as described herein;

[0055] Culture the library cells to express the bispecific binder;

[0056] Expose these bispecific conjugates to a target, optionally to two targets, allowing one or more target conjugates (if present) to recognize the target, optionally recognizing two targets; and

[0057] The test determines whether the target conjugate identifies the target, and two targets can be selected at random.

[0058] Methods in this regard may be referred to as "methods for screening to identify conjugates against a target as described herein" or "methods for screening to identify conjugates," etc. As used herein, "methods described herein" also refers to the aforementioned methods for screening to identify conjugates against a target.

[0059] In these methods, a library is cultured to express a bispecific binder, and the bispecific binder is exposed to one or more targets to allow one or more target binders (if present) to recognize one or more targets, and to detect whether the target binder recognizes one or more targets. When the desired one or more targets are bound, cloned cells containing DNA encoding the target binder can be recovered. Optionally, the DNA encoding the bispecific binder is then isolated from the recovered clones, thus providing DNA encoding a bispecific binder that binds to one or more target targets. Optionally, the DNA encoding the binder that recognizes the targets can be sequenced.

[0060] In this method, bispecific binders can be displayed on cell surfaces, and clones exhibiting bispecific binders with the desired properties can be isolated from a library. Therefore, cells incorporating genes encoding bispecific binders with the desired function or binding characteristics can be identified within the library. These genes can be recovered and used to generate bispecific binders, or further engineered to create derivative libraries of bispecific binders to produce bispecific binders with improved properties.

[0061] In one aspect, bispecific binders identified from libraries as described herein are provided, for example, bispecific binders identified using methods described herein for screening to identify bispecific binders against a target. Preferred bispecific binders are described elsewhere herein.

[0062] The following further describes various features of the aspects and embodiments disclosed herein. It should be noted that the headings used in this specification are for navigational purposes only and should not be construed as definitive, and the features described in different sections may relate to all aspects and embodiments described herein and can therefore be appropriately combined. Detailed Implementation

[0063] Bispecific conjugate

[0064] The term "conjugate" as used herein refers to a conjugating molecule that represents a specific binding chaperone for another molecule. Typical examples of specific binding chaperones are antibody-antigen and receptor-ligand. Preferably, the (bispecific) conjugates described herein are (bispecific) antibodies.

[0065] As used herein, "bispecific conjugate" refers to a molecule containing binding domains targeting two different antigens or two different epitopes on the same antigen. A bispecific conjugate consists of two "monospecific" subunits that can bind together in a bispecific form (heterodimerization) to form the "bispecific conjugate." Therefore, in this disclosure, references to "bispecific conjugate" may be replaced with references to "subunits of a bispecific conjugate." Subunits are often engineered to facilitate the formation of a bispecific / heterodimer form, for example, by using Fc domains engineered to facilitate heterodimerization. Cellular clone libraries encoding such bispecific conjugate libraries allow for the direct identification of the optimal bispecific conjugate in the final (therapeutic) form.

[0066] A library-encoded dual-specific binding library typically shares a common structure (constant domain, such as the Fc domain) and has one or more diverse regions (variable domains). Therefore, the library can select members of molecules belonging to a desired structural class. For example, a dual-specific binding compound can be a polypeptide that shares a common structure (constant domain, such as the Fc domain) and has one or more amino acid sequence diversity regions (variable domains).

[0067] This can be illustrated by considering a bispecific antibody molecule library. These may be antibody molecules of a common structural class, such as scFv-Fc or VHH-Fc, with one or more regions of their sequence differing. The complementarity-determining region (CDR) of an antibody molecule typically exhibits sequence variability; the CDR is the region primarily involved in antigen recognition. The bispecific binding library described herein could be a bispecific antibody molecule library that differs in one or more CDRs, for example, in all six CDRs, or where sequence diversity may exist in one or more specific CDRs, such as heavy chain CDR3 and / or light chain CDR3.

[0068] For multimeric conjugates, such as dimer bispecific conjugates, the use of a bispecific conjugate library is also covered, in which one or more subunits do not have sequence variability. In other words, for example in the case of dimer bispecific conjugates, one subunit can remain constant while the other subunit can have one or more amino acid sequence diversity regions as described above.

[0069] This also covers bispecific binding libraries that use different linkers. The term "linker" is described in more detail elsewhere in this document and includes, for example, linkers that connect the VH and VL domains of scFv and link the binding domain to the Fc domain. Thus, for example, the bispecific binding libraries described herein can be libraries of bispecific antibody molecules having the same CDR or the same variable region but differing in linker sequences. Using bispecific binding libraries that differ in one or more linkers allows for optimization of the molecule for linker length and / or sequence in its final desired (therapeutic) form.

[0070] It also covers the use of bispecific conjugate libraries containing humanized and / or low-immunogenic variants of bispecific conjugates. Libraries of humanized / low-immunogenic variants using such bispecific conjugates allow for optimization of humanized / low-immunogenic molecules in their final desired (therapeutic) form. Techniques for antibody humanization and immunogenicity reduction are known to those skilled in the art, as described below: e.g., Safdari et al., Antibody humanization methods-areview and update. [Antibody Humanization Methods - Review and Update] Biotechnol Genet Eng Rev. [Biotechnology and Genetic Engineering Review] 2013; 29: 175-86; and Ministro et al., Therapeutic Antibody Engineering and Selection Strategies. [Therapeutic Antibody Engineering and Selection Strategies] Adv BiochemEng Biotechnol. [Advances in Biochemical Engineering / Biotechnology] 2020; 171: 55-86, both of which are incorporated herein by reference.

[0071] For multimeric conjugates, such as dimeric bispecific conjugates, the donor DNA encoding the bispecific conjugate can be provided as one or more DNA molecules. For example, in the case where a single bispecific antibody subunit is to be expressed separately, these can be encoded on a separate molecule of donor DNA. The donor DNA is integrated into the cellular DNA at multiple integration sites, for example, a first subunit at a locus and a second subunit at a second locus. Methods for introducing donor DNA encoding individual bispecific conjugate subunits are described in more detail elsewhere herein. Alternatively, and preferably, both subunits of the multimeric (e.g., dimeric) bispecific conjugate can be encoded on the same molecule of donor DNA integrated into the cellular DNA, optionally at a fixed locus.

[0072] Bispecific conjugates can be bispecific antibody molecules or bispecific non-antibody proteins containing two or more antigen-binding sites. Antigen-binding sites can be provided by arranging peptide loops on a non-antibody protein scaffold (such as fibronectin or cytochrome B), or by randomizing or mutating the amino acid residues of the loops within the protein scaffold to confer binding to the desired target (Haan & Maggos. BioCentury 2004; 12(5): A1-A6; Koide et al. Journal of Molecular Biology 1998; 284: 1141-1151; Nygren et al. Current Opinion in Structural Biology 1997; 7: 463-469). Protein scaffolds for antibody mimics are disclosed in WO 0034784, which describes proteins (antibody mimics) comprising a fibronectin type III domain having at least one random loop. Any domain member of the immunoglobulin gene superfamily can provide a suitable scaffold for transplanting one or more peptide loops (e.g., a set of antibody VH-CDR loops). The scaffold can be human or non-human protein.

[0073] In addition to the antibody sequence and / or antigen-binding site, bispecific conjugates may contain other amino acids, such as forming peptides or polypeptides, folded domains, or conferring additional functional characteristics to the molecule besides the ability to bind antigens. Bispecific conjugates may carry a detectable label or may be conjugated to a toxin or targeting moiety or enzyme (e.g., via a peptide bond or linker). For example, a bispecific conjugate may include a catalytic site (e.g., in an enzyme domain) and an antigen-binding site, wherein the antigen-binding site binds to the antigen, thereby targeting the catalytic site to the antigen. The catalytic site may inhibit the biological function of the antigen, for example, by cleavage.

[0074] Bispecific antibodies are preferred bispecific conjugates. More than 100 different forms of bispecific antibodies are known (Brinkmann & Kontermann. MAbs [Monoclonal Antibodies] 2017; 9(2): 182-212). As used herein, bispecific antibodies have defined specificity and are artificial or recombinant molecules not found in nature. Preferably, the bispecific antibodies described herein are dimer bispecific antibodies, i.e., composed of two separate polypeptide chains or "subunits".

[0075] As shown elsewhere, the preferred bispecific antibody described herein comprises a first binding domain coupled to a first Fc domain and a second binding domain coupled to a second Fc domain.

[0076] The binding domain can be coupled directly or indirectly to the Fc domain, i.e., by using a “linker” or “linker peptide.” In particular, in the context of the VHH binding domain (described in detail elsewhere herein), the binding domain can be coupled to the Fc domain via a linker. Those skilled in the art can design appropriate linker lengths and linker sequences, for example, depending on one or more targets bound by one or more binding domains. In a preferred embodiment, the linker that couples the binding domain to the Fc domain is a flexible linker. Those skilled in the art will understand that flexible linkers are typically composed of small, nonpolar (e.g., glycine) or polar (e.g., serine and threonine) amino acids, thereby allowing them to provide flexibility and mobility for the linked functional domain (as reviewed in Chen et al., Adv Drug Deliv Rev, 2013; 65(10): 1357-1369; Chichili et al., Protein Sci., 2013; 22(2): 153-67, both of which are incorporated herein by reference). For example, suitable flexible joints include (GGGGS)n (SEQ ID NO: 25), where “n” is the number of repetitions, which may optionally be 2-7, preferably 3-6.

[0077] A bispecific antibody containing one binding site for each antigen can be represented as a bivalent antibody. For example, a preferred bispecific antibody of this disclosure containing a first binding domain coupled to a first Fc domain and a second binding domain coupled to a second Fc domain can be represented as a bivalent bispecific antibody. Such a bivalent bispecific antibody can be described as having a “1+1” stoichiometry.

[0078] However, this disclosure also covers multivalent bispecific antibodies. For example, adding an additional binding domain to each of the polypeptide chains of the bivalent bispecific antibody described herein produces a tetravalent molecule with a “2+2” stoichiometry. Other forms allow for the production of “1+2” or “1+3” molecules, each having one binding site against one antigen and two or three binding sites against another antigen.

[0079] Therefore, in some embodiments, the bispecific antibody described herein comprises:

[0080] - One, two, three, or four associated domains coupled to the first Fc domain; and

[0081] - One, two, three, or four associated structural domains coupled to the second Fc structural domain.

[0082] As used herein, bispecific antibodies can have symmetrical or asymmetrical architectures, with asymmetrical architectures being preferred. A bispecific antibody with a symmetrical architecture can be understood as having a binding domain coupled to a first Fc domain and a binding domain coupled to a second Fc domain of the same type, while a bispecific antibody with an asymmetrical architecture can be understood as having a binding domain coupled to a first Fc domain and a binding domain coupled to a second Fc domain of different types. In the context of the preferred bispecific antibody disclosed herein comprising a first binding domain coupled to a first Fc domain and a second binding domain coupled to a second Fc domain, a bispecific antibody with a symmetrical architecture can be understood as having first and second binding domains of the same type, while a bispecific antibody with an asymmetrical architecture can be understood as having first and second binding domains of different types. The appropriate type of binding domain in the context of this disclosure is described below.

[0083] Combined structural domain

[0084] A binding domain can be any antibody, antibody fragment, or antibody domain capable of binding to an antigen. Therefore, as used herein, a binding domain can alternatively be referred to as an antigen-binding domain.

[0085] In traditional antibodies, the light and heavy chains are divided into "constant" and "variable" regions based on their structural and functional homology. The variable domains of the light chain (VL) and heavy chain (VH) determine antigen recognition and specificity. Conversely, the constant domains (CL) of the light chain and the constant domains (CH1, CH2, or CH3) of the heavy chain confer other biological properties, such as secretion, transplacental migration, Fc receptor binding, and complement binding. Conventionally, the constant domains are numbered increasing with their distance from the antibody's antigen-binding site or N-terminus. The N-terminal portion is the variable region, and the C-terminal portion is the constant region; the CH3 and CL domains actually contain the carboxyl terms of the heavy and light chains, respectively.

[0086] In addition to conventional antibodies composed of two heavy chains and two light chains, certain vertebrates, particularly camelids including dromedary camels, camels, llamas, and alpacas, and some cartilaginous fish such as sharks, produce so-called heavy chain antibodies, which are antibodies composed of only two heavy chains and lack the two light chains present in conventional antibodies. In camelids, these are designated as VHHs (variable domains) of heavy chain antibodies, and in some cartilaginous fish such as sharks, these are designated as VNARs (variable neoantigen receptors).

[0087] The preferred binding domain in this disclosure is a single-chain binding domain, that is, the binding domain is monolithic.

[0088] In some embodiments, the binding domain includes at least one antibody variable domain. In some embodiments, the first and / or second binding domain includes at least one antibody variable domain.

[0089] In some embodiments, the binding domain includes a single antibody variable domain. In some embodiments, the first and / or second binding domains comprise a single antibody variable domain. A binding domain comprising a single antibody variable domain may also be represented as a single-domain antibody (sdAb). An sdAb is a small monomeric antigen-binding fragment of an antibody, i.e., a variable region of the antibody heavy or light chain.

[0090] Heavy chain antibodies (HCAbs) discovered in camelids consist of two heavy chains, each divided into three domains: CH3-CH2-VHH. The variable domain of the HCAb corresponding to the complementary site for antigen recognition is called VHH. This variable domain can be expressed independently and still recognize the antigen. The amino acid sequence and structure of VHH can be considered to consist of four frame regions, or "FRs," but is not limited thereto. These frame regions or FRs are referred to in the art and hereinafter as "frame region 1" or "FR1"; "frame region 2" or "FR2"; "frame region 3" or "FR3"; and "frame region 4" or "FR4," respectively. These frame regions are interrupted by three complementarity-determining regions, or "CDRs," which are referred to in the art as "complementarity-determining region 1" or "CDR1"; "complementarity-determining region 2" or "CDR2"; and "complementarity-determining region 3" or "CDR3," respectively.

[0091] Heavy chain antibodies (HCAbs) found in sharks consist of two heavy chains comprising five constant domains (CNAR1, CNAR2, CNAR3, CNAR4, and CNAR5). VNARs are the variable domains of these antibodies. For VHHs, VNARs possess complete antigen recognition properties. The main difference between their variable domains is the absence of complementarity-determining region 2 (CDR2), resulting in only two CDRs.

[0092] In some embodiments, the binding domain comprising a single antibody variable domain may be (i) a variable domain of the heavy chain of a heavy chain antibody that naturally lacks a light chain, including but not limited to the variable domain of the heavy chain of a camel antibody (VHH) or the variable domain of the heavy chain of a shark antibody (VNAR), or (ii) a variable domain of the heavy chain of a conventional tetrachain antibody, including but not limited to the camelified variable domain of the heavy chain of a conventional tetrachain.

[0093] In a preferred embodiment, the binding domain containing a single antibody variable domain is VHH or VNAR, preferably VHH.

[0094] In some embodiments, the binding domain comprises two antibody variable domains. In some embodiments, the first and / or second binding domain comprises two antibody variable domains. The binding domain comprising two single antibody variable domains is preferably a single-chain variable fragment (scFv).

[0095] The scFv molecule consists of a VH domain and a VL domain linked by a linker or linker peptide. In a preferred embodiment, the linker coupling the VH and VL domains is a flexible linker. Suitable flexible linkers are described elsewhere herein. In the scFv molecule, the VH and VL domains form a VH-VL pair, wherein the complementarity-determining regions of the VH and VL bind together to form an antigen-binding site.

[0096] In some embodiments, the 1, 2, 3, or 4 binding domains coupled to the first Fc domain comprise a single antibody variable domain as described herein. In some embodiments, the 1, 2, 3, or 4 binding domains coupled to the second Fc domain comprise two antibody variable domains as described herein. In some embodiments, the 1, 2, 3, or 4 binding domains coupled to the first Fc domain comprise a single antibody variable domain as described herein, and the 1, 2, 3, or 4 binding domains coupled to the second Fc domain comprise two antibody variable domains as described herein, or vice versa.

[0097] In some embodiments, the first binding domain comprises a single antibody variable domain as described herein.

[0098] In some embodiments, the second binding domain comprises the two antibody variable domains described herein.

[0099] As explained above, it is preferable that the first and second binding domains are of different types. Therefore, in a preferred embodiment, the first binding domain comprises a single antibody variable domain (e.g., VHH) as described herein, and the second binding domain comprises two antibody variable domains (e.g., scFv) as described herein, or vice versa.

[0100] Constant structure domain

[0101] As indicated elsewhere in this document, the bispecific antibodies described herein may comprise:

[0102] - One, two, three, or four associated domains coupled to the first Fc domain; and

[0103] - One, two, three, or four associated structural domains coupled to the second Fc structural domain.

[0104] For example, the preferred bispecific antibodies described herein comprise a first binding domain coupled to a first Fc domain and a second binding domain coupled to a second Fc domain. The terms Fc region, Fc domain, etc., are used interchangeably herein and refer to the CH2 and CH3 domains that govern the IgG antibody molecule. The CH2 and CH3 domains are responsible for interacting with effector cells and complement components within the immune system. Therefore, bispecific antibodies containing Fc domains have the advantage that they can exhibit Fc-mediated effector functions such as antibody-dependent cell-mediated cytotoxicity (ADCC), antibody-dependent phagocytosis (ADCP), complement fixation, and FcRn-mediated recycling.

[0105] In a preferred embodiment, the Fc domain is engineered to facilitate heterodimerization. Typically, this means that the first Fc domain and the second Fc domain are different (i.e., they have different amino acid sequences). Therefore, in some embodiments, the first and second Fc domains described herein are different from each other. In some embodiments, the first and second Fc domains described herein are different from each other and are engineered to facilitate heterodimerization.

[0106] Various strategies for promoting heterodimerization are known to those skilled in the art, based on spatial or electrostatic directing effects or combinations thereof, and the formation of defined interchain disulfides to produce complementary interfaces that favor heterodimerization rather than homodimerization. Several methods have been described, and these methods may be appropriate in the context of this disclosure, see, for example, Table 1 in Brinkmann & Kontermann. MAbs [Monoclonal Antibodies] 2017; 9(2): 182-212:

[0107]

[0108] In some embodiments, the CH3 domains of the first Fc domain and the CH3 domains of the second Fc domain have corresponding mutations that promote dimerization. For example, such mutations may be selected from the table below.

[0109]

[0110] The mortar and pestle technique is based on generating a pestle at the CH3 domain interface of the first chain by replacing one or more smaller amino acid side chains with a larger amino acid side chain (e.g., T366Y, T366W, or S354C+T366W); and generating a mortar at the CH3 interface of the second chain by replacing one or more larger amino acid side chains with smaller amino acid side chains (e.g., Y407T, T366S+L368A+Y407V, or Y349C+T366S+L368A+Y407V). Because pestle-peggle and mortar-mortar interactions are spatially hindered or energy-inefficient, they favor the formation of heterodimers.

[0111] In a preferred embodiment, the first and second Fc domains, particularly the CH3 regions of the first and second Fc domains, contain corresponding club-and-mortar mutations. Therefore, the first Fc domain may contain a club-and-mortar mutation and the second Fc domain may contain a mortar mutation, or the first Fc domain may contain a mortar mutation and the second Fc domain may contain a club-and-mortar mutation.

