Construction of diverse synthetic peptide and polypeptide libraries
The present invention concerns the design and construction of diverse peptide and polypeptide libraries. In particular, the invention concerns methods of analytical database design for creating datasets using multiple relevant parameters as filters, and methods for generating sequence diversity by directed multisyntheses oligonucleotide synthesis. The present methods enable the reduction of large complex annotated databases to simpler datasets of related sequences, based upon relevant single or multiple key parameters that can be individually directly defined. The methods further enable the creation of diverse libraries based on this approach, using multisynthetic collections of discrete and degenerate oligonucleotides to capture the diverse collection of sequences, or portions thereof.
This is a non-provisional application claiming priority under 35 U.S.C. §119(e) from U.S. provisional patent application No. 60/849,035, filed Oct. 2, 2006, the entire disclosure of which is expressly incorporated by reference herein.
FIELD OF THE INVENTIONThe present invention concerns the design and construction of diverse peptide and polypeptide libraries. In particular, the invention concerns methods of analytical database design for creating datasets using multiple relevant parameters as filters, and methods for generating sequence diversity by directed multisyntheses oligonucleotide synthesis. The present methods enable the reduction of large complex annotated databases to simpler datasets of related sequences, based upon relevant single or multiple key parameters that can be individually directly defined. The methods further enable the creation of diverse libraries based on this approach, using multisynthetic collections of discrete and degenerate oligonucleotides to capture the diverse collection of sequences, or portions thereof.
BACKGROUND OF THE INVENTIONThe development of peptide- or polypeptide-based drug candidates often starts with the screening of libraries of related peptide or polypeptide sequences. Thus the first step for the selection of therapeutic antibody candidates usually is the creation of a highly diverse library of antibody sequences.
Several methods for the design and construction of diverse antibody libraries are known in the art.
It has been described that the diversity of a filamentous phage-based combinatorial antibody library can be increased by shuffling of the heavy and light chain genes (Kang et al., Proc. Natl. Acad. Sci. USA, 88:11120-11123, (1991)) or by introducing random mutations into the library by error-prone polymerase chain reactions (PCR) (Gram et al., Proc. Natl. Acad. Sci. USA, 89:3576-3580, (1992)). The use of defined frameworks as the basis for generating antibody libraries has been described by Barbas et al., Proc. Natl. Acad. Sci. USA 89:4457-4461 (1992) (randomizing CD3-H3); Barbas et al., Gene 137:57-62 (2003) (extending randomization to Vκ CDR3); and Hayanashi et al., Biotechniques 17:310 (1994) (simultaneous mutagenesis of antibody CDR regions by overlap extension and PCR). Others report combination of CDR-H3 libraries with a single VL, gene (Nissim et al., EMBO J. 13:692-698 (1994)), a limited set of VL genes (De Kruif et al., J. Mol. Biol. 248:97-105 (1995)); or a randomized repertoire of VL genes (Griffiths et al., EMBO J. 13:3245-3260 (1994)).
See also U.S. Pat. Nos. 5,667,988; 6,096,551; 7,067,284 describing methods for producing antibody libraries using universal or randomized immunoglobulin light chains.
Knappik et al., J. Mol. Biol. 296:57-86 (2000) describe a different concept for designing and constructing human antibody libraries, designated HuCAL (Human Combinatorial Antibody Libraries). This approach is based on the finding that each of the human VH and VL subfamilies that is frequently used during an immune response is represented by one consensus framework, resulting in seven HuCAL consensus genes for heavy chains and seven HuCAL consensus genes for light chains, which yield 49 possible combinations. All genes are made by total synthesis, taking into consideration codon usage, unfavorable residues that promote protein aggregation, and unique and general restriction sites flanking all CDRs. The approach leads to the generation of modular antibody genes containing CDRs that can be converted into different antibody formats, as needed. The design and synthesis of HuCAL antibody libraries is described in U.S. Pat. Nos. 6,300,064; 6,696,248; 6,706,484; and 6,828,422.
Despite these and other advances there a great need for new, efficient methods for the design and construction of highly diverse (poly)peptide, such as antibody, libraries.
SUMMARY OF THE INVENTIONThe present invention concerns the design and construction of diverse peptide and polypeptide libraries.
In one aspect, the invention concerns a method for diversity analysis of a database comprising related amino acid sequences characterized by at least one shared sequence motif, comprising the steps of:
(a) aligning the related amino acid sequences;
(b) creating a first dataset by applying a predetermined combination of two or more filters to the related amino acid sequences comprising the shared sequence motif;
(c) analyzing the first dataset for positional amino acid usage frequency within the shared sequence motif; and
(d) creating a second dataset characterized by a minimum threshold amino acid usage frequency at one or more amino acid positions within the shared sequence motif
In step (d) a minimum threshold amino acid usage frequency can be assigned to any and all amino acid positions within the shared sequence motif
In one particular embodiment, a minimum threshold amino acid usage frequency is assigned to the majority of amino acid positions within the shared sequence motif In another particular embodiment, a minimum threshold amino acid usage frequency is assigned to all amino acid positions within the shared sequence motif In various embodiments, the minimum threshold amino acid usage frequencies assigned to specific amino acid positions within the shared sequence motif can be identical or different.
In a further embodiment, the minimum threshold amino acid usage frequency is set to provide a minimum sum amino acid usage for the majority of amino acid positions within the shared sequence motif
In a still further embodiment, the minimum threshold amino acid usage frequency is set to provide a minimum sum amino acid usage for all amino acid positions within said shared sequence motif
The minimum sum amino acid usage can be set to any desired level, and in particular embodiments it is at least about 40%, or at least about 45%, or at least about 50%, or at least about 55%, or at least about 60%, or at least about 65%, or at least about 70%, or at least about 75%, or at least about 80%, or at least about 85%, or at least about 90%.
In another embodiment, the related amino acid sequences are antibody sequences.
In yet another embodiment, the related amino acid sequences comprise antibody heavy chain sequences.
In a further embodiment, the related amino acid sequences comprise antibody light chain sequences.
If the related amino acid sequences are antibody sequences, the shared sequence motif may, for example, be a CDR sequence, such as a CDR1, CDR2 or CDR3 sequence.
There are no limitations on the nature or number of the filters that can be used in step (b) of the method of the present invention. In a particular embodiment, in the case of antibody sequences, the predetermined combination of filters can be selected from the group consisting of (1) the isotype of an antibody heavy or light chain; (2) the length of one or more of CDR1, CDR2 and CDR3 sequences; (3) the presence of one or more predetermined amino acid residues at one or more predetermined positions within one or more CDR1, CDR2 and CDR3 sequences; (4) type of framework; (5) antigen to which the antibody binds; (6) affinity of the antibody; and (7) positional amino acid residues outside the CDR sequences.
In a further embodiment, at least one of the antibody heavy and/or light chain CDR1, CDR2 and CDR3 sequences is size matched. This parameter can, for example, be combined with the isotype of the antibody heavy and/or light chain sequences, as an additional filter.
In various embodiments, the positional amino acid usage frequency is at least about 3%, or at least about 5%, or at least about 10%, or at least about 15%, or set between about 3% and about 15%, or between about 5% and about 10%.
In another embodiment of the methods of the present invention, the same positional amino acid usage frequency characterizes each amino acid within said CDR sequence. In an alternative embodiment, the positional amino acid usage frequencies differ at at least two amino acid residues within said CDR sequence.
In another embodiment, the predetermined combination of filters includes the type of framework.
In yet another embodiment, both antibody heavy and light chain sequences are analyzed. Optionally, the antibody heavy chain sequences are paired to predetermined antibody light chain characteristics, or the antibody light chain sequences are paired to predetermined antibody heavy chain characteristics.
In a further embodiment, the related antibody sequences are from at least one functional antibody.
In a still further embodiment, at least one of the filters applied in step (b) of the method of the invention is the germline sequence most similar to the framework sequence of the heavy and/or light chain of a functional antibody.
Without limitation, the functional antibody may, for example, bind to a polypeptide selected from the group consisting of cell surface and soluble receptors, cytokines, growth factors, enzymes; proteases; and hormones. Thus, the antibody may bind to a cytokine, such as an interleukin, e.g. IL-1, IL-2, IL-3, IL-4, IL-5, IL-6, IL-10, IL-11, IL-12, IL-15, IL-17, IL-18, IL-23, and their respective family members. Alternatively, the cytokine may, for example, be selected from the group consisting of interferons-α, -β, and -γ (IFN-α, -β, and -γ), tumor necrosis factor-α, and -β (TNF-α and -β), TWEAK, RANKL, BLys, RANTES, MCP-1, MIP-1α, MIP-1β, SDF-1, colony stimulating factor (CSF), granulocyte colony stimulating factor (G-CSF), and granulocyte macrophage colony stimulating factor (GMCSF).
The polypeptide to which the antibody binds may also be a growth factor, including, without limitation, nerve growth factor (NGF), insulin-like growth factor 1 (IGF-1), epidermal growth factor (EGF), plateled derived growth factor (PDGF), vascular endothelial growth factor (VEGF), placental growth factor (PLGF), tissue growth factor-α (TGF-α), and tissue growth factor-β (TGF-β).
In another embodiment, the functional antibody binds to a hapten, e.g. Dig, Bio, DNP, or FITC.
In yet another embodiment of the methods herein, the related amino acid sequences originate from members of a family of secreted or extracellular proteins, which can be cytokines, for example.
In a specific embodiment, the cytokine is interferon-a, and the related amino acid sequences are sequences of IFN-α subtypes.
In a particular embodiment, the invention further comprises the step of synthesizing a physical library of related amino acid sequences that is designed with the aid of the datasets identified.
In a certain embodiment of this method, the library is synthesized by generating a discrete number of defined or degenerate oligonucleotides such that only defined amino acids are generated.
In a further embodiment, the diversity of the physical library produced exceeds the diversity of a library which is a physical representation of the datasets identified. This can, for example, result from the fact that at least one amino acid not meeting the minimum threshold amino acid usage frequency is also synthesized to provide said diversity.
In a still further embodiment, the diversity of the physical library produced is less than the diversity of a library which is a physical representation of the datasets identified. This can results from the fact that not all amino acids meeting the minimum threshold amino acid usage frequency are synthesized, for example.
In another embodiment, the dataset comprises antibody heavy and/or light chain sequence, which may include one or more CDRs.
In yet another embodiment, the CDRs are cloned into a scaffold of framework sequences, which may, optionally, be the most frequently used framework sequences in the database comprising said CDRs.
The physical library may be expressed using any expression system, including all prokaryotic and eukaryotic expression systems.
In a specific embodiment, the physical library is expressed and displayed using a phagemid display, mRNA display, microbial cell display, mammalian cell display, microbead display technique, antibody array, or display based on protein-DNA linkage.
In another embodiment of the invention, the library is screened for one or more chemical and/or biological properties of its members. Such properties may include, without limitation, half-life, potency, efficacy, binding affinity, and immunogenicity.
In yet another embodiment, amino acid side-chain diversity is introduced into members of the library at one or more amino acid positions.
In a particular embodiment, the amino acid side-chain diversity is introduced by providing amino acid residues with at least two different side-chain chemical functionalities at said amino acid position or positions.
In other embodiments, at least 30%, or at least 50%, or at least 55%, or at least 60% of all amino acid chemistries are represented at each amino acid position.
Preferably, amino acid said side-chain diversity is introduced by using combinatorial degenerate oligonucleotide synthesis.
In another aspect, the invention concerns a method of producing a combinatorial library of peptide or polypeptide sequences, comprising introducing amino acid side-chain chemical diversity into the peptide or polypeptide sequences at two or more amino acid positions, using combinatorial oligonucleotide synthesis.
In one embodiment, the amino acid side-chain chemical diversity is designed to mimic naturally occurring diversity in said peptide or polypeptide sequences.
The library can be any type of library, including, but not limited to, antibody libraries.
In a specific embodiment, the antibody library comprises antibody heavy chain variable domain sequences.
In another embodiment, the library comprises antibody light chain variable domain sequences.
In yet another embodiment, the library is a combinatorial single-chain variable fragment (scFv) library.
In a further embodiment, the antibody library is a library of Fab, Fab′, or F(ab′)2 fragments.
A. Definitions
Unless defined otherwise, technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. Singleton et al., Dictionary of Microbiology and Molecular Biology 2nd ed., J. Wiley & Sons (New York, N.Y. 1994), provides one skilled in the art with a general guide to many of the terms used in the present application.
One skilled in the art will recognize many methods and materials similar or equivalent to those described herein, which could be used in the practice of the present invention. Indeed, the present invention is in no way limited to the methods and materials described. For purposes of the present invention, the following terms are defined below.
The phrase “shared sequence motif” is used herein in the broadest sense and is used to refer to a pattern of amino acid residues common between two or more peptide or polypeptide sequences. Sequence motifs can be readily identified by a variety of pattern discovery algorithms, such as those discussed in the detailed description of the invention.
In the context of the present invention, the term “antibody” (Ab) is used in the broadest sense and includes immunogloblins which exhibit binding specificity to a specific antigen as well as immunoglobulins and other antibody-like molecules which lack antigen specificity. Polypeptides of the latter kind are, for example, produced at low levels by the lymph system and, at increased levels, by myelomas. In the present application, the term “antibody” specifically covers, without limitation, monoclonal antibodies, polyclonal antibodies, and antibody fragments.
“Native antibodies” are usually heterotetrameric glycoproteins of about 150,000 daltons, composed of two identical light (L) chains and two identical heavy (H) chains. Each light chain is linked to a heavy chain by covalent disulfide bond(s), while the number of disulfide linkages varies between the heavy chains of different immunoglobulin isotypes. Each heavy and light chain also has regularly spaced intrachain disulfide bridges. Each heavy chain has, at one end, a variable domain (VH) followed by a number of constant domains. Each light chain has a variable domain at one end (VL) and a constant domain at its other end; the constant domain of the light chain is aligned with the first constant domain of the heavy chain, and the light chain variable domain is aligned with the variable domain of the heavy chain. Particular amino acid residues are believed to form an interface between the light- and heavy-chain variable domains, Chothia et al., J. Mol. Biol. 186:651 (1985); Novotny and Haber, Proc. Natl. Acad. Sci. U.S.A. 82:4592 (1985).
The term “variable” with reference to antibody chains is used to refer to portions of the antibody chains which differ extensively in sequence among antibodies and participate in the binding and specificity of each particular antibody for its particular antigen. Such variability is concentrated in three segments called hypervariable regions both in the light chain and the heavy chain variable domains. The more highly conserved portions of variable domains are called the framework region (FR). The variable domains of native heavy and light chains each comprise four FRs (FR1, FR2, FR3 and FR4, respectively), largely adopting a β-sheet configuration, connected by three hypervariable regions, which form loops connecting, and in some cases forming part of, the β-sheet structure. The hypervariable regions in each chain are held together in close proximity by the FRs and, with the hypervariable regions from the other chain, contribute to the formation of the antigen-binding site of antibodies (see Kabat et al., Sequences of Proteins of Immunological Interest, 5th Ed. Public Health Service, National Institutes of Health, Bethesda, Md. (1991), pages 647-669). The constant domains are not involved directly in binding an antibody to an antigen, but exhibit various effector functions, such as participation of the antibody in antibody-dependent cellular toxicity.
The term “hypervariable region” when used herein refers to the amino acid residues of an antibody which are responsible for antigen-binding. The hypervariable region comprises amino acid residues from a “complementarity determining region” or “CDR” (i.e., residues 30-36 (L1), 46-55 (L2) and 86-96 (L3) in the light chain variable domain and 30-35 (H1), 47-58 (H2) and 93-101 (H3) in the heavy chain variable domain; MacCallum et al,. J Mol Biol. 1996.
The term “framework region” refers to the art recognized portions of an antibody variable region that exist between the more divergent CDR regions. Such framework regions are typically referred to as frameworks 1 through 4 (FR1, FR2, FR3, and FR4) and provide a scaffold for holding, in three-dimensional space, the three CDRs found in a heavy or light chain antibody variable region, such that the CDRs can form an antigen-binding surface.
Depending on the amino acid sequence of the constant domain of their heavy chains, antibodies can be assigned to different classes. There are five major classes of antibodies IgA, IgD, IgE, IgG, and IgM, and several of these may be further divided into subclasses (isotypes), e.g., IgG1, IgG2, IgG3, IgG4, IgA, and IgA2.
The heavy-chain constant domains that correspond to the different classes of immunoglobulins are called α, δ, ε, γ, and μ, respectively.
The “light chains” of antibodies from any vertebrate species can be assigned to one of two clearly distinct types, called kappa (κ) and lambda (λ), based on the amino acid sequences of their constant domains.