[0112] The pestle mutation can be selected from the following groups: T366Y, T366W, and S354C+T366W. The mortar mutation can be selected from the following groups: Y407T, T366S+L368A+Y407V, and Y349C+T366S+L368A+Y407V. The corresponding pestle and mortar mutations can be selected from the following groups:

[0113] -T366Y (pestle) and Y407T (mortar);

[0114] -T366W (mortar) and T366S+L368A+Y407V (mortar); and

[0115] -S354C+T366W (pestle) and Y349C+T366S+L368A+Y407V (mortar).

[0116] In a preferred embodiment, the Fc domain containing the club-shaped mutation is coupled to a binding domain containing a single antibody variable domain (e.g., VHH) as described herein, and / or the Fc domain containing the mortise-shaped mutation is coupled to a binding domain containing two antibody variable domains (e.g., scFv) as described herein. In a more preferred embodiment, the Fc domain containing the club-shaped mutation is coupled to VHH, and the Fc domain containing the mortise-shaped mutation is coupled to scFv.

[0117] Bispecific conjugates can be derived from human antibody molecules. Therefore, where a constant domain is present, these are preferably human constant domains.

[0118] The libraries described herein can be used to select bispecific antibody molecules that bind to one or more target antigens. The selection process from the library is described in detail below. The selected bispecific antibody molecules, their encoding nucleic acids, and their sequences are one aspect of this disclosure. The antibody molecules and their encoding nucleic acids can be provided in isolated form or as cell clones containing said nucleic acids and expressing said molecules.

[0119] Bispecific antibody molecules can be selected from a library and then modified. For example, the in vivo half-life of antibody molecules can be increased by chemical modification (e.g., PEGylation) or by incorporation into liposomes.

[0120] Origin of nucleic acids encoding binding domains

[0121] The VH and VL genes can be amplified from B cells of immunized animals and cloned into suitable vectors for introduction into the eukaryotic libraries described herein. Phage display and ribosome display allow for the construction of very large libraries (>10^9 clones), thus enabling the isolation of human antibodies without immunization. The libraries generated according to this disclosure can also be used in conjunction with such methods. After multiple rounds of phage display selection, the selected conjugate population can be directionally integrated into eukaryotic cells via the nucleases described herein. This allows for the initial enrichment of the conjugate population using very large libraries based on other systems (e.g., phage display), while simultaneously allowing for efficient screening using eukaryotic cells as described above. Therefore, this disclosure combines the best features of phage display and eukaryotic display to obtain a high-throughput system with quantitative screening and sorting capabilities.

[0122] Using phage display and yeast display, it has been previously demonstrated that conjugates can be generated without immunization, provided a sufficiently large display library is used. For example, a variety of conjugates have been generated from non-immune antibody libraries of >10^7 clones (Marks et al. J Mol Biol [Journal of Molecular Biology] 1991; 222(3), 581-597). This, in turn, allows for the generation of target-specific conjugates that are difficult to achieve through conventional immunization pathways, such as antibodies against “self-antigens” or epitopes conserved across species. For example, human / mouse cross-reactive conjugates can be enriched by sequential selection of human and mouse versions of the same target. Since it is not possible to specifically immunize humans against most target antibodies, this method is particularly important in allowing the generation of human antibodies preferred for therapeutic purposes.

[0123] In previous mammalian display examples, the library size and quality were limited, and conjugates were generated only using libraries pre-enriched with conjugates (e.g., by immunization or by engineering of pre-existing conjugates). The ability to prepare large libraries in eukaryotic cells, particularly higher eukaryotes, creates the possibility of directly isolating conjugates from these libraries, starting with non-immune conjugates or conjugates not previously selected in another system. By generating libraries according to this disclosure, conjugates can be generated from non-immune sources. This, in turn, opens up possibilities for using conjugate genes from a variety of sources. Conjugate genes may be derived from PCR of natural sources (such as antibody genes). Conjugate genes may also be re-cloned from existing libraries (such as antibody phage display libraries) and cloned into suitable donor vectors for targeted integration of nucleases into target cells. Conjugates can be synthesized completely or partially. In addition, various types of conjugates are described elsewhere in this article, such as conjugate genes that can encode antibodies or alternative scaffolds (Skerra. Curr Opin Biotechnol [Contemporary Biotechnol] 2007; 18(4): 295-304; Gebauer & Skerra. Curr Opin Chem Biol [Contemporary Chemical Biology] 2009; 13(3): 245-255), peptides or engineered proteins or protein domains.

[0124] eukaryotic cells

[0125] Preferred eukaryotic cells and eukaryotic cell clones for all aspects of this disclosure (including the methods, uses, and libraries described herein) are defined as follows. It should be understood that all preferences relating to eukaryotic cells can also be applied to eukaryotic cell clones.

[0126] The eukaryotic cells are preferably higher eukaryotic cells, defined herein as cells with a genome larger than that of *Saccharomyces cerevisiae*, whose genome size is 12 × 10^6 base pairs (bp). For example, the genome size of a higher eukaryotic cell may be greater than 2 × 10^7 base pairs. In some embodiments, the eukaryotic cell is a higher eukaryotic cell with a genome size greater than 2 × 10^7 base pairs. This includes, for example, mammalian, avian, insect, or plant cells, preferably mammalian cells. Preferably, the eukaryotic cell is a mammalian cell, such as a mouse or human. More preferably, the eukaryotic cell is a human cell. The eukaryotic cell can be a primary cell or a cell line. Chinese hamster ovary (CHO) cells are commonly used for antibody and protein expression, but any alternative stable cell line such as HEK293 cells may be used in this disclosure. There are several methods that can effectively introduce exogenous DNA into primary cells, and these methods can be used (e.g., by electroporation, with an efficiency and viability of up to 95%, Parthiban et al. MAbs [Monoclonal Antibody] 2019; 11(5); 884-898; Dyson et al. MAbs [Monoclonal Antibody] 2020; 12(1): 1829335).

[0127] One particular benefit of nuclease-directed integration in methods for generating libraries involves integrating conjugate genes into higher eukaryotic cells with large genomes, where homologous recombination is less efficient without nuclease cleavage. Yeast (e.g., *Saccharomyces cerevisiae*) has a smaller genome than mammalian cells, and homologous recombination by homologous arms (without nuclease-directed cleavage) is an efficient method for introducing exogenous DNA compared to higher eukaryotes. Nuclease-directed integration has been used in yeast cells to address the problem of efficiently integrating multiple genes into a single yeast cell, for example, for the engineering of metabolic pathways (US2012 / 0277120), but this work does not include the introduction of conjugate libraries, nor does it address the problem of library construction in higher eukaryotes.

[0128] Preferred eukaryotic cells are T lymphocyte lineage cells (e.g., primary T cells or T cell lines) or B lymphocyte lineage cells. Particularly preferred are primary T cells or T cell-derived cell lines used for TCR libraries, including cell lines lacking TCR expression (Letourneur, F., Malissen, B. European Journal of Immunology 1989; 19(12): 2269-2274; Kanayama et al. Biochem Biophys Res Commun 2005; 327(1), 70-75; Lin et al. Nucleic Acids Research 2011; 39(3), e14). Preferred B lymphocyte lineage cells are B cells, pre-B cells, or progenitor B cells, and cell lines derived from any of these.

[0129] Constructing libraries in primary B cells or B cell lines is of particular value for constructing antibody libraries. These eukaryotic cells are preferred among methods for library generation. Breous-Nystrom et al. Methods [Methods] 2014; 65(1): 57-67 Library generation in mouse pre-B cell lines (1624-5). The chicken B cell-derived cell line DT40 (ATCC CRL-2111) shows particular promise for constructing conjugate libraries. DT40 is a small cell line with a relatively rapid cell division rate. Conjugate libraries can be targeted to specific loci using zinc finger nucleases (ZFNs), TALE nucleases, or CRISPR / Cas9 targeting endogenous sequences, or by targeting pre-integrated heterologous sites (which may include a wide range of nuclease recognition sites). DT40 cells express antibodies, so targeting antibody genes within antibody loci with or without disruption of the endogenous chicken antibody variable domains would be advantageous. DT40 cells have also been used as the basis for an in vitro system for generating chicken IgM, called the Autonomous Diversification Library System (ADLib system), which utilizes the intrinsic diversification that occurs at chicken antibody loci. Due to this endogenous diversification, it is possible to generate novel specificities. The nuclease-directed approach described herein can be used in combination with ADLib to combine various conjugate libraries from heterologous sources (e.g., human antibody variable region libraries or synthetically derived alternative scaffolds) with the potential for further diversification at chicken IgG loci. Similar benefits can be applied to human B cell lines, such as Nalm6 (Adachi et al., DNA and Cell Biology, 2006; 25(1), 19-24).

[0130] Other preferred B lineage cell lines used in methods for identifying loci and for generating libraries include cell lines such as mouse pre-B cell line 1624-5 and progenitor B cell line Ba / F3. Ba / F3 is IL-3 dependent (Palacios et al. Cell 1985; 41(3), 727-734), and its use is discussed elsewhere in this document.

[0131] Finally, many cell lines are preferred, including those listed in the "Encyclopedia of Cancer Cell Lines" (Barretina et al., Nature, 2012; 483(7391): 603-607) or the "COSMIC Catalogue of Somatic Mutations in Cancer" (Forbes et al., Nucleic Acids Research, 2011; 39(database release): D945-50).

[0132] In the methods for generating libraries, and in the libraries described herein, eukaryotic cells are preferably of a single type, generated by introducing donor DNA into a population of cloned eukaryotic cells, for example, by introducing donor DNA into cells of a specific cell line. The main significant differences between different library clones will be due to the integration of the donor DNA.

[0133] Eukaryotic virus system

[0134] The advantages of the aspects and embodiments described herein, such as the method for generating eukaryotic cell clone libraries and the resulting libraries, can be applied to virus display systems based on eukaryotic expression systems, such as baculovirus display or retrovirus display (Russell et al. Nucleic Acids Res 1993; 21(5): 1081-1085; Boublik et al. Nature Biotech 1995; 13(10): 1079-1084; Mottershead et al. Biochemical and Biophysical Research Communications 2000; 275(&): 84-90; Oker-Blom et al. Briefings in functional genomics and proteomics 2003; 2(3): 244-253). In this method, each cell will encode a binding compound that can be incorporated into the viral particle. In the case of retroviral systems, the encoded mRNA is packaged, and the encoded conjugate is presented on the cell surface. In the case of baculovirus systems, the gene encoding the conjugate needs to be encapsulated within the baculovirus particle to maintain the association between the gene and the encoded protein. This can be achieved using a host cell carrying an episomal copy of the baculovirus genome. Alternatively, the integrated copy can be released by a specific nuclease (different from the nuclease used for drive site-specific integration). In the case of multimeric conjugate molecules, some chaperones can be encoded in the cellular DNA, and the genes of one or more chaperones are packaged within the virus.

[0135] Introduction of nucleic acids

[0136] The methods described herein include introducing nucleic acids into eukaryotic cells. In methods for generating libraries, donor DNA molecules are introduced. Unless otherwise specified, the introduction of nucleic acids refers to the introduction of DNA molecules into eukaryotic cells.

[0137] Many methods for introducing nucleic acids into eukaryotic cells have been described, including transfection, infection, or electroporation. These methods are well known to those skilled in the art; see, for example, Green & Sambrook, Molecular Cloning: A Laboratory Manual, 4th ed., Cold Spring Harbor Press, New York (2012) (ISBN 978-1-936113-42-2); and Ausubel et al., Current Protocols in Molecular Biology, 3rd ed., John Wiley & Sons Inc. (2003).

[0138] Large-scale cell transfection can be performed using standard methods, including the polyethyleneimine-mediated transfection described herein. Additionally, there are methods for efficient electroporation of 10 cells within 5 minutes. 10 For cell-based methods, see, for example, Parthiban et al. MAbs [Monoclonal Antibody] 2019; 11(5); 884-898; Dyson et al. MAbs [Monoclonal Antibody] 2020; 12(1): 1829-335.

[0139] In methods for generating libraries, combinatorial libraries can be created where members of multimeric binding pairs (e.g., the first and second subunits described herein) or even different portions of the same binding molecule are introduced onto different plasmids. The introduction of individual donor DNA molecules encoding individual binding molecules or binding subunits can be simultaneous or sequential. For example, the first subunit of a multimeric bispecific binding compound can be introduced via transfection or infection, with cell culture and selection as necessary. Other components, such as the second subunit of the multimeric bispecific binding compound, can then be introduced in subsequent infection or transfection steps. One or two steps may involve the directed integration of a nuclease into a specific genomic locus.

[0140] Nucleic acid integration

[0141] The methods for generating libraries described herein may involve integrating nucleic acids into the genome of a eukaryotic cell. In this context, the terms genome and cellular DNA are used interchangeably. Unless otherwise explicitly stated, integration refers to the integration of a DNA molecule into the genome of a eukaryotic cell. Integrating nucleic acids into the genome (i.e., cellular DNA) forms recombinant DNA with a continuous DNA sequence, wherein the nucleic acid is inserted at the integration site. In this disclosure, integration is mediated by a natural DNA repair mechanism that is endogenous to the cell.

[0142] Nucleic acid integration can be random or specific.

[0143] In random integration of nucleic acids, the integration site is not defined by a specific sequence.

[0144] In the specific integration of nucleic acids, the integration site is defined by a specific sequence. In the context of specific integration, nucleic acids can be referred to as donor DNA, donor DNA molecules, or donor DNA sequences, etc.

[0145] Specific integration can be allowed to occur by introducing nucleic acids into cells, allowing site-specific nucleases to create integration sites, and allowing donor DNA to integrate. In this context, specific integration can also be referred to as nuclease-directed integration. Cells can be held in the culture for a sufficient time to allow DNA integration. This typically results in a mixed population of cells, including (i) recombinant cells in which donor DNA has integrated at an integration site generated by a site-specific nuclease, and optionally (ii) cells in which donor DNA has integrated at sites other than the desired integration site, and / or optionally (iii) cells in which donor DNA has not integrated. Thus, the desired recombinant cells and resulting clones can be provided in a mixed population further comprising other eukaryotic cells. The selection methods described elsewhere herein can be used to select desired cells and clones, or to enrich the desired cells and clones in said mixed populations.

[0146] As mentioned above, integration is mediated by natural DNA repair mechanisms that are endogenous to the cell. Endogenous DNA repair mechanisms in eukaryotic cells include homologous recombination, non-homologous end joining (NHEJ), and microhomologous end joining (MMEJ). The integration efficiency of such processes can be improved by introducing double-strand breaks (DSBs) into the cellular DNA, and it has been reported that the efficiency has been increased by 40,000 times using rare cutting endonucleases (a wide range of nucleases) such as I-SceI (Porteus & Baltimore. Science 2003; 300(5620): 763; Rouet et al. Molecular and Cellular Biology 1994; 14(12): 8096-8106; Jasin. Trends ingenetics 1996; 12(6): 224-228).

[0147] Unlike the site-specific recombination involved in systems such as the Flp-In system

[16] , the integration in this disclosure does not require a recognition site for an exogenous recombinase or engineered recombinase. Therefore, the methods for generating libraries preferably do not include the step of recombinase-mediated DNA molecule integration. Furthermore, in the methods for identifying loci and for generating libraries, eukaryotic cells preferably lack the recombination sites of site-specific recombinases. The mechanisms and practicalities of specific integration of donor DNA into cellular DNA by recombinases and nucleases are very different, as discussed in Jasin 1996 (Jasin. Trends in Genetics 1996; 12(6): 224-228).

[0148] Conversely, specific integration, including the use of site-specific nucleases, involves nucleases that create breaks or gaps in cellular DNA, which are then exposed to and repaired by endogenous cellular repair mechanisms such as homologous recombination or NHEJ. Recombinase-based methods require pre-integration at their recognition sites, thus necessitating the engineering of "hotspot" integration sites into cellular DNA as a preliminary step. With nuclease-directed integration, nucleases can be engineered or, in the case of CRISPR:Cas9, guided by guide RNA, to recognize endogenous recognition sequences—nucleic acid sequences naturally present in cellular DNA. Finally, at a practical level, nuclease-directed methods are more efficient at the level required for the specific integration of transgenes to prepare large conjugate libraries.

[0149] In the methods described herein, such as those used to generate libraries, the DNA repair mechanism for integrating donor DNA can be predetermined or biased to some extent by the design of the donor DNA and / or the selection of site-specific nucleases.

[0150] Homologous recombination is a natural mechanism by which cells use homologous sequences (e.g., from another allele) as repair templates to repair double-strand breaks. Homologous recombination has been used in cell engineering to introduce insertions (including transgenes), deletions, and point mutations into the genome. Homologous recombination is facilitated by providing homologous arms on the donor DNA. Therefore, the donor DNA preferably contains homologous arms. Primitive methods for engineering higher eukaryotic cells typically use 5-10 kb homologous arms within the donor plasmid to improve the efficiency of targeted integration into the target site. Homologous recombination is particularly effective in eukaryotes, such as yeast, with a genome size of only 12.5 × 10^6 bp, compared to higher eukaryotes with larger genomes (e.g., mammalian cells with 3000 × 10^6 bp).

[0151] Homologous recombination can also be guided by gaps in cellular DNA (Fujioka et al. Nucleic Acids Res. 1993; 21(3): 407-412), which can also serve as a pathway for the directed integration of nucleases into cellular DNA. Therefore, the integration of donor DNA included in the methods described herein, such as those used to generate libraries, preferably includes the introduction of gaps in cellular DNA. Two different pathways have been shown to facilitate homologous recombination of gapped DNA. One is essentially similar to double-strand break repair using Rad51 / Brca2, while the other is inhibited by Rad51 / Brca2 and preferentially uses single-stranded DNA or gapped double-stranded donor DNA (Davus & Maizels. PLoS ONE. 2011; 6(9): e23981).

[0152] Non-homologous end joining (NHEJ) is an alternative mechanism for repairing double-strand breaks in the genome, in which DNA ends are directly rejoined without the need for a homologous template. Nuclease-directed genomic DNA cleavage can also enhance transgene integration through non-homologous mechanisms. NHEJ provides a simple method for integrating introns and exons into introns, or allows for the integration of promoter-gene cassettes into the genome. Using non-homologous methods allows for the use of donor vectors lacking homologous arms, thus simplifying donor DNA construction.

[0153] It has been pointed out that short regions of terminal homology are used to rejoin DNA ends, and it has been hypothesized that 4 bp of microhomology may be used to guide the repair of double-strand breaks, known as microhomology-mediated end joining (MMEJ) (Jasin. Trends in Genetics 1996; 12(6): 224-228).

[0154] Site-specific nucleases

[0155] In some embodiments, the aspects and embodiments described herein relate to the use of site-specific nucleases and their recognition sequences. For example, methods for generating libraries may involve providing site-specific nucleases that cleave recognition sequences in cellular DNA. Preferred site-specific nucleases are defined below. It should be understood that all preferences relating to site-specific nucleases can also be applied, with necessary modifications, to the corresponding recognition sites.

[0156] Site-specific nucleases cleave cellular DNA upon specifically binding to a recognition sequence, thereby creating an integration site for donor DNA. In this document, the terms site, target site, recognition site, and recognition sequence are used interchangeably. Nucleases may create double-strand breaks or single-strand breaks (gap). Nuclease-mediated DNA cleavage enhances site-specific integration of the bound gene through endogenous cellular DNA repair mechanisms.