“Antibody fragments” comprise a portion of a full length antibody, generally the antigen binding or variable domain thereof. Examples of antibody fragments include, but are not limited to, Fab, Fab′, F(ab′)2, Dab, and Fv fragments, linear antibodies, single-chain antibody molecules, diabodies, and multispecific antibodies formed from antibody fragments.
The term “monoclonal antibody” is used to refer to an antibody molecule synthesized by a single clone of B cells. The modifier “monoclonal” indicates the character of the antibody as being obtained from a substantially homogeneous population of antibodies, and is not to be construed as requiring production of the antibody by any particular method. Thus, monoclonal antibodies may be made by the hybridoma method first described by Kohler and Milstein, Nature 256:495 (1975); Eur. J. Immunol. 6:511 (1976), by recombinant DNA techniques, or may also be isolated from phage antibody libraries.
The term “polyclonal antibody” is used to refer to a population of antibody molecules synthesized by a population of B cells.
“Single-chain Fv” or “sFv” antibody fragments comprise the VH and VL domains of antibody, wherein these domains are present in a single polypeptide chain. Generally, the Fv polypeptide further comprises a polypeptide linker between the VH and VL domains which enables the sFv to form the desired structure for antigen binding. For a review of sFv see Pückthun in The Pharmacology of Monoclonal Antibodies, vol. 113, Rosenburg and Moore eds. Springer-Verlag, New York, pp. 269-315 (1994). Single-chain antibodies are disclosed, for example in WO 88/06630 and WO 92/01047.
As used herein the term “antibody binding regions” refers to one or more portions of an immunoglobulin or antibody variable region capable of binding an antigen(s). Typically, the antibody binding region is, for example, an antibody light chain (VL) (or variable region thereof), an antibody heavy chain (VH) (or variable region thereof), a heavy chain Fd region, a combined antibody light and heavy chain (or variable region thereof) such as a Fab, F(ab′)2, single domain, or single chain antibody (scFv), or a full length antibody, for example, an IgG (e.g., an IgG1, IgG2, IgG3, or IgG4 subtype), IgA1, IgA2, IgD, IgE, or IgM antibody.
The term “threshold frequency of occurrence” refers to a criterion of the invention which requires that a selected sequence for use in a library herein be derived from a sequence which has been determined to be a sequence favored to be expressed. Depending on the ultimate goal, such as the required degree of diversity, the desired size of library, the “threshold frequency of occurrence” can be set at different levels.
The term “amino acid” or “amino acid residue” typically refers to an amino acid having its art recognized definition such as an amino acid selected from the group consisting of: alanine (Ala); arginine (Arg); asparagine (Asn); aspartic acid (Asp); cysteine (Cys); glutamine (Gin); glutamic acid (Glu); glycine (Gly); histidine (His); isoleucine (Ile): leucine (Leu); lysine (Lys); methionine (Met); phenylalanine (Phe); proline (Pro); serine (Ser); threonine (Thr); tryptophan (Trp); tyrosine (Tyr); and valine (Val) although modified, synthetic, or rare amino acids may be used as desired. Thus, modified and unusual amino acids listed in 37 CFR 1.822(b)(4) are specifically included within this definition and expressly incorporated herein by reference. Amino acids can be subdivided into various sub-groups. Thus, amino acids can be grouped as having a nonpolar side chain (e.g., Ala, Cys, Ile, Leu, Met, Phe, Pro, Val); a negatively charged side chain (e.g., Asp, Glu); a positively charged side chain (e.g., Arg, His, Lys); or an uncharged polar side chain (e.g., Asn, Cys, Gln, Gly, His, Met, Phe, Ser, Thr, Trp, and Tyr). Amino acids can also be grouped as small amino acids (Gly, Ala), nucleophilic amino acids (Ser, His, Thr, Cys), hydrophobic amino acids (Val, Leu, Ile, Met, Pro), aromatic amino acids (Phe, Tyr, Trp, Asp, Glu), amides (Asp, Glu), and basic amino acids (Lys, Arg) (see,
The term “conserved amino acid residue” refers to an amino acid residue determined to occur with a high frequency, typically at least 50% or more (e.g., at about 60%, 70%, 80%, 90%, 95%, or higher), for a given residue position in two or more amino acid sequences compared.
The term “semi-conserved amino acid residue” refers to amino acid residues determined to occur with a high frequency between two or more amino acid sequences compared for a given residue position. When 2-3 residues, preferably 2 residues, that together, are represented at a frequency of about 40% of the time or higher (e.g., 50%, 60%, 70%, 80%, 90% or higher), the residues are determined to be semi-conserved.
The term “variable amino acid residue” refers to amino acid residues determined to occur with a variable frequency between two or more sequences compared for a given residue position. When many residues appear at a given position, the residue position is determined to be variable.
The term “variability profile” refers to the cataloguing of amino acids and their respective frequencies of occurrence present at a particular amino acid position within a polypeptide sequence, such as within a CDR of an antibody.
The term “polynucleotide(s)” refers to nucleic acids such as DNA molecules and RNA molecules and analogs thereof (e.g., DNA or RNA generated using nucleotide analogs or using nucleic acid chemistry). As desired, the polynucleotides may be made synthetically, e.g., using art-recognized nucleic acid chemistry or enzymatically using, e.g., a polymerase, and, if desired, be modified. Typical modifications include methylation, biotinylation, and other art-known modifications. In addition, the nucleic acid molecule can be single-stranded or double-stranded and, where desired, linked to a detectable moiety.
The term “mutagenesis” refers to, unless otherwise specified, any art recognized technique for altering a polynucleotide or polypeptide sequence. Preferred types of mutagenesis include error prone PCR mutagenesis, saturation mutagenesis, or other site directed mutagenesis.
The term “vector” is used to refer to a rDNA molecule capable of autonomous replication in a cell and to which a DNA segment, e.g., gene or polynucleotide, can be operatively linked so as to bring about replication of the attached segment. Vectors capable of directing the expression of genes encoding for one or more polypeptides are referred to herein as “expression vectors.”
The term “primer,” as used herein, refers to a polynucleotide whether purified from a nucleic acid restriction digestion reaction or produced synthetically, which is capable of acting as a point of initiation of nucleic acid synthesis when placed under conditions in which synthesis of a primer extension product which is complementary to a nucleic acid strand is induced. Such conditions may include the presence of nucleotides and a DNA polymerase, reverse transcriptase and the like, at a suitable temperature and pH. The primer is preferably single stranded, but may also be in a double stranded form. The primer must be sufficiently long to prime the synthesis of extension products in the presence of the agents for polymerization. The exact lengths of the primers will depend on many factors, including the complexity of the target sequence, temperature and the source of primer. A primer typically contains about 15 to about 25 nucleotides, but shorter and longer primers may also be used. Shorter primers generally require cooler temperatures to form stable complexes with the template.
A “phage display library” is a protein expression library that expresses a collection of cloned protein sequences as fusions with a phage coat protein. Thus, the phrase “phage display library” refers herein to a collection of phage (e.g., filamentous phage) wherein the phage express an external (typically heterologous) protein. The external protein is free to interact with (bind to) other moieties with which the phage are contacted. Each phage displaying an external protein is a “member” of the phage display library.
An “antibody phage display library” refers to a phage display library that displays antibodies or antibody fragments. The antibody library includes the population of phage or a collection of vectors encoding such a population of phage, or cell(s) harboring such a collection of phage or vectors. The library can be monovalent, displaying on average one single-chain antibody or antibody fragment per phage particle or multi-valent displaying, on average, two or more antibodies or antibody fragments per viral particle. The term “antibody fragment” includes, without limitation, single-chain Fv (scFv) fragments and Fab fragments. Preferred antibody libraries comprise on average more than 106, or more than 107, or more than 108, or more than 109 different members.
The term “filamentous phage” refers to a viral particle capable of displaying a heterogenous polypeptide on its surface, and includes, without limitation, fl, fd, Pfl and M13. The filamentous phage may contain a selectable marker such as tetracycline (e.g., “fd-tet”). Various filamentous phage display systems are well known to those of skill in the art (see, e.g., Zacher et al. Gene 9: 127-140 (1980), Smith et al. Science 228: 1315-1317 (1985); and Parmley and Smith Gene 73: 305-318 (1988)).
The term “panning” is used to refer to the multiple rounds of screening process in identification and isolation of phages carrying compounds, such as antibodies, with high affinity and specificity to a target.
B. Detailed Description
Techniques for performing the methods of the present invention are well known in the art and described in standard laboratory textbooks, including, for example, Ausubel et al., Current Protocols of Molecular Biology, John Wiley and Sons (1997); Molecular Cloning: A Laboratory Manual, Third Edition, J. Sambrook and D. W. Russell, eds., Cold Spring Harbor, N.Y., USA, Cold Spring Harbor Laboratory Press, 2001; O'Brian et al., Antibody Phage Display, Methods and Protocols, Humana Press, 2001; Phage Display: A Laboratory Manual, C. F. Barbas III et al. eds., Cold Spring Harbor, N.Y., USA, Cold Spring Harbor Laboratory Press, 2001; and Antibodies, G. Subramanian, ed., Kluwer Academic, 2004. Mutagenesis can, for example, be performed using site-directed mutagenesis (Kunkel et al., Proc. Natl. Acad. Sci USA 82:488-492 (1985)). PCR amplification methods are described in U.S. Pat. Nos. 4,683,192, 4,683,202, 4,800,159, and 4,965,188, and in several textbooks including “PCR Technology: Principles and Applications for DNA Amplification”, H. Erlich, ed., Stockton Press, New York (1989); and “PCR Protocols: A Guide to Methods and Applications”, Innis et al., eds., Academic Press, San Diego, Calif. (1990).
Information concerning antibody sequence analysis using the Kabat database and Kabat conventions may be found, e.g., in Johnson et al., The Kabat database and a bioinformatics example, Methods Mol. Biol. 2004; 248:11-25; and Johnson et al., Preferred CDRH3 lengths for antibodies with defined specificities, Int Immunol. 1998, December; 10(12):1801-5.
Information regarding antibody sequence analysis using Chothia conventions may be found, e.g., in Chothia et al., Structural determinants in the sequences of immunoglobulin variable domain, J Mol Biol. 1998 May 1; 278(2):457-79; Morea et al., Antibody structure, prediction and redesign, Biophys Chem. 1997; 68(1-3):9-16; Morea et al., Conformations of the third hypervariable region in the VH domain of immunoglobulins; J Mol Biol. 1998, 275(2):269-94; Al-Lazikani et al., Standard conformations for the canonical structures of immunoglobulins, J Mol Biol. 1997, 273(4):927-48. Barre et al., Structural conservation of hypervariable regions in immunoglobulins evolution, Nat Struct Biol. 1994, 1(12):915-20; Chothia et al., Structural repertoire of the human VH segments, J Mol Biol. 1992, 227(3):799-817 Conformations of immunoglobulin hypervariable regions, Nature. 1989, 342(6252):877-83; and Chothia et al., Review Canonical structures for the hypervariable regions of immunoglobulins, J Mol Biol. 1987, 196(4):901-17).
1. In SilicoDesign of Diverse (Poly)Peptide Libraries
According to the present invention, the design of diverse (poly)peptide libraries starts with the use of a database of related (poly)peptide sequences of interest, and, typically, the identification of sequence motifs that are shared by individual members of the library. Various computer programs for identifying sequence motifs in polypeptides are well known in the art and can be used on-line. Thus, for example, sequence motifs can be identified using the ELPH (a general-purpose Gibbs sampler for finding motifs in a set of DNA or protein sequences), MEME (Multiple EM for Motif Elicitation system that allows one to discover motifs of highly conserved regions in groups of related DNA or protein sequences); PPSEARCH (allows to search sequences for motifs or functional patterns in the PROSITE database (EBI)); emotif (a research system that forms motifs for subsets of aligned sequences, and ranks the motifs that it finds by both their specificity and the number of supplied sequences that it covers(Stanford Bioinformatics Group)); and the like.
In the next step, one or more sequence motifs identified are aligned to each other, and subdivided into separate datasets, each dataset being characterized by sharing a predetermined combination of parameters characteristic of one or more of the aligned sequence motifs. Such parameter can, for example, be the length, the subfamily in which a particular sequence motif belongs, the species from which the sequence derives, biological function, etc. The datasets characterized by a given combination of two or more parameters are then analyzed by position for amino acid frequency usage to identify key amino acid usage in individual stretches of amino acids within the datasets.
Alignment of the sequence motifs can be achieved in various ways that are within the skill in the art, for instance, using publicly available computer software such as BLAST, BLAST-2, ALIGN or Megalign (DNASTAR) software. Those skilled in the art can determine appropriate parameters for measuring alignment, including any algorithms needed to achieve maximal alignment over the full length of the sequences being compared.
Determination of amino acid frequency usage can be based on the appearance of high degree (typically at least 50%), and preferably complete, identity in all members of the dataset in a given position (conserved amino acid residue), or appearance of an amino acid residue in two or more members (preferably the majority) of the dataset for a given residue position. Additional datasets characterized by one or more additional parameters can then be created, not all of which need to be sequence related.
For example, if the goal is to design a diverse antibody library, antibody heavy and light chain CDR sequences present in the Kabat database, an electronic database containing non-redundant rearranged antibody sequences, can be analyzed for positional frequencies with unique combinations (filters) of predetermined parameters. The Kabat database contains antibody protein sequences, which are annotated upon submission. Information from the Kabat database can be imported into other environments, such as, for example, a Microsoft Access database, which allows for convenient application of filters, and the results can be tabulated and further analyzed in using any other software, including, for example, Excel.
The approach of the present invention allows for simultaneous filtering of both antibody heavy and light chain sequences, using a wide array of parameters (filters) and combination of parameters (filters). Thus, the generation of diversity datasets for particular heavy chains can be linked to light chain restrictions of choice. For example, filters for the analysis of antibody heavy chain CDR sequences may include one or more of (1) pairing with a certain light chain type (e.g. kappa (κ) or lambda (λ)); (2) CDR size. (e.g. CDR1=6 residues; CDR2=13 residues); and (3) CDR3 subfamily (e.g. VH1 vs. VH3). In the light chains, all CDRs may be size matched. For example, it can be pre-determined that CDR1=7, CDR2=10, and CDR3=8 amino acid residues. In addition, or in the alternative, the light chains can be filtered (sub-divided) based on the type of the light chain subfamily (e.g. κ1 or κ3 subfamily).
Thus, for example, heavy chain diversity analysis can be performed based upon pairing with x light chains, but the analysis may also be further restricted to those heavy chain sequences that pair with Vκ3 subfamily light chains, or to κ light chains bearing a CDR3 containing a length of 8 amino acids, or a combination of both.
Additional filters for covariant analysis of antibody heavy and/or light chains may include, without limitation, isotype, antigen type, affinity, and/or positional residues not related to CDR or the type or subtype of antibody chain.
In addition, the present invention enables the design of themed libraries, based upon “productive” heavy and light chain pairings. Thus various antibodies to the same antigen, including commercial antibodies, can be subjected to diversity analysis to identify antibodies that are most likely to succeed in human therapy.
If the goal is the design of themed antibody libraries, based upon productive heavy and light chain pairings, one or more productive (e.g. commercial) antibodies are selected to a selected antigen. Next, the germline origins to both the heavy and light chains are determined, and the heavy and light chain CDR sequences of the same type (e.g. VH3, V78 1) are subjected to the type of multivariate analysis described above to create diversity datasets. Preferably, the analysis should be based only on size-matched CDRs.
In the methods of the present invention, alignment and application of filters are followed by positional analysis in order to determine the positional frequency of individual amino acids or groups of amino acids within the previously created datasets, and to generate diversity datasets, such as CDR diversity datasets. After determining the absolute positional amino acid usage for each amino acid position of interest, the thresholds for percentage usage and for sum usage of amino acids can be lowered, in order to accommodate greater coverage of diversity. Thus, for example, the required total coverage may be set to more than 80%, with no individual amino acid being represented below 10%.
The in silico modeling can be continually updated with additional modeling information, from any relevant source, e.g., from gene and protein sequence and three-dimensional databases and/or results from previously tested polypeptides, such as antibodies, so that the in silico database becomes more precise in its predictive ability.
In addition, the in silico subsets can be supplemented with results of biological assays, such as binding affinity/avidity results, biological activity of previously tested antibodies. In this way, structural features can be more closely correlated with expected performance for an intended use.
Design of CDR diversity datasets is followed by the synthesis of a collection of combinatorial (degenerate) oligonucleotide sequences providing the required diversity, and cloning of the collection on the background of a suitable template.