[0157] In the methods described herein, such as those used to generate libraries, the eukaryotic cells used may contain endogenous sequences recognized by site-specific nucleases, or the recognition sequences may be engineered into the cellular DNA. Furthermore, the site-specific nucleases may be exogenous, i.e., not naturally present in the selected cell type.

[0158] In the methods described herein, such as those used to generate libraries, site-specific nucleases can be introduced before, after, or simultaneously with the introduction of donor DNA. In addition to the conjugate, the donor DNA can conveniently encode the nuclease, or encode the nuclease on a separate nucleic acid co-transfected with or otherwise introduced with the donor DNA. The clone of the library can optionally retain the nucleic acid encoding the site-specific nuclease, or such nucleic acid can be transiently transfected into cells only.

[0159] Any suitable site-specific nuclease may be used in this disclosure. It may be a naturally occurring enzyme or an engineered variant. There are many known nucleases that are particularly suitable, such as those that recognize or can be engineered to recognize sequences that are rarely found in cellular DNA.

[0160] Preferably, the site-specific nuclease recognizes only one or two distinct recognition sequences. This is advantageous because it should ensure that each cell integrates only one or two donor DNA molecules.

[0161] If the recognition sequence is relatively long, the rarity of the sequence recognized by the site-specific nuclease is more likely. Preferably, the recognition sequence is at least 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, or 30 nucleotides long. Preferably, the recognition sequence is 10 to 40, 39, 38, 37, 36, 35, 34, 33, 32, 31, 30, 29, 28, 27, 26, 25, 24, 23, 22, 21, or 20 nucleotides long, or 12 to 40, 39, 38, 37, 36, 35, 34, 33, 32, 31, 30, 29, 28, 27, 26, 25, 24, 23, 22, 21, or 20 nucleotides long.

[0162] Preferred site-specific nucleases are broad-spectrum nucleases, zinc finger nucleases (ZFNs), TALE nucleases, and nucleic acid-guided (e.g., RNA-guided) nucleases, such as the CRISPR / Cas system. Each of these produces double-strand breaks, although engineered forms that produce single-strand breaks are known.

[0163] Large-scale nucleases (also known as homing endonucleases) are nucleases that occur in all kingdoms of life and recognize relatively long sequences (12–40 bp). Given their long recognition sequences, they are absent or relatively rare in eukaryotic genomes. Large-scale nucleases are classified into five families based on their sequence / structure: the LAGLIDADG (SEQ ID NO: 2) family, the GIY-YIG family, the HNH family, the His-Cys box family, and the PD-(D / E)XK family. The best-studied family is the LAGLIDADG (SEQ ID NO: 2) family, which includes the well-characterized I-SceI large-scale nuclease from *Saccharomyces cerevisiae*. I-SceI recognizes and cleaves an 18 bp recognition sequence (5'TAGGGATAACAGGGTAAT, SEQ ID NO: 1), leaving a 4 bp 3' overhang. Another commonly used example is I-CreI, which originates from the chloroplasts of the single-celled green algae *Chlamydomonas reinhardtii* and recognizes a 22 bp sequence (Silva et al., *Current Gene Therapy*, 2011; 11(1): 11-27). Numerous engineered variants with altered recognition sequences have been produced (Epinat et al., *Topics in Current Genetics*, 2013; 23: 147-185). A wide range of nucleases represents the first instance of the use of site-specific nucleases in genome engineering (Rouet et al., *Molecular and Cellular Biology*, 1994; 14(12): 8096-8106; Jasin. *Trends in Genetics*, 1996; 12(6): 224-228). Similar to recombinase-based approaches, the use of I-SceI and other broad-spectrum nucleases requires the prior insertion of an appropriate recognition sequence to be targeted within the genome or the engineered broad-spectrum nuclease to recognize the endogenous recognition sequence (Silva et al., Current Gene Therapy, 2011; 11(1): 11-27). Using this approach, targeting efficiency in HEK293 cells was achieved in 10%–20% of cells using I-SceI (judged by homologous-directed “repair” of the integration of the defective GFP gene) (Szczepek et al., Nature Biotechnology, 2007; 25(7): 786-793).

[0164] A preferred class of broad-spectrum nucleases is the LAGLIDADG (SEQ ID NO: 2) endonuclease. These include I-SceI, I-ChuI, and I-CreI. I, CsmI, PI-SceI, PI-TIiI, PI-MtuI, I-CeuI, I-SceII, I-SceIll, HO, Pi-CivI, PI-CtrI, PI- AaeI, PI-BsuI, PI-DhaI, PI-DraI, PI-MavI, PI-MchI, PI-Mfu, PI-MflI, PI-MgaI, PI-MgoI, PI -MinI, PI-MkaI, PI-MleI, PI-MmaI, PI-MshI, PI-MsmI, PI-MthI, PI-Mtu, PI-MxeI, PI-NpuI, PI-PfuI, PI-RmaI, PI-SpbI, PI-SspI, PI-FacI, PI-MjaI, PI-PhoI, Pi-TagI, PI-ThyI, PI-Tko I, I-Msol and PI-TspI; preferably, I-SceI, I-CreI, I-ChuI, I-DmoI, I-CsmI, PI-SceI, PI-PfuI, PI-TliI, PI-MtuI and I-Ceul.

[0165] In recent years, numerous methods have been developed that allow the design of novel sequence-specific nucleases by fusing sequence-specific DNA-binding domains with non-specific nucleases, creating sequence-specific nucleases designed with custom DNA-binding domain orientation. Binding specificity can be guided by engineered binding domains, such as zinc finger domains. These are small, modular domains stabilized by zinc ions, which participate in molecular recognition and are essentially used to recognize DNA sequences. Arrays of zinc finger domains have been engineered for sequence-specific binding and have been linked to the non-specific DNA-cutting domain of the type II restriction enzyme Fok1 to create zinc finger nucleases (ZFNs). Such ZFNs are the preferred site-specific nucleases discussed in this paper. ZFNs can be used to generate double-strand breaks at specific sites within the genome. Fok1 is a specific dimer and requires two ZFNs to bind tightly for cleavage to occur. By creating two distinct Fok1 variants engineered to form heterodimers only with each other, the specificity of the engineered nuclease was enhanced and its toxicity reduced [Doyon et al., NatMethods, 2011; 8(1): 74-79]. Such specific heterodimer ZFNs have been shown to achieve homologous-directed integration in 5%–18% of target cells without drug selection (Moehle et al., PNAS, 2007; 104(9): 3055-3060; Perez-Pinera et al., Nucleic Acids Research, 2012; 40(8): 3741-3752; Umov et al., Nature, 2005; 435(7042): 646-651). In the absence of selection, incorporation of inserts up to 8 kb at frequencies >5% has been demonstrated.

[0166] The discovery of transcription activator-like effector (TALE) molecules in Xanthomonas bacteria has further simplified the ability to engineer DNA-binding domains with defined specificity. These TALE molecules consist of an array of 33–35 amino acid monomers, each of which recognizes a single base within the target sequence (Bogdanove et al., Science, 2011; 333(6051): 1843–1846). This modular 1:1 relationship makes it relatively easy to design engineered TALE molecules to bind to any target DNA target. By coupling these designed TALEs with Fok1, new sequence-specific TALE-nucleases can be created. TALE nucleases, also known as TALENs, are preferred site-specific nucleases for this application and have been designed for a large number of sites (i.e., recognition sequences) with a high success rate of efficient gene-modification activity (Reyon et al., Nature Biotechnology, 2012; 30(5): 460–465). Other variants and enhancements of TALE nuclease technology have been developed and can be used as site-specific nucleases in methods described herein, such as those used for library generation. These include “giant TALENs” in which the TALE nuclease-binding domain is fused with a large number of nucleases [Boissel et al., Nucleic Acids Research, 2013; 42(4): 2591-2601] and “compact TALENs” in which cleavage is achieved using a single TALE nuclease-recognizing domain (Beurdeley et al., Nature Communications, 2013; 4: 1762).

[0167] In recent years, another system for directing double-stranded or single-stranded breaks to specific sequences in the genome has been described. This system, called the “clustered regularly spaced short palindromic repeats (CRISPR) and CRISPR-associated (Cas)” system, is based on bacterial defense mechanisms (Sampson & Weiss. BioEssays: News and Reviews in Molecular, Cellular and Developmental Biology [BioEssays: News and Reviews in Molecular, Cellular and Developmental Biology] 2014; 36(1): 34-38). The CRISPR / Cas system is a preferred site-specific nuclease in methods for identifying loci or for generating libraries. The CRISPR / Cas system targets DNA for cleavage via a short complementary single-stranded RNA (CRISPR RNA or crRNA) linked to a short palindromic repeat. In the commonly used “Type II” system, the processing of the targeting RNA depends on the presence of trans-activating crRNA (tracrRNA), whose sequence is complementary to the palindromic repeat sequence. Hybridization of tracrRNA with the palindromic repeat sequence triggers processing. The processed RNA activates the Cas9 domain and directs its activity to complementary sequences within the DNA. This system has been simplified to guide Cas9 to cleave from a single RNA transcript and has been directed to many different sequences within the genome (Shalem et al., Science 2014; 343(6166): 84-87; Wang et al., Science 2014; 343(6166): 80-84). An advantage of this genome cleavage approach is that it is guided by short RNA sequences, which makes engineered cleavage specificity relatively simple. Therefore, there are many different methods to achieve site-specific cleavage of genomic DNA. As mentioned above, this enhances the integration rate of the donor plasmid through endogenous cellular DNA repair mechanisms.

[0168] In the methods for generating libraries described herein, using a wide range of nucleases, ZFN, TALE nucleases, or nucleic acid guidance systems (such as the CRISPR / Cas9 system) as site-specific nucleases will enable the targeting of endogenous loci within the genome.

[0169] Alternatively, in the methods described herein, such as those used to generate libraries, heterologous recognition sites (i.e., recognition sequences) of site-specific nucleases can be pre-introduced, including broad-spectrum nucleases, ZFN, and TALE nucleases. Directed targeting of nucleases can be used to drive the insertion of recognition sequences via homologous recombination or using vector DNA or even double-stranded oligonucleotides (Orlando et al., Nucleic Acids Research, 2010; 38(15): e152). As an alternative, non-specific targeting methods can be used to introduce site-specific nuclease recognition sequences by using transposon-directed integration (Cadinanos & Bradley. Nucleic Acids Research, 2007; 35(12): e87). Virus-based systems, such as lentiviruses, applied at low titers, can also be used to introduce recognition sequences.

[0170] Site-specific nucleases can be encoded by a single gene introduced on one plasmid, while the donor DNA is present on a second plasmid. Of course, combinations of two or more of these elements can be used with the same plasmid, which can improve targeting efficiency by reducing the number of plasmids introduced in methods for identifying loci or generating libraries. Furthermore, one or more nucleases may be pre-integrated, which can also be inducible to allow for temporal control of nuclease activity, as demonstrated with respect to transposases (Cadinanos & Bradley. Nucleic Acids Research, 2007; 35(12): e87). Finally, nucleases can be introduced as recombinant proteins or protein:RNA complexes (e.g., in the case of RNA-directed nucleases such as CRISPR:Cas9).

[0171] Identification Sequence

[0172] As described, the method for generating the library involves providing a site-specific nuclease that cuts the recognition sequence in cellular DNA.

[0173] In some embodiments, the identification sequence is located at the AAVS locus, as described, for example, in WO 2015 / 166272, which is incorporated herein by reference in its entirety.

[0174] In some embodiments, the identification sequence is in the ROSA26 locus, as described, for example, in Perez-Pinera et al., Nucleic Acids Research, 2012; 40(8), which is incorporated herein by reference in its entirety.

[0175] Other suitable recognition sequences are described in WO 2023 / 025834, which is incorporated herein by reference in its entirety. In some embodiments, the recognition sequence is in the neurolysin (NLN) gene, as described, for example, in WO 2023 / 025834, which is incorporated herein by reference in its entirety. The eukaryotic cells used may contain an endogenous sequence recognized by a site-specific nuclease, or the recognition sequence may be engineered into the cellular DNA, as previously described herein. The neurolysin gene (human sequence: Uniprot Q9BYT8, ENSEMBL gene id ENSG00000123213) encodes a member of the metallopeptidase M3 protein family that cleaves neurotensin at the Pro10-Tyr1 1 bond to form neurotensin (1-10) and neurotensin (11-13). An exemplary sequence of the neurolysin gene is represented by SEQ ID NO: 3. In some embodiments, the recognition sequence is in a nucleic acid molecule represented by a nucleotide sequence comprising, substantially comprising, or composed of: SEQ ID NO: 1, or represented by a nucleotide sequence having at least 60%, 61%, 62%, 63%, 64%, 65%, 66%, 67%, 68%, 69%, 70%, 71%, 72%, 73%, 74%, 75%, 76%, 77%, 78%, 79%, 80%, 81%, 82%, 83%, 84%, 85%, 86%, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99%, or 100% identity with SEQ ID NO: 3.

[0176] In some embodiments, the identification sequence is in the TRAF2 and NCK-interacting kinase (TNIK) gene (UniprotQ9UKE5, ENSEMBL gene id ENSG00000154310), as described, for example, in WO 2023 / 025834, which is incorporated herein by reference in its entirety. An exemplary sequence of the TNIK gene is represented by SEQ ID NO: 4. In some embodiments, the identification sequence is in the protein mono-ADP ribosyltransferase 11 (PARP11) gene (Uniprot Q9NR21, ENSEMBL gene id ENSG00000111224), as described, for example, in WO 2023 / 025834, which is incorporated herein by reference in its entirety. An exemplary sequence of the PARP11 gene is represented by SEQ ID NO: 5. In some embodiments, the identifying sequence is in the RAB40B gene (member of the RAS oncogene family, Uniprot Q12829, ENSEMBL gene id ENSG00000141542), as described, for example, in WO2023 / 025834, which is incorporated herein by reference in its entirety. An exemplary sequence of the RAB40B gene is represented by SEQ ID NO: 6. In some embodiments, the identifying sequence is in the abl interactor 2 (ABI2) gene (Uniprot Q9NYB9, ENSEMBL gene id ENSG00000138443), as described, for example, in WO 2023 / 025834, which is incorporated herein by reference in its entirety. An exemplary sequence of the ABI2 gene is represented by SEQ ID NO: 7. In some embodiments, the identifying sequence is in the ring finger protein 19B (RNF19B) gene (Uniprot Q6ZMZ0, ENSEMBL gene id ENSG00000116514), as described, for example, in WO 2023 / 025834, which is incorporated herein by reference in its entirety. An exemplary sequence of the RNF19B gene is represented by SEQ ID NO: 8. In some embodiments, the identifying sequence is in the cAMP-dependent protein kinase inhibitor α (PKIA) gene (Uniprot P61925, ENSEMBL gene id ENSG00000171033), as described, for example, in WO 2023 / 025834, which is incorporated herein by reference in its entirety. An exemplary sequence of the PKIA gene is represented by SEQ ID NO: 9. In some embodiments, the identification sequence is in the imine methyltransferase cyclization deaminase (FTCD) gene (Uniprot O95954, ENSEMBL gene idENSG00000160282), as described, for example, in WO 2023 / 025834, which is incorporated herein by reference in its entirety. An exemplary sequence of the FTCD gene is represented by SEQ ID NO: 10.

[0177] In some embodiments, the recognition sequence is in the NLN gene, TNIK gene, or RAB40B gene. In some embodiments, the recognition sequence is in a nucleic acid molecule represented by a nucleotide sequence comprising, substantially comprising, or composed of: SEQ ID NO: 3-10, or represented by a nucleotide sequence having at least 60%, 61%, 62%, 63%, 64%, 65%, 66%, 67%, 68%, 69%, 70%, 71%, 72%, 73%, 74%, 75%, 76%, 77%, 78%, 79%, 80%, 81%, 82%, 83%, 84%, 85%, 86%, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99%, or 100% identity with SEQ ID NO: 3-10.

[0178] In some embodiments, the identifying sequence is in an intron selected from the following genes: NLN, TNIK, PARP11, RAB40B, ABI2, RNF19B, PKIA, or FTCD genes, preferably NLN, TNIK, or RAB40B genes, as described, for example, in WO 2023 / 025834, which is incorporated herein by reference in its entirety. The term "intron" is used herein as is customary and commonly understood by those skilled in the art.

[0179] The recognition sequence in the NLN intron is preferably in intron 1 (introns 1-2), intron 2 (introns 2-3), or intron 6 (introns 6-7) of NLN-207. The recognition sequence in the TNIK intron is preferably in intron 2 (introns 2-3) of TNIK-04 (Ensembl ID ENST00000436636.7). The recognition sequence in the PARP11 intron is preferably in intron 1 (introns 1-2) of PARP11-205 (Ensembl ID ENST00000450737.2). The recognition sequence in the RAB40B intron is preferably in intron 1 (introns 1-2) of RAB40B-206 (Ensembl ID ENST00000571995.6). The preferred recognition sequence in the ABI2 intron is in intron 1 (introns 1-2) of ABI2-203 (Ensembl ID ENST00000261018.12). The preferred recognition sequence in the RNF19B intron is in intron 1 (introns 1-2) of RNF19B-201 (Ensembl ID ENST00000235150.5). The preferred recognition sequence in the PKIA intron is intron 1 (introns 1-2) of PKIA-202 (Ensembl ID ENST00000396418.7). The preferred recognition sequence in the FTCD intron is intron 3 (introns 3-4) of FTCDNL1-201 (Ensembl ID ENST00000416668.5).

[0180] In a preferred embodiment, the recognition sequence is located in an intron of the neurolysin gene. A typical transcript of the human neurolysin (NLN) gene is NLN-201 (Ensembl transcript ID: ENST00000380985.10), which contains 13 exons. Another transcript is NLN-207 (Ensembl transcript ID: ENST00000509935.2), which contains 7 exons. In some embodiments, the recognition sequence is located in intron 1 of the NLN-201 gene (NLN-201 introns 1-2; exemplary sequence: SEQ ID NO: 11). In some embodiments, the recognition sequence is located in intron 2 of the NLN-201 gene (NLN-201 introns 2-3; exemplary sequence: SEQ ID NO: 12). In some embodiments, the recognition sequence is in intron 3 (NLN-201 introns 3-4; exemplary sequence: SEQ ID NO: 13) of the neurolysin gene. In some embodiments, the recognition sequence is in intron 4 (NLN-201 introns 4-5; exemplary sequence: SEQ ID NO: 14) of the neurolysin gene. In some embodiments, the recognition sequence is in intron 5 (NLN-201 introns 5-6; exemplary sequence: SEQ ID NO: 15) of the neurolysin gene. In some embodiments, the recognition sequence is in intron 6 (NLN-201 introns 6-7; exemplary sequence: SEQ ID NO: 16) of the neurolysin gene. In some embodiments, the recognition sequence is in intron 7 (NLN-201 introns 7-8; exemplary sequence: SEQ ID NO: 17) of the neurolysin gene. In some embodiments, the recognition sequence is located in intron 8 of NLN-201 or intron 1 of NLN-207 (NLN-201 introns 8-9 or NLN-207 introns 1-2; exemplary sequence: SEQ ID NO: 18). In some embodiments, the recognition sequence is located in intron 9 of NLN-201 or intron 2 of NLN-207 (NLN-201 introns 9-10 or NLN-207 introns 2-3; exemplary sequence: SEQ ID NO: 19). In some embodiments, the recognition sequence is located in intron 10 of NLN-201 or intron 3 of NLN-207 (NLN-201 introns 10-11 or NLN-207 introns 3-4; exemplary sequence: SEQ ID NO: 20). In some embodiments, the recognition sequence is in NLN-201 intron 11 or NLN-207 intron 4 (NLN-201 introns 11-12 or NLN-207 introns 4-5; exemplary sequence: SEQ ID NO: 21) of the neurolysin gene.In some embodiments, the recognition sequence is in intron 12 (NLN-201 introns 12-13; exemplary sequence: SEQ ID NO: 22) of the NLN-201 neurolysin gene. In some embodiments, the recognition sequence is in intron 5 (NLN-207 introns 5-6; exemplary sequence: SEQ ID NO: 23) of the NLN-207 neurolysin gene. In some embodiments, the recognition sequence is in intron 6 (NLN-207 introns 6-7; exemplary sequence: SEQ ID NO: 24) of the NLN-207 neurolysin gene.