2. Construction of Diverse (Poly)Peptide Libraries
After the creation of combinatorial positional diversity data sets as described above, physical combinatorial diversity sets can be generated by multisyntheses oligonucleotide synthesis. According to the present invention, instead of using a mutagenic code or mixed codon trimers, discrete degenerate oligonucleotide collections are generated that can be quantitatively restricted or relaxed to physically represent the combinatorial diversity sets produced through the foregoing analysis and design. Relaxing the criteria enables capture of the desired diversity through synthesis of fewer oligonucleotide probes, or to rationally expand the diversity set if the ability to clone the collection exceeds the predicted collection generated through diversity analysis. In addition, the physical combinatorial diversity sets can include side-products not found in the virtual diversity sets, with or without additional rule sets. This approach is most helpful in the field of combinatorial antibody library generation, but can also be rationally extended into other appropriate applications, such as to generating libraries of various polypeptide classes (e.g. growth factor libraries), etc. It is important to note that the physical library does not necessarily need to contain members comprising all amino acids at any given position that were identified by setting the threshold percentage usage as described above. For various reasons, for example in order to reduce the number of oligonucleotides needed, it may be advantageous to omit certain amino acid(s) at a given position. Alternatively or in addition, it is possible to increase coverage and diversity of the library by synthesizing members with amino acid residue(s) at a given position that did not meet the pre-set threshold frequency usage. The two approaches can be combined, i.e. certain amino acid residues present in the in silico diversity data sets may be omitted while others, not represented in the in silico diversity data sets at a given position may be added.
The first step in the creation of peptide or polypeptide libraries herein is the reverse translation of a collection of amino acids for multiplexed synthesis to contain an entire positional collection. Reverse translation tools are well known in the art and are commercially available. For example, the Java-based backtranslation tool of Entelechon (DE) translated proteins into nucleotides sequences with adapted codon usage, and allows optimization of a sequence for expression in specific organisms. In a preferred embodiment, the methods of the present invention employ an automated reverse translation algorithm capable of synthesizing discrete and degenerate sets of oligonucleotides to represent the diversity tables created by in silico analysis. This algorithm can include or exclude particular codons and even incorporate non-equimolar degeneracies to more accurately achieve not only the diversity of the dataset but also the relative distributions.
The number of oligonucleotides needed can be restricted by selecting degenerate bases to simultaneously encoding more than one of the frequently used amino acids at a time. In addition, such degenerate bases can be restricted to avoid rare codons of the species of interest. For example, if the collection is synthesized in E. coli, the use of rare arginine codon usage for E. coli can be restricted in the reverse translation. In addition, it is known that not all amino acids are used with the same frequency. Therefore, non-equimolar mixes can be used to more accurately reflect the profile of the virtual (in silico) diversity tables.
Where positional diversity requires the synthesis of more oligonucleotides than desired, diversity can be arbitrarily defined with a chemical probe collection. Thus, amino acid side-chain chemistries can be captured within subsets of amino acids, such as small, hydrophobic, aromatic, basic, acidic, amide, nucleophilic, etc. amino acids can constitute such subsets. As the Examples will illustrate, such chemically probed diversity positions can be synthesized by using a much smaller number of oligonucleotides than would otherwise be required. Chemically probed diversity covers much of the naturally occurring diversity, and provides broad interactive chemistries.
When constructing the diverse antibody libraries of the present invention, modified amino acid residues, for example, residues outside the traditional 20 amino acids used in most polypeptides, e.g., homocysteine, can be incorporated into the antibody sequences, such as CDRs, as desired. This can be carried out using art recognized techniques which typically introduce stop codons into the polynucleotide where the modified amino acid residue is desired. The technique then provides a modified tRNA linked to the modified amino acid to be incorporated (a so-called suppressor tRNA of, e.g., the stop codon amber, opal, or ochre) into the polypeptide (see, e.g., Köhrer et al., PNAS, 98, 14310-14315 (2001)).
In a preferred embodiment, one or more of the above steps are computer-assisted. In a particular embodiment, the computer assisted step comprises, e.g., mining the Kabat database and, optionally, cross-referencing the results against the Vbase sequence directory (Tomlinson, I M. et al., .
The oligonucleotides for generation of the libraries herein can be synthesized by known methods for DNA synthesis. Known synthesis methods include the phosphoramidite chemistry (Beaucage and Caruthers, Tetrahedron Letts., 22(20):1859 1862 (1981)), which permits effective oligo preparation, especially in the most common 40 80 bp size range, using an automated synthesizer, as described, for example, in Needham-VanDevanter et al. Nucleic Acids Res., 12:6159 6168 (1984)). In addition, oligonucleotides can be synthesized by the triester, phosphite, and H-phosphonate methods, all well known in the art. For a review of the oligonucleotide synthesis methods see, for example, “Oligonucleotide Synthesis: A Practical Approach”, ed. M. J. Gait, JRL Press, New York, N.Y. (1990). Oligonucleotides can also be custom ordered from a variety of commercial sources, such as, for example, The Midland Certified Reagent Company (Midland, Tex.), The Great American Gene Company (Salt Lake City, Utah), ExpressGen Inc. (Chicago, Ill.), Operon Technologies Inc. (Alameda, Calif.).
If the library is an antibody library, in the next step diversity is cloned into frameworks to produce a diverse antibody library.
The framework scaffold can be selected by methods well known in the art. Thus, the most frequently used frameworks in the database can be chosen for use as a scaffold, and diversity is cloned into the germline frameworks. For framework sequence selection, a subset of all available framework scaffolds determined to have been expressed in response to a particular antigen are arrayed. By determining the frameworks that are most frequently expressed in nature in response to a given antigen class an appropriate framework acceptor is selected. For example, to determine the preferred acceptor frameworks expressed in response to protein-based antigens, the Kabat database is searched for “protein-directed” frameworks. If preferred acceptor sequences are needed for presenting CDRs against a different antigen class, and/or, acceptor sequences of a particular species, the Kabat protein sequence filter is set accordingly. Thus, to determine sequences for use as human therapeutics against protein-based targets, the filter is set to focus only on human antibody sequences that recognize protein/peptide antigens. This greatly reduces redundancy in the dataset and sequence information that would bias results. Such analysis can be performed for VH, Vκ and/or Vλ, genes in a similar manner.
The diverse collections can be incorporated on an acceptor that is target specific to generate variant collections for antibody engineering.
The CDR diversity generated can be incorporated into framework regions by methods known in the art, such as polymerase chain reaction (PCR). For example, the oligonucleotides can be used as primers for extension. In this approach, oligonucleotides encoding the mutagenic cassettes corresponding to the defined region (or portion thereof), such as a CDR, are complementary to each other, at least in part, and can be extended to form a large gene cassette (e.g., a scFv) using a polymerase, e.g., a Taq polymerase.
In another approach, partially overlapping oligonucleotides are designed. The interal oligonucleotides are annealed to their complementary strand to yield a double-stranded DNA molecule with single-stranded extensions useful for further annealing. The annealed pairs can then be mixed together, extended, and ligated to form full-length double-stranded molecules using PCR. Convenient restriction sites can be designed near the ends of the synthetic gene for cloning into a suitable vector. In this approach, degenerate nucleotides can also be directly incorporated in place of one of the oligonucleotides. The complementary strand is synthesized during the primer extension reaction from a partially complementary oligonucleotide from the other strand by enzymatic extension with the aid of a polymerase. Incorporation of the degenerate polynucleotides at the stage of synthesis simplifies cloning, for example, where more than one domain or defined region of a gene is mutagenized or engineered to have diversity.
Regardless of the method used, after conversion into double stranded form, the oligonucleotides can be ligated into a suitable expression vector by standard techniques. By means of an appropriate vector, such as a suitable plasmid, the genes can be introduced into a cell-free extract, or prokaryotic cell or eukaryotic cell suitable for expression of the antibodies.
In a different approach, the desired coding sequence can be cloned into a phage vector or a vector with a filamentous phage origin of replication that allows propagation of single-stranded molecules with the use of a helper phage. The single-stranded template can be annealed with a set of degenerate oligonucleotides representing the desired mutations, elongated and ligated, thus incorporating each analog strand into a population of molecules that can be introduced into an appropriate host (see, e.g., Sayers, J. R. et al., Nucleic Acids Res. 16: 791-802 (1988)).