[0181] The preferred introns are introns 1, 2 and 6 of NLN-207 in the NLN gene. In some embodiments, the identification sequence is in a nucleic acid molecule represented by a nucleotide sequence comprising, substantially comprising, or composed of: SEQ ID NO: 18, 19, 24, or represented by a nucleotide sequence having at least 60%, 61%, 62%, 63%, 64%, 65%, 66%, 67%, 68%, 69%, 70%, 71%, 72%, 73%, 74%, 75%, 76%, 77%, 78%, 79%, 80%, 81%, 82%, 83%, 84%, 85%, 86%, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99%, or 100% identity with SEQ ID NO: 18, 19, 24.

[0182] In some embodiments, particularly when the recognition sequence is in an intron of a gene selected from the NLN, TNIK, PARP11, RAB40B, ABI2, RNF19B, PKIA, or FTCD genes as described above, the recognition sequence is in an open chromatin region of the intron.

[0183] In some embodiments, particularly when the recognition sequence is in an intron of a gene selected from the NLN, TNIK, PARP11, RAB40B, ABI2, RNF19B, PKIA, or FTCD genes as described above, the recognition sequence is in the enhancer region of the intron.

[0184] As used in this article, “open chromatin” or “euchromatin” or “loose chromatin” refers to structures that allow transcription, while “heterochromatin” or “tight” or “closed” chromatin is more compact and more tolerant to factors that require access to the DNA template.

[0185] Identify the distribution of sequences

[0186] In the method described herein, the recognition sequence of the site-specific nuclease can be present in genomic DNA or in episodic DNA that is stably inherited in the cell. Therefore, donor DNA can be integrated into the genomic or episodic loci within the cellular DNA.

[0187] In its simplest form, the donor DNA targets a single site in the eukaryotic genome. Identifying cells exhibiting specific binding activity or cellular phenotypes allows for the direct isolation of genes encoding the desired properties (e.g., from mRNA or genomic DNA via PCR). This is facilitated by the use of unique recognition sequences of site-specific nucleases that occur only once in the cellular DNA. Thus, cells used to create libraries can contain nuclease recognition sequences at a single fixed locus, i.e., at the same locus in all cells. Libraries produced from such cells will contain donor DNA integrated at a fixed locus, i.e., at the same locus in the cellular DNA of all clones in the library.

[0188] Optionally, the recognition sequence can occur multiple times in the cellular DNA, giving the cell more than one potential integration site for donor DNA. This would be typical of diploid or polyploid cells, where the recognition sequence is located at a corresponding location on a pair of chromosomes, namely the replication locus. Libraries produced from such cells may contain donor DNA integrated at the replication locus. For example, a library produced from diploid cells may have donor DNA integrated at the duplication locus, while a library produced from triploid cells may have donor DNA integrated at the triplet locus. Many suitable mammalian cells are diploid, and the mammalian cell clonal libraries described herein can contain donor DNA integrated at the duplication locus.

[0189] The sequence recognized by site-specific nucleases may be located at more than one independent locus in cellular DNA. Therefore, donor DNA can be integrated at multiple independent loci. Diploid or polyploid cell libraries can contain donor DNA integrated at multiple independent fixed loci and / or repetitive fixed loci.

[0190] In cells containing recognition sequences at multiple loci (whether repetitive or independent), each locus represents a potential integration site for a donor DNA molecule. Introducing donor DNA into a cell can result in integration at all nuclease recognition sequences present in the cell, or the donor DNA can integrate at some, but not all, of these potential sites. For example, when generating a library from diploid cells containing recognition sequences at first and second fixed loci (e.g., repetitive fixed loci), the resulting library can contain clones of donor DNA integrated at the first fixed locus, clones of donor DNA integrated at the second fixed locus, and clones of donor DNA integrated at both the first and second fixed loci.

[0191] Therefore, the library generation methods described herein may involve site-specific nuclease cleavage of multiple fixed loci in cells and integration of donor DNA into multiple fixed loci. As mentioned above, in the presence of multiple copies of the same recognition sequence (e.g., when targeting endogenous loci in diploid or polyploid cells), nucleic acid sequences encoding binders (or binder subunits) may be integrated, especially when using effective targeting mechanisms, where only one nucleic acid sequence encoding a binder (or binder subunit) is specific to the target. Once the nucleic acid sequence encoding the binder (or binder subunit) is isolated, this can be addressed during subsequent screening.

[0192] Donor DNA

[0193] The methods described herein for generating libraries may include integrating donor DNA. Preferred donor DNA molecules are described in this section.

[0194] Donor DNA is typically circular DNA and can be provided as a plasmid or vector. Linear DNA is another possibility. In addition to one or more donor DNA sequences integrated into the cellular DNA, the donor DNA molecule can also include regions that are not integrated into the cellular DNA. DNA is usually double-stranded, but single-stranded DNA can be used in some cases. The donor DNA contains one or more transgenes encoding binding agents; for example, it can contain promoters: gene cassettes.

[0195] In its simplest form, double-stranded circular plasmid DNA can be used to drive homologous recombination. This requires DNA regions flanking the transgene that are homologous to DNA sequences flanking cleavage sites in the genomic DNA. Linearized double-stranded plasmid DNA, PCR products, or synthetic genes can be used to drive homologous recombination and the NHEJ repair pathway. As an alternative to double-stranded DNA, single-stranded DNA can be used to drive homologous recombination (Fujioka et al., Nucleic Acids Res, 1993; 21(3): 407-412). A common method for generating single-stranded DNA is to include the single-stranded origin of replication from filamentous phages in the plasmid.

[0196] Single-stranded DNA viruses such as adeno-associated virus (AAV) have been used to drive efficient homologous recombination, with efficiency shown to be improved by several orders of magnitude (Khan et al., Nat Prot [Nature-Lab Handbook] 2011; 6(4): 482-501; Deyle & Russell. Current Opinion in Molecular Therapeutics [Current Opinion in Molecular Therapeutics] 2009; 11(4): 442-447). Systems such as AAV can be used in conjunction with nuclease-directed cleavage in both locus identification and library generation methods. The advantages of both systems can be applied to both locus identification and library generation methods. The packaging limit of AAV vectors is 4.7 kb, but digestion of target genomic DNA with nucleases will reduce this limitation, thus allowing the introduction of larger transgenic constructs.

[0197] The donor DNA molecule may encode a single binder or multiple binders. Optionally, each donor DNA molecule may encode multiple subunits of the binder. In some embodiments, the donor DNA encodes a subunit of a multimeric binder.

[0198] Selection of promoters in donor DNA and clones with integrated donor DNA

[0199] In the methods described herein, such as those used to generate libraries, the donor DNA contains one or more nucleic acid sequences encoding a binding domain coupled to an Fc domain. Transcription from the donor DNA to the binding compound or binding subunit is typically achieved by placing the sequence encoding the binding compound or binding subunit under the control of a promoter and optionally one or more enhancer elements for transcription. The promoter (and optionally other genetic control elements) may be contained within the donor DNA molecule itself. Alternatively, the sequence encoding the binding compound may lack a promoter on the donor DNA but may be operatively linked to a promoter on cellular DNA (e.g., an endogenous promoter or a pre-integrated exogenous promoter) as a result of its insertion into an integration site generated by a site-specific nuclease.

[0200] In some embodiments, the donor DNA molecule described herein includes a first promoter operatively linked to a first nucleic acid sequence and / or a second promoter operatively linked to a second nucleic acid sequence.

[0201] In some embodiments, the first and second promoters are identical. In a preferred embodiment, the first and second promoters are different. In some embodiments, each of the first and second promoters can be independently a constitutive or inductive promoter, preferably a constitutive promoter. Suitable constitutive promoters include CMV promoters, CAG promoters, and EF1α promoters, or variations thereof. Preferred promoters are CMV and EF1α promoters. More preferred promoters are CMV promoters. Therefore, each of the first and second promoters can be independently selected from the group consisting of CMV promoters, CAG promoters, and EF1α promoters, preferably from the group consisting of CMV promoters and EF1α promoters.

[0202] In a preferred embodiment, the first promoter is a CMV promoter and the second promoter is an EF1α promoter, or the first promoter is an EF1α promoter and the second promoter is a CMV promoter. More preferably, the first promoter is a CMV promoter and the second promoter is an EF1α promoter.

[0203] In some embodiments, the donor DNA molecule described herein includes a bidirectional promoter operatively linked to first and second nucleic acid sequences. A preferred bidirectional promoter is a bidirectional CMV promoter.

[0204] The donor DNA may further contain one or more additional coding sequences, such as genetic elements capable of selecting cells containing or expressing the donor DNA. Such elements can be called selection markers. Similar to the sequences encoding binders discussed above, these elements can bind to a promoter on the donor DNA, or, as a result of the donor DNA integrating into a fixed locus, can be placed under the control of a promoter. The latter arrangement provides a convenient method for specifically selecting cells that have integrated donor DNA at the desired site, since these cells should express the genetic element used for selection. For example, this could be a gene conferring resistance to negative selectors such as blast fungicides or puromycin. One or more selection steps can be applied to remove unwanted cells, such as those lacking donor DNA or those that have not integrated donor DNA at the correct location.

[0205] The expression of membrane-anchored conjugates can themselves be used as a form of selection marker. For example, for the library described herein, ten cells expressing the conjugate can be selected using secondary reagents that recognize Fc expressed on the surface using the methods described herein. Transient expression of the conjugate (and cell surface expression) will occur after initial transfection with donor DNA encoding the conjugate under the control of an exogenous promoter, and the transient expression must be allowed to attenuate (to achieve, for example, targeted integration of 1-2 antibody genes / cells).

[0206] As an alternative, constructs encoding membrane tethering elements or membrane anchors (e.g., the Fc domain of this example fused with the transmembrane domain of the PDGF receptor) can be pre-integrated before the introduction of the binding sequence.

[0207] Therefore, in some embodiments, the methods described herein enable the first and / or second nucleic acid sequences described herein to further encode membrane anchors, such as transmembrane domains or membrane localization signals. A preferred transmembrane domain is the PDGF receptor transmembrane domain. A preferred membrane localization signal is a GPI recognition sequence. Preferably, the membrane anchor described herein is fused to a binding unit (subunit), more preferably to an Fc domain.

[0208] In some embodiments, the first nucleic acid sequence, rather than the second nucleic acid sequence, further encodes the membrane anchor. In some embodiments, the second nucleic acid sequence, rather than the first nucleic acid sequence, further encodes the membrane anchor.

[0209] If the membrane tethering element lacks a promoter or is encoded within an exon that does not fit within the frame of the preceding exon, surface expression will be impaired. This defect can then be corrected by targeted integration of the incoming donor molecule (e.g., by targeting the promoter or “in-frame” exon to an intron upstream of the defective tethering element). If the frame “correcting exon” also encodes the conjugate, a fusion will occur between the conjugate and the membrane tethering element, resulting in surface expression of both. Therefore, proper targeted integration will result in in-frame expression of the membrane tethering element, either alone or as part of a fusion with the incoming conjugate. Furthermore, if the incoming conjugate library lacks membrane tethering elements and these elements are incorrectly integrated, they will not be selected. Therefore, the expression of the conjugate itself on the cell surface can be used to select cell populations for proper targeted integration.

[0210] Clone count and library diversity

[0211] The methods described herein for generating libraries are eukaryotic clonal libraries containing DNA encoding a diverse library of bispecific binders. In the context of this application, a library refers to a eukaryotic clonal library containing DNA encoding a diverse library of binders, obtainable by one of these methods, unless otherwise expressly stated. Preferred libraries and their characteristics are defined in this section.

[0212] 10 7 -10 10 Yeast display libraries have previously been constructed and demonstrated to produce binders in the absence of herd immunity or preselection (Chao et al., Nat Protoc [Nature-Laboratory Handbook] 2006; 1(2): 755-768; Benatulil et al., Protein Engineering, Design and Selection 2010; 23(4): 155-159; Feldhaus et al., Nat Biotechnol [Nature Biotechnology] 2003; 21: 163-170; Zhao et al., Journal of Immunological Methods 2011; 363(2): 221-232). Given the limitations on library size and variability when using cells from higher eukaryotes, many previously published mammalian display libraries utilize antibody genes from immune donors or even enriched antigen-specific B lymphocytes. Due to the efficiency of gene targeting described in this paper, large original libraries can be constructed in higher eukaryotes (e.g., mammalian cells) that match the libraries described for simpler eukaryotes (e.g., yeast).

[0213] After the donor DNA is integrated into the cellular DNA, the resulting recombinant cells are cultured to allow them to replicate, thereby generating cell clones from each initially generated recombinant cell. Thus, each clone is derived from a progenitor cell into which the donor DNA is integrated at an integration site created by a site-specific nuclease. The method described herein is associated with high efficiency and high fidelity of donor DNA integration, and the libraries described herein can contain at least 100, 10 3 10 4 10 5 10 6 10 7 10 8 10 9 Or 10 10 One clone.

[0214] Unbound by this theory, targeted integration using nucleases can target 10% or more of transfected mammalian cells. Growth and transformation >10 10 Individual cells are also feasible (e.g., from 2 × 10⁻⁶ cells). 6 (5 liters of cells per ml). Transfection of such large numbers of cells can be performed using standard methods, including polyethyleneimine-mediated transfection as described herein. Additionally, there are methods for efficient electroporation of 10 cells per ml within 5 minutes. 10Methods for creating individual cells, for example, http: / / www.maxcyte.com. Therefore, using the methods disclosed herein, it is possible to create more than 10 9 A cloned library.

[0215] When the donor DNA molecule population used to create a library contains multiple copies of the same sequence, two or more clones containing DNA encoding the same conjugate can be obtained. It is also possible for clones to contain donor DNA encoding more than one different conjugate, for example, if there is recognition sequence for more than one site-specific nuclease, as detailed elsewhere in this document. Therefore, the diversity of the library may differ from the number of clones obtained in terms of the number of different conjugates encoded or expressed.

[0216] Clones in a library preferably contain donor DNA encoding one or two members of the conjugate library, and / or preferably express only one or two members of the conjugate library. A limited number of different conjugates per cell is advantageous when it comes to identifying clones and / or DNA encoding specific conjugates identified when screening a library against a given target. This is simplest when cloning a single member of the conjugate library. However, it is also simple to identify the relevant encoding DNA of the desired conjugate if the clones selected from the library encode a small number of different conjugates, for example, if the clones could encode two members of the conjugate library. As discussed elsewhere in this document, clones encoding one or two conjugates are particularly convenient to generate by selecting recognition sequences of site-specific nucleases that occur once per chromosome copy in a diploid genome, because diploid cells contain repetitive fixed loci, one on each chromosome copy, and donor DNA can be integrated at one or two fixed loci. Therefore, clones of a library may each express only one or two members of the conjugate library.

[0217] The conjugates displayed on the cell surface of the library can be identical to other conjugates displayed on the same cell (having the same amino acid sequence). The library can consist of cell clones, each displaying a single member of the conjugate library, or clones displaying multiple members of the conjugate library per cell. Alternatively, the library can contain some clones displaying a single member of the conjugate library, and some clones displaying multiple members (e.g., two) of the conjugate library.

[0218] Therefore, the libraries described herein may include clones encoding more than one member of a binding library, wherein donor DNA is integrated at a repetitive fixed locus or multiple independent fixed loci.

[0219] As mentioned above, if the corresponding clone expresses only one conjugate, the corresponding coding DNA of the conjugate is most easily identified. Typically, the donor DNA molecule will encode a single conjugate. The conjugate can be multimeric, such that the donor DNA molecule includes multiple genes or open reading frames corresponding to multiple subunits of the multimeric conjugate.

[0220] The library described in this article can be encoded with at least 100, 10 3 10 4 10 5 Or 10 6 10 7 10 8 10 9 Or 10 10 Different conjugates. In the case of a multimer, diversity can be provided by one or more subunits of the conjugate. Multimer conjugates can combine one or more variable subunits with one or more constant subunits, where the constant subunits are the same (or have more limited diversity) across all clones of the library. In generating multimer conjugate libraries, combinatorial diversity is possible, where a first conjugate subunit library can be paired with any of the second conjugate subunit libraries.

[0221] Features and forms of a library

[0222] The method described herein enables the construction of eukaryotic cell libraries with many advantageous features. The library preferably has one or more of the following features:

[0223] 1. Diversity. The library can encode and / or express at least 100, 10 3 10 4 10 5 10 6 10 7 10 8 Or 10 9 Different combinations.

[0224] 2. Uniform Integration. The library may consist of clones containing donor DNA integrated at one fixed locus or a limited number of fixed loci in the cellular DNA. Thus, each clone in the library contains donor DNA at that fixed locus or at least one of the fixed loci. Preferably, the clone contains donor DNA integrated at one or two fixed loci in the cellular DNA. As explained elsewhere herein, the integration site is located at the recognition sequence of a site-specific nuclease. The integration of donor DNA to produce recombinant DNA is described in detail elsewhere herein and can produce different results depending on the number of integration sites. In the case where a single potential integration site is present in the cell used to generate the library, the library will be a clonal library containing donor DNA integrated at a single fixed locus. Thus, all clones in the library contain the binding gene at the same location in the cellular DNA. Alternatively, in the case where multiple potential integration sites are present, the library may be a clonal library containing donor DNA integrated at multiple and / or different fixed loci. Preferably, each clone in the library contains donor DNA integrated at a first and / or second fixed locus. For example, a library may include clones incorporating donor DNA at a first fixed locus, clones incorporating donor DNA at a second fixed locus, and clones incorporating donor DNA at both the first and second fixed loci. In a preferred embodiment, the clones in the library contain only one or two fixed loci, but donor DNA may be incorporating at multiple loci if required for a specific application. Thus, in some libraries, each clone may contain donor DNA incorporating at any one or more of several fixed loci, for example, three, four, five, or six fixed loci. For a library containing a binding subunit integrated at a single site, the clones of the library may contain DNA encoding a first binding subunit integrated at the first fixed locus and DNA encoding a second binding subunit integrated at the second fixed locus, wherein the clones express a multimeric binding compound containing the first and second subunits.