Various phagemid cloning systems suitable for producing the libraries, such as synthetic human antibody libraries, herein are known in the art, and have been described, for example, by Kang et al., Proc. Natl. Acad. Sci., USA, 88:4363 4366 (1991); Barbas et al., Proc. Natl. Acad. Sci. USA, 88:7978 7982 (1991); Zebedee et al., Proc. Natl. Acad. Sci., USA, 89:3175 3179 (1992); Kang et al., Proc. Natl. Acad. Sci., USA, 88:11120 11123 (1991); Barbas et al., Proc. Natl. Acad. Sci., USA, 89:4457 4461 (1992); and Gram et al., Proc. Natl. Acad. Sci., USA, 89:3576 3580 (1992).
The size of the library will vary depending upon the CDR length and the amount of CDR diversity which needs to be represented. Preferably, the library will be designed to contain less than 1015, 1014, 1013, 1012, 1011, 1010, 109, 108, 107, and more preferably, 106 or less antibodies or antibody fragments.
The libraries constructed in accordance with the present invention may be also attached to a solid support, such as a microchip, and preferably arrayed, using art recognized techniques.
The libraries constructed in accordance with the present invention can be expressed using any methods known in the art, including, without limitation, bacterial expression systems, mammalian expression systems, and in vitro ribosomal display systems.
In a preferred embodiment, the present invention encompasses the use of phage vectors to express the diverse libraries herein. The method generally involves the use of a filamentous phage (phagemid) surface expression vector system for cloning and expression. See, e.g., Kang et al., Proc. Natl. Acad. Sci., USA, 88:4363-4366 (1991); Barbas et al., Proc. Natl. Acad. Sci., USA, 88:7978-7982 (1991); Zebedee et al., Proc. Natl. Acad. Sci., USA, 89:3175-3179 (1992); Kang et al., Proc. Natl. Acad. Sci., USA, 88:11120-11123 (1991); Barbas et al., Proc. Natl. Acad. Sci., USA, 89:4457-4461 (1992); Gram et al., Proc. Natl. Acad. Sci., USA, 89:3576-3580 (1992); Brinkman et al., J. Immunol. Methods 182:41-50 (1995); Ames et al., J. Immunol. Methods 184:177-186 (1995); Kettleborough et al., Eur. J. Immunol. 24:952-958 (1994); Persic et al., Gene 187 9-18 (1997); Burton et al., Advances in Immunology 57:191-280 (1994); and U.S. Pat. Nos. 5,698,426; 5,233,409; 5,580,717; 5,427,908; 5,750,753; 5,821,047; 5,403,484; 5,571,698; 5,516,637; 5,780,225; 5,658,727; 5,733,743; 5,837,500; 5,969,108; 6,326,155; 5,885,793; 6,521,404; 6,492,160, 6,492,123; 6,489,123; 6,342,588; 6,291,650; 6,225,447; 6,180,336; 6,172,197; 6,140,471; 5,994,519; 6,969,108; 5,871,907; and 5,858,657.
The vector is used to transform a recombinant host cell, which is cultured to allow the introduced phage genes and display protein genes to be expressed, and for phage particles to be assembled and shed from the host cell. The shed phage particles are then harvested (collected) from the host cell culture media and screened for desirable antibody binding properties. Typically, the harvested particles are “panned” for binding with a preselected antigen. The strongly binding particles are collected, and individual species of particles are clonally isolated and further screened for binding to the antigen. Phages which produce a binding site of desired antigen binding specificity are selected.
It is emphasized that the methods of the present invention are not limited by any particular technology used for the expression and display of antibody libraries. Other display techniques, such as ribosome or mRNA display (Mattheakis et al., Proc. Natl. Acad. Sci. USA 91:9022-9026 (1994); Hanes and Pluckthun, Proc. Natl. Acad. Sci. USA 94:4937-4942 (1997)), microbial cell display, such as bacterial display (Georgiou et al., Nature Biotech. 15:29-34 (1997)), or yeast cell display (Kieke et al., Protein Eng. 10:1303-1310 (1997)), display on mammalian cells, spore display, viral display, such as retroviral display (Urban et al., Nucleic Acids Res. 33:e35 (2005), display based on protein-DNA linkage (Odegrip et al., Proc. Acad. Natl. Sci. USA 101:2806-2810 (2004); Reiersen et al., Nucleic Acids Res. 33:e10 (2005)), and microbead display (Sepp et al., FEBS Lett. 532:455-458 (2002)) are also suitable.
In ribosome display, the antibody and the encoding mRNA are linked by the ribosome, which at the end of translating the mRNA is made to stop without releasing the polypeptide. Selection is based on the ternary complex as a whole.
In a mRNA display library, a covalent bond between an antibody and the encoding mRNA is established via puromycin, used as an adaptor molecule (Wilson et al., Proc. Natl. Acad. Sci. USA 98:3750-3755 (2001)). For use of this technique to display antibodies, see, e.g., Lipovsek and Pluckthun, J. Immunol. Methods. 290:51-67 (2004).
Microbial cell display techniques include surface display on a yeast, such as Saccharomyces cerevisiae (Boder and Wittrup, Nat. Biotechnol. 15:553-557 (1997)). Thus, for example, antibodies can be displayed on the surface of S. cerevisiae via fusion to the α-agglutinin yeast adhesion receptor, which is located on the yeast cell wall. This method provides the possibility of selecting repertoires by flow cytometry. By staining the cells by fluorescently labeled antigen and an anti-epitope tag reagent, the yeast cells can be sorted according to the level of antigen binding and antibody expression on the cell surface. Yeast display platforms can also be combined with phage (see, e.g., Van den Beucken et al., FEBS Lett. 546:288-294 (2003)).
For a review of techniques for selecting and screening antibody libraries see, e.g., Hoogenboom, Nature Biotechnol. 23(9):1105-1116 (2005).
The invention will be illustrated by the following, non-limiting Examples.
EXAMPLESTechniques for performing the methods of the present invention are well known in the art and described in standard laboratory textbooks, including, for example, Ausubel et al., Current Protocols of Molecular Biology, John Wiley and Sons (1997); Molecular Cloning: A Laboratory Manual, Third Edition, J. Sambrook and D. W. Russell, eds., Cold Spring Harbor, N.Y., USA, Cold Spring Harbor Laboratory Press, 2001; O'Brian et al., Antibody Phage Display, Methods and Protocols, Humana Press, 2001; Phage Display: A Laboratory Manual, C. F. Barbas III et al. eds., Cold Spring Harbor, N.Y., USA, Cold Spring Harbor Laboratory Press, 2001; and Antibodies, G. Subramanian, ed., Kluwer Academic, 2004. Mutagenesis can, fore example, be performed using site-directed mutagenesis (Kunkel et al., Proc. Natl. Acad. Sci USA 82:488-492 (1985)); DNA Cloning, Vols. 1 and 2, (D. N. Glover, Ed. 1985); Oligonucleotide Synthesis (M. J. Gait, Ed. 1984); PCR Handbook Current Protocols in Nucleic Acid Chemistry, Beaucage, Ed. John Wiley & Sons (1999) (Editor); Oxford Handbook of Nucleic Acid Structure, Neidle, Ed., Oxford Univ Press (1999); PCR Protocols: A Guide to Methods and Applications, Innis et al., Academic Press (1990); PCR Essential Techniques: Essential Techniques, Burke, Ed., John Wiley & Son Ltd (1996); The PCR Technique: RT-PCR, Siebert, Ed., Eaton Pub. Co. (1998); Antibody Engineering Protocols (Methods in Molecular Biology), 510, Paul, S., Humana Pr (1996); Antibody Engineering: A Practical Approach (Practical Approach Series, 169), McCafferty, Ed., Irl Pr (1996); Antibodies: A Laboratory Manual, Harlow et al.; C.S.H.L. Press, Pub. (1999); Large-Scale Mammalian Cell Culture Technology, Lubiniecki, A., Ed., Marcel Dekker, Pub., (1990). Border et al., Yeast surface display for screening combinatorial polypeptide libraries, Nature Biotechnology, 15(6):553-7 (1997); Border et al., Yeast surface display for directed evolution of protein expression, affinity, and stability, Methods Enzymol., 328:430-44 (2000); ribosome display as described by Pluckthun et al. in U.S. Pat. No. 6,348,315, and Profusion™ as described by Szostak et al. in U.S. Pat. Nos. 6,258,558; 6,261,804; and 6,214,553;, and bacterial periplasmic expression as described in US20040058403A1.
Further details regarding antibody sequence analysis using Kabat conventions may be found, e.g., in Johnson et al., The Kabat database and a bioinformatics example, Methods Mol Biol. 2004; 248:11-25; Johnson et al., Preferred CDRH3 lengths for antibodies with defined specificities, Int Immunol. 1998, December; 10(12):1801-5; Johnson et al., SEQHUNT. A program to screen aligned nucleotide and amino acid sequences, Methods Mol Biol. 1995; 51:1-15. and Wu et al., Length distribution of CDRH3 in antibodies; and Johnson et al., Proteins. 1993 May; 16(1):1-7. Review).
Further details regarding antibody sequence analysis using Chothia conventions may be found, e.g., in Chothia et al., Structural determinants in the sequences of immunoglobulin variable domain, J Mol Biol. 1998 May 1; 278(2):457-79; Morea et al., Antibody structure, prediction and redesign, Biophys Chem. 1997 October; 68(1-3):9-16; Morea et al., Conformations of the third hypervariable region in the VH domain of immunoglobulins; J Mol Biol. 1998 Jan. 16; 275(2):269-94; Al-Lazikani et al., Standard conformations for the canonical structures of immunoglobulins, J Mol Biol. 1997 Nov. 7; 273(4):927-48. Barre et al., Structural conservation of hypervariable regions in immunoglobulins evolution, Nat Struct Biol. 1994 December; 1(12):915-20; Chothia et al., Structural repertoire of the human VH segments, J Mol Biol. 1992 Oct. 5; 227(3):799-817 Conformations of immunoglobulin hypervariable regions, Nature. 1989 Dec. 21-28; 342(6252):877-83; and Chothia et al., Review Canonical structures for the hypervariable regions of immunoglobulins, J Mol Biol. 1987 Aug. 20; 196(4):901-17).
Further details regarding Chothia analysis are described, for example, in Morea V, Tramontano A, Rustici M, Chothia C, Lesk A M. Conformations of the third hypervariable region in the VH domain of immunoglobulins. J Mol Biol. 1998 Jan. 16; 275(2):269-94; Chothia C, Lesk A M, Gherardi E, Tomlinson I M, Walter G, Marks J D, Llewelyn M B, Winter G. Structural repertoire of the human VH segments. J Mol Biol. 1992 Oct. 5; 227(3):799-817; Chothia C, Lesk A M, Tramontano A, Levitt M, Smith-Gill S J, Air G, Sheriff S, Padlan E A, Davies D, Tulip W R, et al. Conformations of immunoglobulin hypervariable regions. Nature. 1989 Dec. 21-28; 342(6252):877-83; Chothia C, Lesk A M. Canonical structures for the hypervariable regions of immunoglobulins. J Mol Biol. 1987 Aug. 20; 196(4):901-17; and Chothia C, Lesk A M. The evolution of protein structures. Cold Spring Harb Symp Quant Biol. 1987; 52:399-405.
Further details regarding CDR contact considerations are described, for example, in MacCallum R M, Martin A C, Thornton J M. Antibody-antigen interactions: contact analysis and binding site Topography. J Mol Biol. 1996 Oct. 11; 262(5):732-45.
Further details regarding the antibody sequences and databases referred to herein are found, e.g., in Tomlinson I M, Walter G, Marks J D, Llewelyn M B, Winter G. The repertoire of human germline VH sequences reveals about fifty groups of VH segments with different hypervariable loops. J Mol Biol. 1992 Oct. 5; 227(3):776-98; Li W, Jaroszewski L, Godzik A. Clustering of highly homologous sequences to reduce the size of large protein databases. Bioinformatics. 2001 March; 17(3):282-3; [VBDB] www.mrc-cpe.cam.ac.ukkbase-ok.php?menu=901; [KBTDB] database.com; [BLST] www.ncbi.nlm.nih.gov/BLAST/[CDHIT] bioinformatics.ljcrf.edu/cd-hi/; [EMBOSS] www.hgmp.mrc.ac.uk/Software/EMBOSS/; [PHYLIP] evolution.genetics.washington.edu/phylip.html; and [FASTA] fasta.bioch.virginia.edu.
Example 1Frequency Analysis of Antibody Light Chain Vκ1 CDR1, 2, and 3 Sequences
In a first step, 2374 human antibody Vκ1 light chain variable domain sequences were collected from the Kabat Database of Sequences of Proteins of Immunological Interest. For each sequence, the gene sequence was translated into the corresponding amino acid sequence, and the amino acid sequences were positionally aligned, following the Kabat numbering system.
Next, the collection of Vκ1 light chain sequences obtained was filtered by selecting sequences having amino acids “RV” at positions and 18-19, and applying the following length restrictions: CDR1=7 amino acids, CDR2=10 amino acids, and CDR3=8 amino acids. By applying these filters, the originally 2374-member collection was reduced to 771 members.
By using only entries containing complete unambiguous sequences from the “RV” motif preceding DR1 through a complete CDR3 sequence, the number of Vκ1 light chain variable domain sequences was further reduced to 383.
Subsequently, the sequences were aligned, the occurrence amino acids at each position was tabulated, and the distribution of the 20 naturally occurring amino acids at each position was calculated to produce the positional frequency-based database of CDR domain diversity based on absolute usage of amino acids by position. The results of this tabulation are shown in
The datasets set forth in
As shown in
Design of VH3 Heavy Chain Synthetic Library Diversity
Analyzing VH3 heavy chain polypeptide sequences, 10 amino acids in length, obtained from the Kabat database of antibody sequences essentially as described in Example 1, the data shown in
Making a Semi-Synthetic Antibody Library
As previously described, the analysis and generation of the VH CDR diversity can be tuned to contextually reflect compositions for productive and specific pairings with κ or λ light chains, (i.e. pairings resulting in antibodies specifically binding a target antigen). These synthetic VH repertoires need not exclusively be paired with synthetic light chain repertoires, but can be combinatorially cloned with collections of lymphocyte derived light chains. In practice, a collection of κ and λ light chains is separately cloned into a phage display vector followed by cloning of either the individual heavy chain variable region frameworks for subsequent introduction of diversity or a collection of pre-diversified variable region frameworks. In either case, the light chain compatibly paired heavy chain variable regions are expected to more productively pair with the corresponding light chains.
Example 4Engineering Improved Antibodies by Creating Libraries of Variants Upon a Base Clone.
In a matter analogous to introducing productive diversity upon a germline acceptor frameworks for creating de novo immunoglobulin repertoires, target specific mutagenesis libraries are created for a specific antibody or defined collection of antibodies. Such libraries are useful in the task of antibody engineering especially in the field of affinity maturation. Starting with a monoclonal antibody of interest, the defining characteristics are determined, which are captured in the previously defined diversity influencing elements of the present invention, such as, germline framework origin, light chain type, and light chain and heavy chain CDR lengths. After determining these or similar characteristics the next step is to refer to database sequences that correspond these parameters. Subsequent to identifying the corresponding sets of sequences an analysis, similar to that described earlier, is performed to examine subset repertoire diversity and then generate the corresponding multi-degenerate oligonucleotides necessary to encode the desired diversity. These multi-degenerate oligonucleotides are then cloned as single or combinatorial CDR collections. As it is more likely to find synergistic improvements with multi-CDR mutagenesis the creation of combinatorial CDR mutagenesis libraries is preferred. Using the multi-degenerate oligonucleotides from the analysis described above rationally creates and re-diversifies antibodies according to positional diversity with respect to human bias and preference. It is important to note that in instances where any of the light chain CDR sequences or heavy chain CDR1 or CDR2 sequences diverge from the germline sequence the corresponding germline-encoding oligonucleotide is also included into the combinatorial CDR library. This inclusion of germline encoding oligonucleotides allows for backcrossing of germline sequences to create more productive CDR combinations.
This “diversity reincorporation scheme” is also useful in engineering a re-diversified set of antibodies from an existing synthetic antibody clone. As the potential diversity of the synthetic libraries created in accordance with the present invention exceeds the limits of currently available techniques to display and select all members, it is very likely that any discovered target specific clones represent only a fraction of the possible solutions present and accessible at the DNA-level of any of the typically screened libraries. Thus, after identifying in a library of the present invention, following four rounds of panning, an anti-EGF antibody, the originally designed diversity was combinatorially reintroduced into the clone to create a new set of variants. These new sets of variants were then re-screened by panning on EGF and in each successive round the stringency of binding and washing was increased. The net result created pools of EGF binding phage that enriched to greater levels over background than those found in the original panning.