[0225] 3. Consistent Transcription. The relative transcriptional levels of the conjugates among different clones of the library are maintained within controlled limits due to the integration of donor DNA at a controlled number of loci and at the same loci (fixed loci) in different clones. The relatively consistent transcription of the conjugate genes results in comparable expression levels of the conjugates on or from clones within the library. The conjugates displayed on the cell surface of the library can be identical (having the same amino acid sequence) to other conjugates displayed on the same cell. The library can consist of cell clones, each displaying a single member of the conjugate library, or clones displaying multiple members of the conjugate library per cell. Alternatively, the library can contain some clones displaying a single member of the conjugate library, and some clones displaying multiple members (e.g., two) of the conjugate library. Preferably, the clones of the library express one or two members of the conjugate library. For example, a eukaryotic cell clone library according to this disclosure can express at least 10 3 10 4 10 5 10 6 10 7 10 8 Or 10 9 A library of different conjugates, each cell containing donor DNA integrated at a fixed locus in the cell's DNA. The donor DNA encodes the conjugate and may further contain genetic elements for selecting cells in which the donor DNA is integrated at the fixed locus. The cells in the library may contain DNA encoding exogenous site-specific nucleases.

[0226] These and other features of the library are described further elsewhere in this article.

[0227] This disclosure extends to libraries in their pure form, as a population of library clones in the absence of other eukaryotic cells, or mixed with other eukaryotic cells. Other cells can be eukaryotic cells of the same type (e.g., the same cell line) or different cells. Further advantages can be obtained by combining two or more libraries described herein, or by combining libraries described herein with a second library or a second cell population, to facilitate or expand screening or for other uses described herein or obvious to those skilled in the art.

[0228] The libraries described herein, one or more clones obtained from the libraries, or host cells in which DNA encoding conjugates from the libraries is introduced, can be provided in cell cultures. Cells can be cultured and then concentrated to form a cell pellet for easy transport or storage.

[0229] The libraries described herein are typically provided in vitro. The libraries may be in containers, such as cell culture flasks containing library cells suspended in a culture medium, or containers containing a precipitate or concentrated suspension of eukaryotic cells containing the library. The library may constitute at least 75%, 80%, 85%, or 90% of the eukaryotic cells in the container.

[0230] It should be understood that the fixed loci of the integrated donor DNA in the library described herein correspond to the location of the recognition sequence in the method described herein for generating the library. Therefore, all preferences for the recognition sequence locations described herein also apply to the fixed loci of the library described herein.

[0231] Combined display

[0232] Libraries constructed in the context of this disclosure can be cultured to express bispecific conjugates in soluble, secreted or transmembrane form, preferably transmembrane form. If the expressed bispecific conjugates remain on the surface of the cells encoding them, they are considered "displayed." In this context, terms such as "conjugate display," "displayed on a surface / site," "displayed on a cell," and "conjugate display" are used interchangeably. In this context, a library may also be referred to as a displayed library or a display library.

[0233] Preferably, the library displaying the expressed bispecific binder is a bispecific binder library provided for screening against a target.

[0234] Bispecific conjugates may contain or be linked to membrane anchors, such as transmembrane domains, for extracellular display of the conjugate on the cell surface. This may involve the conjugate (i.e., one or both of its subunits) fusing directly to a membrane localization signal (such as a GPI recognition sequence) or to a transmembrane domain (such as the transmembrane domain of the PDGF receptor), as described elsewhere in this paper (Gronwald et al., Proc Natl Acad Sci USA, 1988; 85(10): 3435-9). Retention of the conjugate on the cell surface can also be accomplished indirectly by association with another cell surface-retaining molecule expressed within the same cell. This association molecule may itself be part of a heterodimeric conjugate, such as a tethered antibody heavy chain that associates with a light chain chaperone that is not directly tethered.

[0235] While cell surface immobilization facilitates conjugate selection, many applications require the preparation of cell-free, secretory conjugates. Membrane tethering and soluble secretion can be combined using recapture methods that attach secretory conjugates to cell surface receptors. One approach is to format a conjugate library as a secretory molecule that can bind to a membrane-anchoring molecule expressed within the same cell, which can then be used to capture the secretory conjugate. For example, in the case where an antibody or conjugate molecule is fused to an antibody Fc domain, the membrane-tethered Fc can “sample” secretory conjugate molecules expressed in the same cell, thus revealing the monomeric fraction of the expressed conjugate molecule while the remainder is secreted in a divalent form (US 8,551,715). An alternative is to use a tethered IgG binding domain, such as protein A.

[0236] Kumar et al. (Methods 2012; 56(3), 366–374) reviewed other methods for retaining secretory antibodies on cells that produce them, including encapsulating cells in droplets, matrix-assisted capture, affinity capture surface display (ACSD), secretion and capture technique (SECANT), and “cold capture.” In the examples of ACSD and SECANT, biotinylation is used to promote the immobilization of streptavidin or capture antibodies on the cell surface. The captured molecules then capture the secreted antibodies. In the example of SECANT, biotinylation of the secreted molecules occurs in vivo. Using the “cold capture” technique, secretory antibodies can be detected on producing cells using antibodies against the secreted molecules. This has been proposed because the secreted antibodies bind to the glycocalyx of the cell

[86] . Alternatively, it has been shown that the secreted products are captured by staining antibodies on the cell surface before being endocytosed

[87] . The above methods have been used to identify highly expressed clones in a population, but may be suitable for identifying binding specificity, provided that the binding has a sufficient lifetime on the cell surface.

[0237] Even if the conjugate is directly tethered to the cell surface, a soluble product can be generated. For example, the gene encoding the selected conjugate can be recovered and cloned into an expression vector lacking a membrane-anchoring sequence. Alternatively, expression constructs can be used where the transmembrane domain is encoded in exons flanked by recombination sites, such as the ROX recognition site of the Ore recombinase (Anastassiadis et al., Disease Models & Mechanisms, 2009; 2(9-10): 508-515). The exons encoding the transmembrane domain can be removed by transfection with a gene encoding the Dre recombinase to convert the expression to a secretory form.

[0238] Any of the methods described above or other suitable methods can be used to ensure that the bispecific bindings of the library clone are displayed on the surface of the cells expressing it.

[0239] Derivative Library

[0240] According to the methods for generating libraries described herein, one or more library clones can be selected and used to generate further second-generation libraries. When a library is generated by introducing DNA into a eukaryotic cell as described herein, the library can be cultured to express a bispecific binder, and one or more clones expressing the target bispecific binder can be recovered, for example, by selecting a bispecific binder against a target (optionally two targets) using the methods described herein for identifying bispecific binders against a target. These clones can then be used to generate derivative libraries containing DNA encoding a second bispecific binder library, preferably using the methods described herein for generating libraries.

[0241] To generate a derivative library, one or more recovered clones of DNA are mutated to provide a second bispecific binding library. Mutations can be the addition, substitution, or deletion of one or more nucleotides. When the binding compound is a polypeptide, the mutation will alter the sequence encoding the binding compound by adding, substituting, or deleting one or more amino acids. Mutations can be concentrated in one or more regions, such as one or more CDRs of an antibody molecule, providing a bispecific binding library of different common structural classes in one or more diverse regions, as described elsewhere in this document.

[0242] Generating a derivative library may include isolating DNA from one or more recovered clones, introducing mutations into the DNA to provide a derivative population of donor DNA molecules encoding a second bispecific binding library, and introducing the derivative population of donor DNA molecules into cells to create a cellular derivative library containing DNA encoding a second bispecific binding library.

[0243] DNA isolation may involve obtaining and / or identifying DNA from clones. These methods may include amplifying DNA encoding bispecific binders from recovered clones, such as by PCR and introducing mutations. The DNA can be sequenced and the mutated DNA synthesized.

[0244] Alternatively, mutations can be introduced into the DNA of one or more recovered clones by inducing intraclonal DNA mutations. Thus, derivative libraries can be created from one or more clones without isolating the DNA, for example, through endogenous mutations in avian DT40 cells.

[0245] Antibody display is particularly suitable for creating derivative libraries. Once the antibody gene is isolated, various mutagenesis methods (e.g., error-prone PCR, oligonucleotide directed mutagenesis, strand shuffling) can be used to create a display library from which improved variants can be selected. For example, by strand shuffling, DNA subcloning encoding a selected VH clonal population, oligoclonal mixture, or population can be incorporated into a vector encoding a suitable antibody format and a properly formatted VL chain library (Dyson et al., Anal Biochem, 2011; 417(1): 25-35). Alternatively and again, instances of VH can be introduced into a population of eukaryotic cells encoding and expressing a properly formatted light chain chaperone population (e.g., a VL-CL chain for binding to a heavy chain formatted with IgG or Fab). The VH population can be derived from any of the above sources, including B cells from immunized animals or scFv genes from a selected phage population. In the latter example, cloning the selected VH into the light chain library can combine chain shuffling and reformatting (e.g., to IgG format) in one step.

[0246] A particular advantage of display on eukaryotic cells is the ability to control the rigor of the selection / screening process. By reducing the antigen concentration, cells expressing the highest affinity binder can be distinguished from clones with lower affinity in the population. Visualizing and quantifying the affinity maturation process using flow cytometry is a major advantage of eukaryotic display, as it indicates the percentage of positives in the original library early on and allows for direct comparison of the affinity of selected clones to the parent population during sorting. After sorting, the affinity of individual clones can be determined by pre-incubation with a series of antigen concentrations followed by flow cytometry, or by homogeneous time-resolved fluorescence (TRF) assays, or by surface plasmon resonance (SPR) (Biacore).

[0247] Screening to identify or select conjugates targeting specific targets.

[0248] As described, eukaryotic cell libraries can be used in methods for screening bispecific binders that identify a target (optionally two target targets). Therefore, in one aspect, the libraries described herein are provided as demonstrative libraries for selecting bispecific binders against a target (optionally two target targets).

[0249] Such methods can include:

[0250] The library is provided via the methods for generating the library described herein, or as described herein.

[0251] Culture the library cells to express the bispecific binder.

[0252] Expose the bispecific conjugate to the target, optionally two targets, allowing one or more target conjugates (if present) to recognize the target, optionally recognizing two targets, and

[0253] The test determines whether the target conjugate identifies the target, and two targets can be selected at random.

[0254] The method according to this aspect can be referred to as a method for identifying binders against a target in the context of this application. In this context, the selection or screening of binders also refers to such a method.

[0255] Methods for identifying bispecific binders to targets can utilize a range of target molecule classes, such as proteins, nucleic acids, carbohydrates, lipids, and small molecules. Targets can be provided in soluble form. Targets can be labeled to facilitate detection, for example, they can carry a fluorescent label, or they can be biotinylated. Cells expressing target-specific binders can be separated using directly or indirectly labeled target molecules, where the binder captures the labeled molecule. For example, cells bound to a fluorescently labeled target via binder:target interaction can be detected and sorted by flow cytometry or FACS to separate the desired cells. The choice of cytometry requires directly fluorescently labeled or target molecules labeled with molecules detectable by secondary reagents; for example, a biotinylated target can be added to the cells, and binding to the cell surface can be detected using fluorescently labeled streptavidin (such as streptavidin-phycoerythrin). Another possibility is to immobilize the target molecule or a target-binding secondary reagent on a solid surface, such as magnetic beads or agarose beads, to allow enrichment of target-bound cells. For example, cells bound to biotinylated targets via conjugate-target interactions can be separated on streptavidin-coated substrates, such as streptavidin-coated beads.

[0256] In the libraries used in methods for identifying bispecific binders against a target, oversampling is preferred, i.e., screening for a greater number of clones than are present in the library to ensure efficient representation of the library. Identification of bispecific binders from the very large libraries provided in this disclosure can be performed by flow cytometry sorting, but this takes several days, especially when the library is oversampled. As an alternative, initial selection can be based on the use of retrievable antigens, such as biotinylated antigens recovered on streptavidin-coated magnetic beads. Thus, streptavidin-coated magnetic beads can be used to capture cells bound to biotinylated antigens. Selection using magnetic beads can be used as the sole selection method or can be combined with flow cytometry to obtain better resolution, for example, to distinguish clones with higher expression levels from clones with higher affinity (Feldhaus et al. Nat Biotechnol 2003; 21(2): 163-170; Zhao et al. Journal of Immunological Methods 2011; 363(2): 221-232).

[0257] The in vitro properties of display techniques allow for controlled selection in ways that are impossible with immunotherapy, such as selecting specific conformational states of targets (Biffi et al., Nature Chemistry, 2013; 5(3): 182-186; Gao et al., PNAS, 2009; 106(9): 3071-3076). Targets can be labeled by chemical modifications (fluorescein, biotin) or gene fusions (e.g., proteins fused with epitope tags such as FLAG tags or another protein domain or the entire protein). Tags can be nucleic acids (e.g., DNA, RNA, or non-biological nucleic acids) where the tag portion is fused to the target nucleic acid, or they can be chemically linked to another type of molecule (e.g., proteins). This can be done through chemical conjugation or enzymatic attachment (Gu et al., Nat Biotechnol, 2013; 30(2): 144-152). Nucleic acids can also be fused to targets through translational processes such as ribosome display. The “tag” may be another modification occurring intracellularly (e.g., glycosylation, phosphorylation, ubiquitination, alkylation, PAS, SUMOylation, and other modifications described in the post-translational database (db-PTM) at https: / / awi.cuhk.edu.cn / dbPTM, see Li et al., Nucleic Acids Research, 2022; 50(D1): D471-479), which can be detected by secondary reagents. This will produce conjugates that bind to unknown target proteins based on specific modifications.

[0258] The target can be detected using existing conjugates (e.g., target-specific antibodies) that bind to the target molecule. Using existing conjugates for detection has the added advantage of identifying conjugates in a conjugate library that recognize epitopes different from those used for detection. In this way, paired conjugates can be identified for applications such as sandwich ELISA. Purified target molecules are preferred where possible. Alternatively, the target can be displayed on the surface of a target cell population, and the conjugate on the surface of library cells, a method involving exposing the conjugate to the target by contacting the library cells with the target cells. Recovery of cells expressing the target (e.g., using biotinylated cells expressing the target) will allow for the enrichment of cells expressing conjugates targeting them. This approach is useful in cases involving low-affinity interactions due to the potential for strong affinity effects.

[0259] Target molecules can also be unpurified recombinant or unpurified natural targets, provided that the detection molecule is available for identifying cell binding (as described above). Furthermore, binding of a target molecule to a cell expressing a bispecific binder can be detected indirectly by binding the target molecule to another molecule being detected; for example, cell lysates containing a labeled molecule can be incubated with a library of binders to identify binders against the labeled molecule as well as binders against their associated chaperone proteins. This will generate a panel of antibodies against these chaperones, which can be used to detect or identify the chaperones (e.g., using mass spectrometry). Cell fractionation can be used to enrich targets from specific subcellular locations. Alternatively, differential biotinylation of surface or cytoplasmic fractions can be used in conjunction with streptavidin detection reagents for eukaryotic display (Cho & Shusta. Protein Eng Des Sel [Protein Engineering Design and Selection] 2010; 23(7): 567-577). Detergent-dissolved target formulations are a particularly useful method for intact membrane proteins such as GPCRs and ion channels, which are otherwise difficult to prepare. The presence of detergents may have adverse effects on eukaryotic cells that exhibit conjugate genes that require recycling without causing additional growth of selected cells.

[0260] After target recognition of the target conjugate is detected, cloned cells containing DNA encoding the target conjugate can be recovered. The DNA encoding the conjugate can then be isolated (e.g., identified or amplified) from the recovered clones to obtain the DNA encoding the conjugate that recognizes the target. Optionally, the DNA encoding the conjugate that recognizes the target can be sequenced.

[0261] Exemplary conjugates and targets are detailed elsewhere in this document. A typical example is a bispecific antibody molecule library, which can be screened for binding to a target antigen (optionally two target antigens).

[0262] Phenotypic screening

[0263] In a preferred method for identifying bispecific binders against a target (or optionally two targets), the binder is capable of altering cell signaling and / or cell behavior due to the action of the bispecific binder on one or more targets. In a more preferred method, the bispecific binder is a bispecific antibody.

[0264] Therefore, eukaryotic cell libraries can be used in methods for screening cells with a desired phenotype, wherein the phenotype is generated by the expression of a bispecific binder by the cells. Thus, in one aspect, the use of the libraries described herein for screening cells exhibiting a desired cell phenotype, wherein the phenotype is generated by the expression of a bispecific binder by the cells, is provided.

[0265] Such methods can include:

[0266] The library is provided via the methods for generating the library described herein, or as described herein;

[0267] Culture the library cells to express the bispecific binder; and

[0268] The test determines whether the desired phenotype is exhibited.

[0269] The method described in this respect can be referred to as a method for screening cells with a desired phenotype in the context of this application.

[0270] Bispecific antibodies, which alter cell signaling by binding to ligands or receptors, have a well-documented track record in drug development, and the demand for such therapeutic antibodies continues to grow. Such antibodies, along with other classes of functional bispecific conjugates, also hold promise for controlling cell behavior in vivo and in vitro. However, the ability to control and direct cell behavior depends on the availability of natural ligands that control specific signaling pathways. Unfortunately, many natural ligands, such as those controlling stem cell differentiation (e.g., members of the FGF, TGF-β, Wnt, and Notch superfamily), often exhibit mixed interactions and have limited availability due to their poor expression / stability profiles. Bispecific antibodies, due to their specificity, hold immense potential for controlling cell behavior.

[0271] Historically, the identification of functional antibodies that alter cell signaling has been relatively laborious, involving clone selection, antibody expression, characterization based on sequence and binding properties, conversion to mammalian expression systems, and integration into functional cell-based assays. The eukaryotic display method described herein reduces this effort, but still requires antibody production and addition to separate reporter cell cultures. Therefore, a preferred alternative might be to directly screen for the effects of binding to a bispecific binder library expressed in eukaryotic cells on cell signaling or cell behavior by using the producing cells themselves as reporter cells. Following the introduction of the antibody gene, clones exhibiting altered reporter gene expression or phenotypic changes in the resulting cell population can be identified.

[0272] Many recent publications describe the construction of antibody libraries by cloning antibody gene libraries into reporter cells (Zhang et al., Chemistry & Biology, 2013; 20(5), 734-741; Melidoni et al., Proceedings of the National Academy of Sciences, 2013; 110(44), 17802-1780; Zhang et al., Proceedings of the National Academy of Sciences, 2012; 109(39), 15728-15733). These systems combine expression and reporting within a single cell and typically introduce a population of antibodies selected against a predefined target (e.g., using phage display).

[0273] Antibody gene populations can be introduced into reporter cells to generate libraries using the methods described herein, and clones within the population with antibody-directed phenotypic alterations (e.g., altered gene expression or survival) can be identified. For this phenotypic-directed selection to be performed, it is necessary to maintain the link between the antibody gene present in the expressing cell (genotype) and the antibody expression outcome (phenotype). This has previously been achieved by tethering antibodies to the cell surface (Zhang et al., Chemistry & Biology, 2013; 20(5), 734-741) (as described in the antibody presentation) or by using a semi-solid culture medium to retain secreted antibodies near the producing cells (Melidoni et al., Proceedings of the National Academy of Sciences, 2013; 110(44), 17802-1780). Alternatively, antibodies and other conjugates can be retained intracellularly (Xie et al., Chemistry & Biology, 2014; 2(2), 274-283).

[0274] Bound compounds retained on or around the cell surface can interact with endogenous or exogenous receptors on the cell surface, leading to receptor activation. This can then cause changes in reporter gene expression or cell phenotype. Alternatively, antibodies can block the receptor or ligand to reduce receptor activation. The gene encoding the bound compound that causes the alteration in cell behavior can then be recovered for production or further engineering.

[0275] As an alternative to this “targeted” approach, a population of “raw” antibodies that is not pre-selected for a specific target can be introduced (Yea et al., PNAS [Proceedings of the National Academy of Sciences] 2013; 110: 14966-14971). Cell reporter systems are used to identify members of the behaviorally altered population. This non-targeted approach places particular demands on a large antibody library due to the lack of existing knowledge about the target, as it is not possible to pre-enrich a population of antibodies targeting the target. This approach would benefit from nuclease-directed transgene integration as described herein.

[0276] The “functional selection” approach can be used for other applications involving eukaryotic cell libraries, particularly for higher eukaryotes such as mammalian cells. Antibodies can be fused to signal transduction domains, such that binding to a target leads to receptor activation. Kawahara et al. constructed a chimeric receptor in which an extracellular scFv targeting luciferin was fused to a spacer domain (the D2 domain of the Epo receptor) and various intracellular cytokine receptor domains, including thrombopoietin (Tpo) receptor, erythropoietin (Epo) receptor, gp130, IL-2 receptor, and EGF receptor (Kawahara et al. Biochem Biophys ResCommun 2004; 315(1): 132-138; Sogo et al. Cytokine 2009; 46(1): 127-139; Kawahara et al. Cytokine 2011; 55(3): 402-408). These were introduced into the IL-3-dependent proB cell line (BaF3) (Palacios & Steinmetz. Cell 1985; 41(3): 727-734), in which the chimeric receptor exhibited antigen-dependent activation of the chimeric receptor, leading to IL-3-independent growth. The same approach was used in model experiments to demonstrate antigen-mediated chemoattraction of BaF3 cells

[110] . This approach was extended from stable cultured cells to primary cells, such as Tpo-responsive hematopoietic stem cells (Kawahara et al. Cytokine 2011; 55(3): 402-408) or IL-2-dependent primary T cells, in which normal stimulation by Tpo and IL-2 was replaced by luciferase-directed stimulation of the scFv chimeric receptor, respectively. Thus, chimeric antibody-receptor chimera-based systems can be used to drive target-dependent gene expression or phenotypic changes in primary or stable reporter cells. This capability can be used to identify fusion conjugates that drive signal transduction responses or conjugates that inhibit responses.

[0277] In a modification of the above method, the individual VH and VL domains from the anti-lysozyme antibody were fused into the intracellular domain of Epo cells (Ueda et al., J Immunol Methods, 2000; 241(1-2): 159-170). Cells responded to the addition of lysozyme and grew, indicating the dimerization or stabilization of the antigen-induced individual VH and VL fusion chaperones. Thus, the three interacting components are combined to achieve optimal response in this system.

[0278] Although described here with reference to antibody molecules, the methods described above for identifying conjugates (i.e., antibody molecules) targeting a target can also be adapted and applied to libraries of other types of conjugates.

[0279] Protein fragment complementation represents an alternative system for studying and selecting protein-protein interactions in mammalian cells (Kerppola. Chemical Society Reviews 2009; 38(10): 2876-2886, Michnick et al. Nature Reviews Drug Discovery 2007; 6(7): 569 / 582). This involves restoring the function of cleavage reporter proteins through protein-protein interactions. Reporter proteins used include ubiquitin, DNAE intein, β-galactosidase, dihydrofolate reductase, GFP, firefly luciferase, β-lactamase, and TEV protease. A recent example of this approach is the mammalian membrane 2-hybridization (MaMTH) method, in which the binding of a bait protein:cleavage ubiquitin:transcription factor fusion with a chaperone protein:cleavage ubiquitin restores ubiquitin recognition and releases transcription factors for reporter gene expression (Petschnigg et al. Nat Methods 2014; 11(5): 585-92). Similarly, combinations that interfere with or enhance this interaction can be identified by interfering with signal transduction.

[0280] After the desired phenotype is detected, cloned cells exhibiting the desired phenotype can then be recovered. The DNA encoding the binder can then be isolated (e.g., identified or amplified) from the recovered clones, thus obtaining DNA encoding a bispecific binder that produces the desired phenotype when expressed in cells. Optionally, the DNA encoding the binder that recognizes the target can be sequenced.

[0281] When the desired phenotype is detected, cloned cells exhibiting the desired phenotype can be recovered. Optionally, DNA encoding a bispecific binder is then isolated from the recovered clones, thereby providing DNA encoding a bispecific binder that produces the desired phenotype when expressed in cells. Optionally, the DNA encoding the bispecific binder that produces the desired phenotype can be sequenced.

[0282] Recovery and reformatting of conjugates and encoding DNA

[0283] After identifying the conjugate using methods for identifying conjugates targeting a target, a common next step is to isolate (e.g., identify or amplify) the DNA encoding the conjugate. Optionally, it may be necessary to modify the nucleic acid encoding the conjugate, such as reconstructing the conjugate and / or inserting the coding sequence into a different vector. Therefore, a preferred method for identifying conjugates targeting a target includes isolating the DNA encoding the conjugate that recognizes the target. More preferred methods are described below.

[0284] When the conjugate is an antibody molecule, preferred methods for identifying conjugates targeting a target include isolating DNA encoding the antibody molecule from cloned cells, amplifying DNA encoding at least one antibody variable region (preferably VH and VL domains), and inserting the DNA into a vector to provide a vector encoding the antibody molecule.

[0285] Further aspects and embodiments of this disclosure are set forth in the following numbered paragraphs, which form part of this specification and may be suitably combined with aspects and embodiments described elsewhere herein.

[0286] 1. A method for generating a eukaryotic clonal library containing DNA encoding a diverse library of bispecific binders, the method comprising:

[0287] - Provide eukaryotic cells;

[0288] - Provides a plurality of donor DNA molecules, each donor DNA molecule comprising a first nucleic acid sequence encoding a first binding domain coupled to a first Fc domain; and a second nucleic acid sequence encoding a second binding domain coupled to a second Fc domain;

[0289] - Introduce the donor DNA into these cells; and

[0290] - Culture recombinant cells to generate clones, thereby providing a eukaryotic cell clone library containing DNA encoding a bispecific binding library.

[0291] 2. A method for generating a eukaryotic clonal library containing DNA encoding a diverse library of multimers, such as dimer bispecific binders, the method comprising:

[0292] - A eukaryotic cell containing DNA encoding a first subunit of these bispecific binders, the DNA comprising a first nucleic acid sequence encoding a first binding domain coupled to a first Fc domain;

[0293] - Provide multiple donor DNA molecules encoding the second subunit of these bispecific binders, each donor DNA molecule containing a second nucleic acid sequence encoding a second binding domain coupled to the second Fc domain;

[0294] - Introducing the donor DNA into these cells to produce recombinant cells containing donor DNA integrated into the cell DNA; and

[0295] - Culture recombinant cells to generate clones containing DNA encoding these first and second subunits, thereby providing a eukaryotic clonal library containing DNA encoding a bispecific binding library.

[0296] 3. A method for generating a eukaryotic clonal library containing DNA encoding a diverse library of multimers, such as dimer bispecific binders, the method comprising:

[0297] - Provide eukaryotic cells;

[0298] - Provide a plurality of first donor DNA molecules encoding first subunits of these bispecific binders, each donor DNA molecule containing a first nucleic acid sequence encoding a first binding domain coupled to a first Fc domain;

[0299] - Introduce the first donor DNA into these cells to produce a first set of recombinant cells containing the first donor DNA integrated into the cell DNA;

[0300] - Culturing the first set of recombinant cells to produce the first set of clones containing DNA encoding the first subunit;

[0301] - Provides a plurality of second donor DNA molecules encoding second subunits of these bispecific binders, each donor DNA molecule containing a second nucleic acid sequence encoding a second binding domain coupled to a second Fc domain;

[0302] - Introducing the second donor DNA into the cells of the first set of clones to produce a second set of recombinant cells containing the first and second donor DNA integrated into the cell DNA; and

[0303] - Culturing this second set of recombinant cells to produce a second set of clones containing DNA encoding these first and second subunits, thereby providing a eukaryotic clonal library containing DNA encoding a bispecific binding library.

[0304] 4. A method for generating a eukaryotic clonal library containing DNA encoding a diverse library of multimers, such as dimer bispecific binders, the method comprising:

[0305] - Provide eukaryotic cells;

[0306] - Provide a plurality of first donor DNA molecules encoding first subunits of these bispecific binders, each donor DNA molecule containing a first nucleic acid sequence encoding a first binding domain coupled to a first Fc domain;

[0307] - Provides a plurality of second donor DNA molecules encoding second subunits of these bispecific binders, each donor DNA molecule containing a second nucleic acid sequence encoding a second binding domain coupled to a second Fc domain;

[0308] - Introducing the first and second donor DNAs into these cells to produce recombinant cells containing the first and second donor DNAs integrated into the cell DNA; and

[0309] - These recombinant cells are cultured to generate clones, thereby providing a eukaryotic cell clone library containing donor DNA encoding a bispecific antibody library.

[0310] 5. The method according to any one of the preceding paragraphs, wherein the first and second Fc domains are engineered to facilitate heterodimerization.

[0311] 6. The method according to any one of the preceding paragraphs, wherein one or more steps of introducing the donor DNA into these cells include providing a site-specific nuclease within these cells, wherein the nuclease cleaves a recognition sequence in the cell DNA to produce an integration site at which the donor DNA integrates into the cell DNA, the integration occurring through an endogenous DNA repair mechanism for these cells, thereby producing a recombinant cell containing the donor DNA integrated into the cell DNA.

[0312] 7. The method according to paragraph 6, wherein the recognition sequence is in the NLN gene, TNIK gene, or RAB40B gene.

[0313] 8. The method according to paragraph 6 or 7, wherein the recognition sequence is in the NLN gene.

[0314] 9. The method according to any one of paragraphs 6-8, wherein the recognition sequence is in an intron of a gene.

[0315] 10. The method according to any one of paragraphs 6-9, wherein the recognition sequence is in the open chromatin region of the intron.

[0316] 11. The method according to any one of paragraphs 6-10, wherein the identification sequence is in the enhancer region of the intron.

[0317] 12. The method according to any one of paragraphs 6-11, wherein the recognition sequence is in intron 1, 2 or 6 of the NLN-207 gene.

[0318] 13. The method according to any one of the preceding paragraphs, wherein these bispecific conjugates are bispecific antibody molecules.

[0319] 14. The method according to any one of the preceding paragraphs, wherein these bispecific conjugates are polymers, such as dimers, comprising at least first and second subunits.

[0320] 15. The method according to any one of the preceding paragraphs,

[0321] The first and / or second binding domains contain a single antibody variable domain, preferably derived from the variable domains of a heavy chain antibody, such as VHH or V. NARAnd / or

[0322] The first and / or second binding domains contain two antibody variable domains, preferably single-chain variable fragments (scFv).

[0323] 16. The method according to any one of the preceding paragraphs, wherein the first binding domain comprises a single antibody variable domain, preferably derived from a variable domain of a heavy chain antibody such as VHH or V. NAR More preferably, VHH.

[0324] 17. The method according to any one of the preceding paragraphs, wherein the second binding domain comprises two antibody variable domains, preferably single-chain variable fragments (scFv).

[0325] 18. The method according to any one of the preceding paragraphs, wherein the first Fc domain contains a club-shaped mutation and the second Fc domain contains a mortar-shaped mutation, or wherein the first Fc domain contains a mortar-shaped mutation and the second Fc domain contains a club-shaped mutation.

[0326] 19. The method according to any one of the preceding paragraphs, wherein the flanking sides of these donor DNA molecules are homologous arms.

[0327] 20. The method according to any one of the preceding paragraphs, wherein the donor DNA molecules contain a first promoter operatively linked to the first nucleic acid sequence, and / or wherein the donor DNA molecules contain a second promoter operatively linked to the second nucleic acid sequence.

[0328] 21. The method according to any one of the preceding paragraphs, wherein the donor DNA molecules contain a bidirectional promoter operatively linked to the first and second nucleic acid sequences.

[0329] 22. The method according to any one of the preceding paragraphs, wherein the first and / or second nucleic acid sequence encodes a membrane anchor, such as a transmembrane domain or a membrane localization signal.

[0330] 23. The method according to any one of the preceding paragraphs, wherein the bispecific binding library is a plurality of polypeptides sharing a common structure and having one or more regions of amino acid sequence diversity.

[0331] 24. The method according to any one of the preceding paragraphs, wherein the bispecific binding library is a library of different bispecific antibody molecules in one or more complementarity-determining regions.

[0332] 25. The method according to any one of the preceding paragraphs, wherein the bispecific conjugate further comprises one or more additional subunits that may be introduced onto the same donor DNA as the first or second subunit, or may be integrated into a separate site in the cellular DNA.

[0333] 26. The method according to any one of the preceding paragraphs, wherein the cells are higher eukaryotic cells with a genome size greater than 2 x 10^7 base pairs.

[0334] 27. The method according to any one of the preceding paragraphs, wherein the cells are mammalian, bird, insect, or plant cells.

[0335] 28. The method according to paragraph 27, wherein the cells are mammalian cells, preferably human cells.

[0336] 29. The method described in paragraph 28, wherein the cells are HEK293 cells, Chinese hamster ovary (CHO) cells, T lymphocyte lineage cells or B lymphocyte lineage cells or any cell line listed in the “Encyclopedia of Cancer Cell Lines” or the “COSMIC Catalogue of Cancer Somatic Mutations”.

[0337] 30. The method described in paragraph 29, wherein these cells are primary T cells or T cell lines.

[0338] 31. The method described in paragraph 29, wherein these cells are primary B cells, B cell lines, pre-B cell lines, or progenitor B cell lines.

[0339] 32. The method described in paragraph 31, wherein the cells are mouse pre-B cell line 1624-5, IL-3 dependent progenitor B cell line Ba / F3, or chicken DT40B cells.

[0340] 33. The method according to any one of paragraphs 6-32, wherein the recognition sequence of the site-specific nuclease is in the genomic DNA of these cells.

[0341] 34. The method according to any one of paragraphs 6-33, wherein the recognition sequence of the site-specific nuclease is in the episomal DNA within these cells.

[0342] 35. The method according to any one of paragraphs 6-34, wherein the recognition sequence of the site-specific nuclease occurs only once or twice in the cellular DNA.

[0343] 36. The method according to any one of paragraphs 6-35, wherein a site-specific nuclease cuts cellular DNA to produce a double-strand break as an integration site.

[0344] 37. The method according to any one of paragraphs 6-36, wherein the nuclease is a broad-spectrum nuclease.

[0345] 38. The method according to any one of paragraphs 6-37, wherein the nuclease is a zinc finger nuclease (ZFN).

[0346] 39. The method according to any one of paragraphs 6-38, wherein the nuclease is a TALE nuclease.

[0347] 40. The method according to any one of paragraphs 6-39, wherein the nuclease is a nucleic acid-guided nuclease.

[0348] 41. The method according to paragraph 40, wherein DNA cutting is directed by a CRISPR / Cas system.

[0349] 42. The method according to any one of the preceding paragraphs, wherein the donor DNA is integrated into the cell DNA via homologous recombination.

[0350] 43. The method according to any one of the preceding paragraphs, wherein the donor DNA is integrated into the genomic DNA by non-homologous end joining or microhomologous directional end joining.

[0351] 44. The method according to any one of the preceding paragraphs, wherein the donor DNA comprises a genetic element for selecting cells into which the donor DNA is integrated.

[0352] 45. The method according to any one of the preceding paragraphs, wherein the donor DNA is integrated into the cell DNA and the expression of the bispecific binder or its subunits and / or the expression of the genetic selection element are placed under the control of a promoter present within the cell DNA.

[0353] 46. ​​The method according to any one of the preceding paragraphs, wherein the donor DNA comprises a sequence operatively linked to a promoter encoding the bispecific binder or a subunit thereof.

[0354] 47. The method according to any one of the preceding paragraphs, wherein the library contains at least 100, 10^3, 10^4, 10^5, 10^6, 10^7, 10^8 or 10^9 clones, each clone being derived from a single recombinant cell produced by integrating donor DNA.

[0355] 48. The method according to any one of the preceding paragraphs, wherein the library encodes at least 100, 10^3, 10^4, 10^5, 10^6, 10^7, 10^8 or 10^9 different bispecific binders.

[0356] 49. The method according to any one of the preceding paragraphs, wherein each clone contains integrated donor DNA encoding only one or both members of the bispecific binding library.

[0357] 50. The method according to any one of the preceding paragraphs, wherein the eukaryotic cells are diploid and contain a site-specific nuclease recognition sequence at a repetitive fixed locus in the cell DNA.

[0358] 51. The method according to any one of the preceding paragraphs, wherein each clone contains integrated donor DNA encoding a single member of the bispecific binding library.

[0359] 52. The method according to any one of the preceding paragraphs, wherein each of the donor DNA molecules encodes a single bispecific binder or a subunit thereof.

[0360] 53. The method according to any one of the preceding paragraphs, wherein these bispecific binders are displayed on the cell surface.

[0361] 54. The method according to any one of the preceding paragraphs, wherein these bispecific binders are secreted from these cells.

[0362] 55. The method according to any one of the preceding paragraphs, the method further comprising:

[0363] The library of cells was cultured to express these bispecific binders.

[0364] Recover clones expressing one or more bispecific binding compounds, and

[0365] A derivative library is generated from one or more recovered clones, wherein the derivative library contains DNA encoding a second bispecific binding library.

[0366] 56. The method according to paragraph 55, wherein generating the derivative library comprises isolating donor DNA from the one or more recovered clones, introducing mutations into the DNA to provide a derivative population of donor DNA molecules encoding a second bispecific binding library, and introducing the derivative population of the donor DNA molecules into a cell to create a cellular derivative library containing DNA encoding the second bispecific binding library.

[0367] 57. The method according to paragraph 55, wherein generating the derivative library includes introducing mutations into the donor DNA in the one or more recovered clones by inducing intraclonal DNA mutations.

[0368] 58. A method for generating a diverse library of bispecific binders, the method comprising generating the library by means of any one of the preceding paragraphs, and culturing library cells to express the bispecific binders.

[0369] 59. A library generated by the method according to any one of the preceding paragraphs.

[0370] 60. A method for screening cells having a desired phenotype, wherein the phenotype is generated by the cell expressing a bispecific binder, the method comprising:

[0371] - Provide the library via the method for generating the library according to any one of the preceding paragraphs, or provide the library according to paragraph 59;

[0372] -Culturing cells containing this library to express these bispecific binders; and

[0373] - Detect whether the desired phenotype is exhibited.

[0374] 61. The method according to paragraph 60, wherein the phenotype is the expression of a reporter gene in a cell expressing the conjugate.

[0375] 62. The method according to paragraph 60 or paragraph 61, the method further comprising recovering cells of a clone that expresses a bispecific binder that produces the desired phenotype.

[0376] 63. The method according to paragraph 62, further comprising isolating DNA encoding the bispecific binder from the recovered clone to obtain DNA encoding the bispecific binder that produces the desired phenotype.

[0377] 64. A method for screening bispecific binders that identify a target, the method comprising:

[0378] - Provide a library via the method for generating a library according to any one of paragraphs 1 to 58, or provide a library according to paragraph 59;

[0379] -Culturing cells containing this library to express these bispecific binders;

[0380] - Expose these bispecific binders to the target, optionally two targets, allowing the target to be recognized by one or more target binders if present; and

[0381] - Detect whether the target conjugate recognizes the target, optionally both targets.

[0382] 65. The method according to paragraph 64, wherein the target is provided in a soluble form.

[0383] 66. The method according to paragraph 64, wherein the target is displayed on the surface of a target cell population and the bispecific binders are displayed on the surface of the library cells, the method comprising exposing the bispecific binders to the target by contacting the library cells with the target cells.

[0384] 67. The method according to any one of paragraphs 60 to 66, wherein the bispecific conjugates are bispecific antibody molecules and the target is an antigen.

[0385] 68. The method according to any one of paragraphs 64 to 67, the method further comprising detecting target identification of the target bispecific binder and recovering cloned cells containing DNA encoding the target bispecific binder.

[0386] 69. The method according to paragraph 68, further comprising isolating DNA encoding the bispecific binder from the recovered clone to obtain DNA encoding the bispecific binder that recognizes the target.

[0387] 70. The method according to paragraph 63 or paragraph 69, the method comprising introducing a mutation or converting the DNA into modified DNA encoding a reconstructive bispecific binder.

[0388] 71. The method according to paragraph 70, wherein the bispecific conjugate comprises scFv, and the method comprises converting DNA encoding the scFv into DNA encoding Ig or a fragment thereof, while maintaining the original variable VH and VL strand pairing.

[0389] 72. The method according to paragraphs 63, 69, 70 or 71, the method further comprising introducing the DNA into a host cell.

[0390] 73. An in vitro library of eukaryotic cell clones expressing a diverse library of at least 100, 10^3, 10^4, 10^5, 10^6, 10^7, 10^8 or 10^9 different bispecific binders, each cell containing recombinant DNA wherein donor DNA encoding the bispecific binder or its subunits is integrated into at least a first and / or a second fixed locus in the cell DNA.

[0391] 74. The library according to paragraph 59 or 73, wherein the clones further contain DNA comprising a genetic element for selecting cells that express a encoded bispecific binder.

[0392] 75. The libraries described in paragraphs 59, 73, or 74, wherein these bispecific binders are antibody molecules.

[0393] 76. A container containing eukaryotic cells, the eukaryotic cells containing a library according to any one of paragraphs 59 or 73 to 75.

[0394] 77. The container according to paragraph 76, wherein the library constitutes at least 75%, 80%, 85%, or 90% of the eukaryotic cells in the container.

[0395] 78. The container according to paragraph 76 or 77, wherein the container is a cell culture flask containing cells of the library suspended in a culture medium.

[0396] 79. The container as described in paragraph 76 or 77, the container containing a precipitate or concentrated suspension of eukaryotic cells containing the library.

[0397] 80. The use of site-specific nucleases for targeted cleavage of cellular DNA in the construction of eukaryotic cell libraries containing DNA encoding bispecific binding libraries, wherein nuclease-mediated DNA cleavage enhances site-specific integration of binding genes through endogenous cellular DNA repair mechanisms.

[0398] 81. The use of the library according to any one of paragraphs 59 or 73 to 75 as a demonstrative library for selecting bispecific bindings against a desired target.

[0399] 82. The use of the library according to any one of paragraphs 59 or 73 to 75 for screening cells exhibiting a desired cell phenotype, wherein the phenotype is generated by the expression of a bispecific conjugate by the cells.

[0400] General information

[0401] Unless otherwise stated, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure pertains and should be read in accordance with this disclosure.

[0402] As used herein, the term "promoter" or "regulatory sequence" refers to a nucleic acid fragment that controls the transcription of one or more coding sequences, is upstream of the transcription start site relative to the coding sequence, and is structurally characterized by the presence of a DNA-dependent RNA polymerase binding site, a transcription start site, and any other DNA sequence, including but not limited to transcription factor binding sites, repressor and activator protein binding sites, and any other nucleotide sequence known to those skilled in the art to directly or indirectly regulate the amount of transcription by the promoter. A "constitutive" promoter is a promoter that is active in most tissues under most physiological and developmental conditions. An "inducible" and / or "repressive" promoter is a promoter that is physiologically or developmentally regulated to be induced and / or repressed, for example, by the application of chemical inducing agents or repressive signals.

[0403] As used herein, the term "operably ligated" refers to the ligation of polynucleotide elements in a functional relationship. A nucleic acid is "operably ligated" to another nucleic acid sequence when it is positioned to have a functional relationship with that sequence. For example, if a transcriptional regulatory sequence, such as a promoter, affects the transcription of a coding sequence, it is operably ligated to the coding sequence. Operable ligation means that the ligated DNA sequences are typically contiguous, and in cases where the joining of two protein-coding regions is required, contiguous and within the reading frame.

[0404] The terms “protein” or “polypeptide” are used interchangeably and refer to molecules composed of chains of amino acids without specifying a particular mode of action, size, three-dimensional structure, or origin.

[0405] Sequence identity

[0406] It should be understood that each nucleic acid molecule, protein fragment, polypeptide, peptide, derivative peptide, or construct identified herein by a given sequence identifier (SEQ ID NO) is not limited to the specific sequence disclosed herein. Each coding sequence identified herein encodes a given protein fragment, polypeptide, peptide, derivative peptide, or construct, or is itself a protein fragment, polypeptide, construct, peptide, or derivative peptide.

[0407] Throughout this application, whenever a specific nucleotide sequence encoding a given protein fragment, polypeptide, peptide, or derived peptide is mentioned (e.g., SEQ ID NO: X), it may be replaced with:

[0408] i. A nucleotide sequence comprising a nucleotide sequence having at least 60% sequence identity with SEQ ID NO: X;

[0409] ii. A nucleotide sequence that differs from the sequence of a nucleic acid molecule in (i) due to the degeneracy of the genetic code; or

[0410] iii. A nucleotide sequence that encodes an amino acid sequence having at least 60% amino acid identity or similarity to the amino acid sequence encoded by the nucleotide sequence SEQ ID NO: X.

[0411] Another preferred level of sequence identity or similarity is 65%. Another preferred level of sequence identity or similarity is 70%. Another preferred level of sequence identity or similarity is 75%. Another preferred level of sequence identity or similarity is 80%. Another preferred level of sequence identity or similarity is 85%. Another preferred level of sequence identity or similarity is 90%. Another preferred level of sequence identity or similarity is 95%. Another preferred level of sequence identity or similarity is 97%. Another preferred level of sequence identity or similarity is 99%. Another preferred level of sequence identity or similarity is 99.5%.

[0412] In this application, each reference to a specific amino acid sequence SEQ ID NO (for example, SEQ ID NO: Y) can be replaced with: a polypeptide represented by an amino acid sequence containing a sequence having at least 60% sequence identity or similarity to the amino acid sequence SEQ ID NO: Y. Another preferred level of sequence identity or similarity is 65%. Another preferred level of sequence identity or similarity is 70%. Another preferred level of sequence identity or similarity is 75%. Another preferred level of sequence identity or similarity is 80%. Another preferred level of sequence identity or similarity is 85%. Another preferred level of sequence identity or similarity is 90%. Another preferred level of sequence identity or similarity is 95%. Another preferred level of sequence identity or similarity is 97%. Another preferred level of sequence identity or similarity is 99%. Another preferred level of sequence identity or similarity is 99.5%.

[0413] In a further preferred embodiment, each nucleotide sequence or amino acid sequence described herein (based on its percentage of identity or similarity with a given nucleotide sequence or amino acid sequence, respectively) has a similarity to a given nucleotide or amino acid sequence of at least 60%, at least 61%, at least 62%, at least 63%, at least 64%, at least 65%, at least 66%, at least 67%, at least 68%, at least 69%, at least 70%, at least 71%, at least 72%, at least 73%, at least 74%, or more. At least 75%, at least 76%, at least 77%, at least 78%, at least 79%, at least 80%, at least 81%, at least 82%, at least 83%, at least 84%, at least 85%, at least 86%, at least 87%, at least 88%, at least 89%, at least 90%, at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99%, at least 99.5%, or 100% similarity or identity.

[0414] Each non-coding nucleotide sequence (i.e., promoter or another regulatory region) can be replaced by a nucleotide sequence that has at least 60% sequence identity or similarity to a specific nucleotide sequence SEQ ID NO (e.g., SEQ ID NO: A). The preferred nucleotide sequence has at least 60%, at least 61%, at least 62%, at least 63%, at least 64%, at least 65%, at least 66%, at least 67%, at least 68%, at least 69%, at least 70%, at least 71%, at least 72%, at least 73%, at least 74%, at least 75%, at least 76%, at least 77%, at least 78%, at least 79%, at least 80%, at least 81%, at least 82%, at least 83%, at least 84%, at least 85%, at least 86%, at least 87%, at least 88%, at least 89%, at least 90%, at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99%, at least 99.5%, or 100% identity with SEQ ID NO: A. In a preferred embodiment, such a non-coding nucleotide sequence, such as a promoter, at least exhibits or exerts the activity of such a non-coding nucleotide sequence, such as the activity of promoters known to those skilled in the art.

[0415] The terms “homology,” “sequence identity,” etc., are used interchangeably herein. “Sequence identity” describes the relationship between two or more amino acid (peptide or protein) sequences or two or more nucleic acid (polynucleotide) sequences, as determined by sequence comparison. In preferred embodiments, sequence identity is based on the full length or a portion thereof of two given SEQ ID NOs, preferably calculated based on the full length. A portion preferably refers to at least 50%, 60%, 70%, 80%, 90%, or 100% of the two SEQ ID NOs. In the art, “identity” also refers to the degree of sequence correlation between amino acid or nucleic acid sequences, as determined, such as by matching strings of these sequences. “Similarity” between two amino acid sequences is determined by comparing the amino acid sequence of a polypeptide and its conserved amino acid substitutions with the sequence of a second polypeptide. "Identity" and "similarity" can be readily calculated by known methods, including but not limited to those described in Bioinformatics and the Cell: Modern Computational Approaches in Genomics, Proteomics and Transcriptomics, Xia X., Springer International Publishing, New York, 2018; and Bioinformatics: Sequence and Genome Analysis, Mount D., Cold Spring Harbor Laboratory Press, New York, 2004, each incorporated herein by reference.

[0416] "Sequence identity" and "sequence similarity" can be determined by aligning two peptide or nucleotide sequences using a global or local alignment algorithm, depending on the length of the two sequences. Sequences of similar length are preferably aligned using a global alignment algorithm (e.g., Needleman-Wunsch), which optimizes sequence alignment across the entire length, while sequences with significantly different lengths are preferably aligned using a local alignment algorithm (e.g., Smith-Waterman). Sequences can be described as "substantially identical" or "substantially similar" when they share at least a minimum percentage of sequence identity (as described below) when optimally aligned using, for example, programs such as EMBOSS needle or EMBOSS water with default parameters.

[0417] When two sequences are similar in length, global alignment is suitable for determining sequence identity. When sequences have significantly different total lengths, local alignment, such as those using the Smith-Waltman algorithm, is preferred. EMBOSS needle uses the Niederman-Onsch global alignment algorithm to align two sequences across the entire length (full length), maximizing the number of matches and minimizing the number of gaps. EMBOSS water uses the Smith-Waltman local alignment algorithm. Generally, the default parameters for EMBOSSneedle and EMBOSS water are used: gap opening penalty = 10 (nucleotide sequence) / 10 (protein) and gap extension penalty = 0.5 (nucleotide sequence) / 0.5 (protein). For nucleotide sequences, the default scoring matrix used is DNAfull; for proteins, the default scoring matrix is ​​Blosum62 (Henikoff & Henikoff, 1992, PNAS 89, 915-919, incorporated herein by reference).

[0418] Alternatively, percentage similarity or identity can be determined by searching public databases using algorithms such as FASTA and BLAST. Therefore, the nucleic acid and protein sequences of some embodiments of this disclosure can be further used as “query sequences” to search public databases, for example, to identify other family members or related sequences. Such searches can be performed using the BLASTn and BLASTx programs (version 2.0) of Altschul et al. (1990) J. Mol. Biol. [Journal of Molecular Biology] 215: 403-10 (incorporated by reference). NBLAST nucleotide searches can be performed using the BLAST program (score = 100, word length = 12) to obtain nucleotide sequences homologous to the nucleic acid molecules of this disclosure. BLAST protein searches can be performed using the BLASTx program (score = 50, word length = 3) to obtain amino acid sequences homologous to the protein molecules of this disclosure. To obtain gapped alignments for comparison targets, gapped BLAST can be used as described in Altschul et al., (1997) Nucleic Acids Res. [Nucleic Acids Research] 25(17): 3389-3402 (incorporated by reference). When using BLAST and gapped BLAST procedures, the default parameters of each procedure (e.g., BLASTx and BLASTn) can be used. See the homepage of the National Center for Biotechnology Information, accessible on the World Wide Web. www.ncbi.nlm.nih.gov / .

[0419] Optionally, when determining amino acid similarity, those skilled in the art can also consider so-called conserved amino acid substitutions. As used herein, a “conservative” amino acid substitution refers to the interchangeability of residues having similar side chains. The table below provides examples of categories of amino acid residues used for conserved substitutions.

[0420]

[0421] Classification of conserved amino acid residues that can be substituted:

[0422] 1 A S T 2 D E 3 N Q 4 R K 5 I L M 6 F Y W

[0423] Physical and functional classification of amino acid residue substitutability:

[0424] Alcohol group containing residues S and T Aliphatic residues I, L, V and M Cycloalkenyl related residues F, H, W and Y Hydrophobic residues A, C, F, G, H, I, L, M, R, T, V, W and Y Negatively charged residues D and E Polar residues C, D, E, H, K, N, Q, R, S and T Positively charged residues H, K and R Small residues A, C, D, G, N, P, S, T and V Very small residues A, G and S Residues involved in corner formation A, C, D, E, G, H, K, N, Q, R, S, P and T Flexible residues Q, T, K, S, G, P, D, E and R

[0425] For example, a group of amino acids with aliphatic side chains includes glycine, alanine, valine, leucine, and isoleucine; a group of amino acids with aliphatic-hydroxyl side chains includes serine and threonine; a group of amino acids with amide-containing side chains includes asparagine and glutamine; a group of amino acids with aromatic side chains includes phenylalanine, tyrosine, and tryptophan; a group of amino acids with basic side chains includes lysine, arginine, and histidine; and a group of amino acids with sulfur-containing side chains includes cysteine ​​and methionine. Preferred conserved amino acid substitution groups are: valine-leucine-isoleucine, phenylalanine-tyrosine, lysine-arginine, alanine-valine, and asparagine-glutamine. Substitution variants of the amino acid sequences disclosed herein are those in which at least one residue has been removed from the disclosed sequence and a different residue has been inserted at its position. Preferably, the amino acid changes are conserved. Preferred conservative substitutions for each naturally occurring amino acid are as follows: Ala to Ser; Arg to Lys; Asn to Gln or His; Asp to Glu; Cys to Ser or Ala; Gln to Asn; Glu to Asp; Gly to Pro; His to Asn or Gln; Ile to Leu or Val; Leu to Ile or Val; Lys to Arg; Gln or Glu; Met to Leu or Ile; Phe to Met, Leu or Tyr; Ser to Thr; Thr to Ser; Trp to Tyr; Tyr to Trp or Phe; and Val to Ile or Leu.

[0426] In this document and its claims, the verb “comprising” and its variations are used in a non-limiting sense to mean that which includes or contains the items following the word, but does not exclude items not specifically mentioned. Thus, as used herein, the terms “comprising” (“comprising”, “comprises”, “comprised of”, etc.) are synonymous with “including” (“includes”) or “containing” (“contains”) and are inclusive or open-ended, and do not exclude additional, unlisted members, elements, or method steps.

[0427] Additionally, the verb "composed of..." can be replaced with "substantially composed of...", meaning that the composition described herein may contain one or more additional components besides the specifically specified components, which do not alter the unique characteristics of this disclosure. Furthermore, the verb "composed of..." can be replaced with "substantially composed of...", meaning that the method described herein may include one or more additional steps besides the specifically specified steps, which do not alter the unique characteristics of this disclosure.

[0428] As used herein, the singular forms “a / an” and “the” include both singular and plural indicators unless the context clearly indicates otherwise; for example, “antibody” should be understood to mean one or more antibodies. Therefore, the terms “a / an,” “one or more,” and “at least one” are used interchangeably herein.

[0429] The range of values ​​represented by the endpoints includes all numbers and fractions contained within the corresponding range, as well as the listed endpoints.

[0430] As used in this article, "at least" means that specific value or more. For example, "at least 2" is understood as "2 or more", i.e., 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, etc.

[0431] Furthermore, the terms first, second, third, etc., used in the specification and claims are used to distinguish similar elements and are not necessarily used to describe order or chronological sequence. It should be understood that such terms are interchangeable where appropriate, and the embodiments described herein can operate in orders other than those described or illustrated herein.

[0432] When the word “about” or “approximately” is used in conjunction with a numerical value (e.g., about 10), it preferably means that the value is a given value (10) plus or minus 5% of that value, preferably plus or minus 1%.

[0433] As used herein, the term “and / or” means that one or more of the stated circumstances may occur individually or in combination with at least one of the stated circumstances, up to and including all of the stated circumstances.

[0434] This document describes various embodiments. Unless otherwise stated, each embodiment described herein can be combined together. Titles, subtitles, headings, etc., are used herein for ease of reading only and are not intended to limit or constrain the content of this disclosure in any way.

[0435] All patent applications, patents, and print publications cited herein are incorporated herein in their entirety by reference, except for any definitions, subject matter disclaimers, or disclaimers, and the extent of exclusion is such that the incorporated material is inconsistent with the express disclosure herein, in which case the language of this disclosure shall prevail.

[0436] Those skilled in the art will recognize that many methods and materials are similar to or equivalent to those described and materials herein, and can be used in the practice of this disclosure. In fact, this disclosure is by no means limited to the methods and materials described herein. It should also be understood that this disclosure encompasses aspects of the foregoing disclosure as exemplified by the following examples.

[0437] This disclosure is further illustrated by the following examples, which should not be construed as limiting the scope. Attached Figure Description

[0438] Figure 1 A schematic diagram of the bispecific antibody expression cassette pINT177 (a dual CMV promoter expression cassette for surface expression).

[0439] pINT178 is a dual-promoter expression cassette for surface expression. Expression of each antibody is provided by a different promoter (i.e., CMV and pEF).

[0440] pINT179 is a dual CMV promoter surface expression cassette similar to pINT177, but in which scFv is attached to the "mortar" portion of the bispecific molecule and VHH is attached to the "pestle" portion of the bispecific molecule.

[0441] Figure 2 Non-reducing protein blots of the bispecific antibodies described in Table 1. Arrows indicate the locations of heterodimers. Protein blots were detected using anti-Fc(a) and anti-Myc(b) detection antibodies.

[0442] Figure 3Expression of monospecific (A, B) and bispecific (C, D) antibodies on the cell surface. Cells were stained with human anti-Fc phycoerythrin antibody (A, C) or anti-Myc-FITC antibody (B, D).

[0443] Figure 4 Expression of monospecific (B, C) and bispecific (D) antibodies on the cell surface. Cells were stained with anti-Flag-FITC antibody, and untransfected HEK293F cells were shown as a comparison (A).

[0444] Figure 5 The binding of antigens and antibodies on cell surfaces. Cells expressing monospecific (B, C) or bispecific (D) antibodies were incubated with 1 nM biotinylated antigen 1 and streptavidin-APC. Untransfected HEK293F cells are shown as a comparison (A).

[0445] Figure 6 The binding of antigens and antibodies on cell surfaces. Cells expressing monospecific (B, C) or bispecific (D) antibodies were incubated with 59 nM antigen 2-FITC-labeled antigen. Untransfected HEK293F cells are shown as a comparison (A).

[0446] Figure 7 The antibody binds to two antigens expressed on the cell surface as bispecific antibodies. Cells are transfected with expression cassettes 3-6 listed in Table Z, where the antibody gene is expressed under either a dual CMV promoter (A, B) or a CMV-pEF promoter (C, D). The antibody is incubated with 59 nM antigen 2-FITC-labeled antigen and 1 nM biotinylated antigen 1. In the latter case, detection is performed using streptavidin-APC.

[0447] Figure 8 Titration binding of antigen 1 was performed with or without a saturated concentration of unlabeled antigen 2. Cells were transfected with a pINT178-4B12VHH-“Flag”:3G01 scFv-“Myc” construct, and staining with increased concentrations of antigen 1 was detected using streptavidin-APC. A saturated concentration of antigen 2 (180 nM) was applied prior to antigen 1 staining.

[0448] Figure 9 FACS selects conjugates from bispecific mammalian display libraries. In the presence of saturated antigen 2-FITC, mammalian display libraries containing VHH libraries (obtained directly from immunized llama clones (A) or after two rounds of phage display selection of antigen 1 (B)) were stained with biotinylated antigen 1 (APC secondary detection) to enrich bispecific molecules capable of binding both antigens simultaneously. Boxes indicate sorting gates.

[0449] Figure 10 Monoclonal conjugates were identified and confirmed from mammalian display bispecific libraries. FACS selected outputs were presented as single clones, and the binding of the bispecific antibody to antigen 1 was tested using a DELFIA TRF assay. Signals against antigen 1-biotin immobilized on streptavidin were displayed as gray bars, while control values ​​for streptavidin-only antigen (i.e., without antigen 1) were covered by black bars.

[0450] Figure 11 A schematic diagram of the carrier region between the dual-specific carrier pestle and mortar technology and the homologous arm. Example

[0451] The exact antigen targets in the following examples are not disclosed and are referred to as "antigen 1" and "antigen 2". Both antigen 1 and antigen 2 are representative transmembrane proteins expressed by certain immune cells.

[0452] Example 1: Construction of an expression vector for bispecific antibodies

[0453] To enable genetic selection of conjugates (e.g., antibodies, proteins, or peptides), it is necessary to introduce a gene encoding that conjugate and drive its expression from an exogenous or endogenous promoter. Antibodies represent one of the most commonly used classes of conjugates, and they can be formatted for expression in different forms. In the following examples, we describe expression in a dual-gene form, where different classes of variable binding entities (e.g., VHH and scFv) are fused to the same Fc domain. In this example and others, VHH, denoted as 4B12, is an antibody that specifically binds to antigen 1, and scFv, denoted as 4F07 and 3G01, are antibodies (fragments) that specifically bind to antigen 2.

[0454] In order to express these bispecific antibodies in production cells (such as higher eukaryotes), it is necessary to express individual monospecific binding polypeptide chains and promote the proper pairing of these chains. This can be achieved by introducing separate plasmids encoding each chain or by introducing them into a single plasmid using multiple promoters. Figure 1 a-1c and Figure 11This image shows an organization of similar expression cassettes developed for expressing membrane-anchored bispecific antibodies, within the same vector backbone. These expression cassettes are created using a combination of gene synthesis, restriction enzyme digestion, and polymerase chain reaction amplification using standard elements such as promoters, antibiotic resistance genes, and polyA sequences. A first plasmid was created to bind to the cytomegalovirus promoter (CMV promoter), which drives the expression of VHH antibody and a heavy chain constant region containing the "mortar" mutation (T366S:L368A:Y407V; see also P Carter, Bispecific human IgG by design. J Immunol Methods 248, 7-15 (2001)), and is used to express scFv, followed by a heavy chain constant region containing the "mortar" mutation (T366W; see also P Carter, Bispecific human IgG by design. J Immunol Methods 248, 7-15 (2001)). The latter cistron contains the PDGF receptor transmembrane domain to provide membrane anchoring.

[0455] Secretion of isolated entities in the endoplasmic reticulum is directed by two distinct leader sequences. VHH secretion is directed by the BM40 leader sequence [Holden, P., Keene, DR, Lunstrum, GP, HP, & Horton, WA (2005). Secretion of cartilage oligomeric matrix protein is affected by the signal peptide. [J Biol Chem, 280(17), 17172-17179]. Following this are the ApaLI and BstEII restriction sites, which allow in-frame cloning of any antibody. scFv secretion is directed by a leader derived from the mouse VH gene, which divides an intron. The coding sequence for the signal peptide is followed by the NcoI and NotI sites, allowing in-frame cloning of genes encoding antibodies with different binding properties than the primary antibody.

[0456] In a similar manner, individual VHH and scFv antibodies were combined into the same backbone, but using different promoters. Elongation factor-1α protein is ubiquitous and highly expressed in most eukaryotic cells, and its promoter (pEF promoter, also known as EF1α promoter) is commonly used to drive transgene expression [Kim, DW, Uetsuki, T., Kaziro, Y., Yamaguchi, N., & Sugano, S. (1990). Use of the human elongation factor 1 alpha promoter as aversatile and efficient expression system. [Use of the human elongation factor 1α promoter as a versatile and efficient expression system] Gene, 91(2), 217-223]. In pINT178, the pEF promoter is used to drive scFv antibody expression. A polyadenylation site (BGH polyA) derived from bovine growth hormone is present at the end of each expression cassette.

[0457] Example 2: Western blot of bispecific antibodies expressed in mammalian cells

[0458] Western blotting was performed to confirm bispecific antibody formation in cells transfected with the bispecific antibody display vector. HEK293 cells were transfected with the vectors listed in Table 1 using TALEN-directed integration via a homologous arm targeting the NLN locus (see also Example 3 of WO 2023 / 025834 and Examples 6 and 7 of WO 2015 / 166272, both of which are incorporated herein by reference). Protein fractions were prepared using manufacturer-recommended RIPA buffer (Thermo Fisher Scientific, catalog 89901). The antibody-containing protein samples were resolved by gel electrophoresis under non-reducing conditions and then transferred to a PVDF membrane. The proteins separated on the membrane were then detected using anti-Fc and anti-Myc antibodies conjugated with horseradish peroxidase (HRP) and detected with an HRP-enhanced chemiluminescent substrate.

[0459] The detection of anti-Fc and anti-Myc antibodies in soluble and membrane-anchored antibodies under non-reducing conditions showed that pINT178 mainly contained heterodimers. Figure 2 (a, b, lane 3). Meanwhile, pINT179 (ibid., lane 4) showed a higher concentration of soluble scFv-“mortar” homodimers, but no detectable VHH-anchored homodimers. Therefore, both vectors are able to produce heterodimers anchored to the cell surface, making it possible to construct bispecific mammalian display libraries.

[0460] Table 1. Vectors used for transfecting HEK293 cells

[0461]

[0462] Example 3: Bispecific antibody expression and antigen binding via flow cytometry

[0463] To determine whether bispecific antibody expression occurred on the cell surface, cells were transfected using the monospecific or bispecific vectors listed in Table 2. Twenty-four hours after transfection, the volume of the large-volume culture was doubled, and blast fungicide (7.5 μg / ml) was added 24 hours later. Cells were cultured with the same concentration of blast fungicide, and the culture medium was changed every 3–4 days.

[0464] Fifteen days post-transfection, the display level of the bispecific construct was analyzed using only 2.5 x 10^5 cells / sample and 0.1 mL incubation volume (following the protocols described in Example 3 of WO 2023 / 025834 and Examples 6 and 7 of WO 2015 / 166272). Figure 3 The results showed that 15 days post-transfection, cells expressed a bispecific 4B12:3G01 antibody on their cell surface, which could be detected using a phycoerythrin-labeled anti-Fc antibody. The binding of the anti-Myc FITC-labeled antibody to the bispecific 4B12:3G01 antibody was lower than that to the monospecific bispecific 3G01 antibody containing two Myc tags, indicating that the former antibody construct presented a heterodimer containing only one Myc tag.

[0465] Next, to further verify the formation of heterodimers on the cell surface, staining was performed using an antibody labeled with anti-Flag FITC. Figure 4 The positive staining shown confirms that the non-anchored bispecific antibody arm with the Flag tag interacts with the anchored arm via a mortar and pestle mechanism, leading to the formation of a bispecific heterodimer on the cell surface. Combined with these results, this demonstrates the expression of the bispecific antibody.

[0466] Fifteen days after transfection, cells were labeled with antigen 1 according to the following protocol:

[0467] 1. Harvest and wash the cells, adjusting the sample size to 2.5 × 10⁶ cells per sample. Centrifuge the cells at 250g for 4 min at room temperature, wash the cells with 1 ml PBS + 0.1% BSA (4°C), centrifuge the cells at 250g for 4 min at room temperature, and resuspend them in 1 ml PBS + 1% BSA.

[0468] 2. Add biotinylated antigen to a final concentration of 1 nM and incubate at 4°C for 30 min.

[0469] 3. The cells were centrifuged at 250g for 5 min and washed twice with 1 ml of 0.1% BSA.

[0470] 4. Add 1 μl of APC-labeled streptavidin (Thermo Scientific, catalog number 21629) to 0.1 ml PBS + 1% BSA and incubate in the dark at 4°C for 15 minutes.

[0471] 5. After centrifuging at 250g for 5 min, wash the cells twice with 1 ml of 0.1% BSA.

[0472] 6. Resuspend them in 55 μl of ice-cold PBS + 0.1% BSA.

[0473] 7. Perform viability staining with 7-AAD (eBioscience, catalog number 00-6993-50) at a 1:50 dilution.

[0474] The result is Figure 5 As shown in the figure, flow cytometry analysis of labeled antigen 1 bound to the non-anchored arm clearly shows that, at antigen concentrations as low as 1 nM, the heterodimers formed on the cell surface retain the binding properties of the VHH arm.

[0475] Next, similar experiments were conducted to investigate the binding of the scFv anchoring arm to FITC-labeled antigen 2. For example... Figure 6 As shown, the binding of labeled antigen 2 to scFv is detected by bispecific and monospecific antibodies.

[0476] Figure 7 The binding of the bispecific antibodies to the bispecific antibodies expressed under the dual CMV or CMV-pEF promoters is shown. In short, essentially as described above, cells expressing 4B12VHH and 4F07scFv or 4B12VHH and 3G01scFv are stained with antigens 1 and 2. Double staining reveals the detection of the bispecific antibodies, independent of the promoter used to express the individual arms. Optimal antigen binding and bispecificity display of 4B12:4F07 were observed using the pCMV pEF promoter combination.

[0477] Cell lines stained with both antigens showed a double-positive population. However, the antigen concentration was unsaturated, making it unclear whether the bispecific antibody bound one antigen at a time or both simultaneously. Therefore, according to the protocol described above, antigen 1 titration was performed with and without saturated unlabeled antigen 2. For bispecific antibodies capable of binding both antigens simultaneously, similar results are expected under both conditions, while antibodies unable to bind simultaneously may exhibit a weakened VHH arm for antigen 1 binding due to high occupancy of the scFv arm by antigen 2. Figure 8The data shown indicates that there is almost no competition for antigen 1 binding, suggesting that both binding arms are occupied simultaneously by their respective antigens.

[0478] Table 2. Vectors used for transfecting HEK293 cells

[0479]

[0480] Example 4: Identification of novel bispecific antibody conjugates from mammalian display libraries

[0481] To validate the applicability of this system in selecting monoclonal conjugates from bispecific molecular libraries displayed on the surface of mammalian cells, restriction cloning was performed using ApaL1 and BstEII to clone anti-antigen 1VHH libraries directly from immunized llamas or after two rounds of phage display into the pINT178 vector system. As previously described in detail, mammalian display libraries of bispecific molecules were generated from the resulting pINT178 plasmid libraries (WO 2015 / 166272, WO 2023 / 025834, Parthiban et al. MAbs[Monoclonal Antibody] 2019; 11(5); 884-898; Dyson et al. MAbs[Monoclonal Antibody] 2020; 12(1): 1829335). Briefly, HEK293 cells were transfected using the MaxCyte electroporation system and passaged with antibiotic selection (7.5 μg / ml blastonium). At least 6 days post-transfection, magnetically activated cell sorting (MACS) was performed to target antibody expression in order to generate 10 5 -10 7 An integrated mammalian display library of clones. Figure 9 FACS selection from these bispecific mammalian display libraries is shown (inset A shows FACS of a library directly cloned from an immunized llama, inset B shows the phage display-enriched input library). Cell staining was performed as previously and above (Parthiban et al. MAbs [Monoclonal Antibody] 2019; WO 2015 / 166272). Cells were sorted against the binding of the antibody to biotinylated antigen 1 in the presence of saturated antigen 2-FITC (alphacocyanin (APC) secondary assay) (). Figure 9The sorting gates are indicated by boxes in the image, used to enrich bispecific molecules capable of binding to two antigens simultaneously. As expected, antigen 2 binding was retained as a fixed form in the bispecific construct (y-axis) relative to the directly cloned immune library, while antigen 1 binding (x-axis) was enriched in the phage display input. Low-frequency VHH antigen 1 bindings compatible with the bispecific form and antigen 2 occupancy in the scFv-Fc arm were clearly detectable in the llama immune library. The enriched bispecific antibody clones were isolated and screened for binding to antigen 1 using DELFIA TRF assay. Figure 10 ). (Dissociation-enhanced lanthanide fluorescence immunoassay) is a time-resolved fluorescence (TRF) intensity technique (PerkinElmer) that is commonly used industrially as an alternative binding assay to ELISA. Ninety-seven monoclonal conjugates were directly identified from the immunological library, and 81 were identified from the phage enrichment library. In summary, Examples 1-4 demonstrate the display of bispecific antibody libraries on mammalian cells and their use in identifying novel target bispecific molecules. Sequence Overview

[0482] SEQ ID NO SEQUENCE DESCRIPTION 1 I-SceI recognition sequence 2 Conserved structural motif 3 NLN gene 4 TNIK gene 5 PARP 11 gene 6 RAB40B gene 7 ABI2 gene 8 RNF19B gene 9 PKIA gene 10 FTCD gene 11 NLN-201 intron 1 12 NLN-201 intron 2 13 NLN-201 intron 3 14 NLN-201 intron 4 15 NLN-201 intron 5 16 NLN-201 intron 6 17 NLN-201 intron 7 18 NLN-201 intron 8 19 NLN-201 intron 9 20 NLN-201 intron 10 21 NLN-201 intron 11 22 NLN-201 intron 12 23 NLN-207 intron 5 24 NLN-207 intron 6 25 GGGS flexible linker unit

[0483] .

Claims

1. A method for generating a eukaryotic clonal library containing DNA encoding a diverse library of bispecific binders, the method comprising: - Provide eukaryotic cells; - Provides a plurality of donor DNA molecules, each donor DNA molecule comprising a first nucleic acid sequence encoding a first binding domain coupled to a first Fc domain; and a second nucleic acid sequence encoding a second binding domain coupled to a second Fc domain, wherein the first and second Fc domains are engineered to facilitate heterodimerization; - Introduce the donor DNA into these cells; and - Culture recombinant cells to generate clones, thereby providing a eukaryotic clonal library containing DNA encoding a bispecific binding library.

2. The method of claim 1, wherein the first and / or second binding domain comprises a single antibody variable domain, preferably derived from a variable domain of a heavy chain antibody such as VHH or V. NAR And / or wherein the first and / or second binding domains comprise two antibody variable domains, preferably single-chain variable fragments (scFv).

3. The method according to claim 1 or 2, wherein the first binding domain comprises a single antibody variable domain, preferably derived from a variable domain of a heavy chain antibody such as VHH or V. NAR More preferably, VHH.

4. The method according to any one of claims 1-3, wherein the second binding domain comprises two antibody variable domains, preferably single-chain variable fragments (scFv).

5. The method according to any one of claims 1-4, wherein the first Fc domain contains a club-shaped mutation and the second Fc domain contains a mortar-shaped mutation, or wherein the first Fc domain contains a mortar-shaped mutation and the second Fc domain contains a club-shaped mutation.

6. The method according to any one of claims 1-5, wherein the step of introducing the donor DNA into these cells comprises providing a site-specific nuclease within these cells, wherein the nuclease cleaves a recognition sequence in the cellular DNA to produce an integration site at which the donor DNA integrates into the cellular DNA, the integration occurring through an endogenous DNA repair mechanism for these cells, thereby producing recombinant cells containing donor DNA integrated into the cellular DNA.

7. The method according to any one of claims 1-6, wherein the flanking sides of these donor DNA molecules are homologous arms.

8. The method according to any one of claims 1-7, wherein the donor DNA molecules contain a first promoter operatively linked to the first nucleic acid sequence, and / or wherein the donor DNA molecules contain a second promoter operatively linked to the second nucleic acid sequence.

9. The method according to any one of claims 1-8, wherein the donor DNA molecules comprise a bidirectional promoter operatively linked to the first and second nucleic acid sequences.

10. The method according to any one of claims 1-9, wherein the first and / or second nucleic acid sequence encodes a membrane anchor, such as a transmembrane domain or a membrane localization signal.

11. The method according to any one of claims 1-10, wherein the eukaryotic cells have a genome size greater than 2 × 10⁻⁶. 7 Higher eukaryotic cells with 1 base pair, preferably wherein these cells are mammalian, bird, insect or plant cells, more preferably wherein these cells are mammalian cells.

12. A method for screening bispecific binders that identify a target, the method comprising: - The library is generated by the method according to any one of claims 1-11; -Culturing cells containing this library to express these bispecific binders; - Expose these bispecific binders to the target, optionally two targets, allowing the target to be recognized by one or more target binders if present; as well as - Detect whether the target conjugate recognizes the target, optionally both targets. Optionally, it further includes recovering cells of a clone containing DNA encoding the target conjugate, and optionally further includes isolating the DNA encoding the conjugate from the recovered clone.

13. A method for screening cells having a desired phenotype, wherein the phenotype is generated by the cell expressing a bispecific binder, the method comprising: - The library is generated by the method according to any one of claims 1-11; -Culturing cells containing this library to express these bispecific binders; as well as - Detect whether the desired phenotype is exhibited. Optionally, it may further include recovering cells from a clone that expresses a conjugate that produces the desired phenotype. Optionally, it may further include isolating the DNA encoding the conjugate from the recovered clone.

14. A library generated by the method according to any one of claims 1-11.

15. An in vitro library of eukaryotic cell clones expressing a diverse library of at least 10^3, 10^4, 10^5, 10^6, 10^7, 10^8 or 10^9 different bispecific binders, each cell containing recombinant DNA, wherein donor DNA encoding a bispecific binder or a subunit of a bispecific binder is integrated into at least a first and / or a second locus in the cell DNA; optionally, wherein the locus is a fixed locus.

Images