Example 5Design of a Cytokine Theme Library
In order to create productive libraries for the discovery of new anti-cytokine antibodies, a productive anti-TNF-α antibody, HUMIRA® (adalimumab) was selected as a basic theme. HUMIRA® (adalimumab) is a recombinant human IgG1 monoclonal antibody created using phage display technology resulting in an antibody with human-derived heavy and light chain variable regions and human IgG1:κ constant regions.
To determine the germline origin of the heavy chain of parental antibody D2E7, the framework region was analyzed. This was accomplished by masking the CDRs of D2E7 and of the human germline VH genes. Next the remaining sequences between FR1 and FR3 for D2E7 was aligned by BLAST algorithm against all of the human germline VH genes. As shown in
The frequency analysis of antibody light chain Vκ1 CDR1, CDR2, and CDR3 sequences described in Example 1 was modified by setting the threshold percent usage filter to 6%. As shown in
Next, 5971 human antibody heavy chain variable domain sequences were collected from the Kabat Database of Sequences of Proteins of Immunological Interest. For each sequence, the gene sequence was translated into the corresponding amino acid sequence, and the amino acid sequences were positionally aligned, following the Kabat numbering system.
The heavy chain variable domain collection was then subjected to the following filters:
1. VH3 sequences containing “CAAS” at amino acid positions 22-25 (1530 of 5971 members);
2. sequences combined with κ light chains, CDR1=6 amino acids and CDR2=13 amino acids (226 of 1530 members)
3. Including only members containing complete sequences from “CAAS” preceding CDR1 through a complete CDR2 sequence (180 of 226 members).
Subsequently, the sequences were aligned, the occurrence amino acids at each position was tabulated, and the distribution of the 20 naturally occurring amino acids at each position was calculated to produce the positional frequency-based database of CDR domain diversity based on absolute usage of amino acids by position. The results of this tabulation are shown in
The datasets set forth in
Applying the 5% usage filter both CDR1 and CDR2 4 degenerate oligonucleotides required for the synthesis of the CDR1 regions, CDR2 diversity can be encoded by 28 degenerate oligonucleotides (see
In the next step, from the 5971 human antibody heavy chain variable domain sequences described above, sequences VH3 sequences 13 amino acids in length were compiled, regardless of isotype. The required percent amino acid usage for each position was set to 4%, except at amino acid positions 93, 94 and 101, where the threshold was set to 4% usage. The results are set forth in
Design of a Hapten Themed Antibody Library
The objective of this method is to design productive libraries for the identification of new anti-hapten antibodies.
The design started with an anti-digoxigenin (anti-DIG) antibody (Dorsam, H. et al., FEBS Lett. 414:7-13 (1997)). The Ig λ light chain variable region sequence (SEQ ID NO: 1) and the heavy chain variable region sequence (SEQ ID NO: 2) for this antibody are shown in
In order to determine the germline origin of the heavy and light chains of this parental antibody were analyzed. As shown in
Next, the light chain CDR1 and CDR2 sequences were analyzed as described in the previous examples for X length matched VL framework residues. The required percent amino acid usage for each position was set to 6%, so that no individual sequences were reported below 6%. As shown in
Cytokine (IFN-α) Analysis and Library Creation
IFN is a generic term for cytokines having anti-viral activity, among which those produced from leukocytes or lymphoblastic cells by stimulation with virus or double stranded nucleic acids are termed as IFN-α. IFN-═ has a variety of activities including anti-viral activity and cellular growth-suppressing activity, which activities have been found to be useful in the treatment of a variety of diseases such as hepatitis type B and type C infections, and cancer. Analysis of sequences of IFN-α genes cloned from a variety of DNA libraries has revealed that IFN-α exists in several subtypes. For example, for the IFN-α2 gene, three additional types (α2a, .α2b, and .α2c) have been identified. Altogether, there are over 20 currently known IFN-α subtypes. Additional known subtypes include, for example, IFN-α1a, IFN-α1b, IFN-α4a, IFN-α4b, IFN-α5, IFN-α6, etc. It has been demonstrated that many of the IFN-α subtypes differ in their biological activities and other biological properties. Therefore, libraries created based upon the existing natural diversity among members of the IFN-α family find utility in generating IFN-α polypeptides with new and improved properties, such as increased potency, decreased immunogenicity, increased half-life, improved proteolytic stability.
As a first step for creating a diverse IFN-α library, eleven 189 amino acids long gene products were identified. Amino acid residues 32-38 of these IFN-α polypeptides were aligned with each other and the residue frequency usage was determined, as shown in
Once the library is prepared, screening for desired novel properties can be performed by methods known in the art. Thus, increased potency can be tested in standard biological assays, such as by biopanning a phage-displayed IFN-α library. members with increased half-life can be identified, for example, by biopanning a phage-displayed library against an IFN-α receptor, or by exposing members of the library to one or more serum proteases. Decreased immunogenicity can be tested, for example, by identifying the peptides or polypeptides present in the library that show the least binding to MHC molecules, or by testing T cell epitope presentation of whole proteins directly.
These and numerous additional tests are well known to those of ordinary skill in the pertinent art.
Example 8Chemically Probed Antibody Collections
The present example shows the creation of CDR3 heavy chain diversity using probe sets designed based on chemical principles.
Amino acids can be divided into seven groups, characterized by small, nucleophilic, hydrophobic, aromatic, acidic, amide, and basic side-chain chemical functionalities, respectively (
The native heavy chain CDR3 sequence contains a high degree of chemical diversity (around 60% or more). It has been determined that a similar chemical diversity can be generated, by combinatorial denegerate oligonucleotide synthesis, using 128 degenerate oligonucleotides. The design of the corresponding degenerate oligonucleotides is shown in
This chemically probed diversity approach can be used on its own, or in combinations with any of the other methods of the present invention, in order to produce combinatorial libraries with desired properties.
Although in the foregoing description the invention is illustrated with reference to certain embodiments, it is not so limited. Indeed, various modifications of the invention in addition to those shown and described herein will become apparent to those skilled in the art from the foregoing description and fall within the scope of the appended claims. Thus, while the invention is illustrated with reference to antibody libraries, it extends generally to all peptide and polypeptide libraries.
All references cited throughout the specification are hereby expressly incorporated by reference.
Claims
1-18. (canceled)
19. A method of making a library of antibody heavy or light chains, comprising the steps of
- (a) aligning a plurality of antibody heavy or light chain sequences comprising one or more CDR sequence motifs;
- (b) creating a first dataset by applying a predetermined combination of two or more filters to said antibody heavy or light chain sequences;
- (c) analyzing said first dataset for positional amino acid usage frequency within at least one of said CDR sequence motifs;
- (d) creating a second dataset characterized by a minimum threshold amino acid usage frequency at one or more amino acid positions within said CDR sequence motif; and
- (e) synthesizing an antibody heavy or light chain library designed with the aid of the first and second datasets identified.
20. The method of claim 20 wherein said library is synthesized by generating a discrete number of defined or degenerate oligonucleotides such that only defined amino acids are generated.
21. The method of claim 20 wherein the diversity of the physical library produced exceeds the diversity of a library which is a physical representation of the datasets identified.
22. The method of claim 22 wherein at least one amino acid not meeting the minimum threshold amino acid usage frequency is also synthesized to provide said diversity.
23. The method of claim 20 wherein the diversity of the physical library produced is less than the diversity of a library which is a physical representation of the datasets identified.
24. The method of claim 24 wherein not all amino acids meeting the minimum threshold amino acid usage frequency are synthesized.
25. The method of claim 20 wherein said CDRs are cloned into a scaffold of framework sequences.
26. The method of claim 26 wherein said framework sequences are the most frequently used framework sequences in the database comprising said CDRs.
27. The method of claim 20 wherein said physical library is expressed using a prokaryotic or eukaryotic expression system.
28. The method of claim 20 wherein said physical library is expressed and displayed using a phagemid display, mRNA display, microbial cell display, mammalian cell display, microbead display technique, antibody array, or display based on protein-DNA linkage.
29. The method of claim 20 wherein said library is screened for one or more chemical and/or biological properties of its members:
30. The method of claim 30 wherein said biological property is selected from the group consisting of half-life, potency, efficacy, binding affinity, and immunogenicity.
31. The method of claim 20 comprising the introduction of amino acid side-chain diversity at one or more amino acid positions.
32. The method of claim 32 wherein said amino acid side-chain diversity is introduced by providing amino acid residues with at least two different side-chain chemical functionalities at said amino acid position or positions.
33. The method of claim 33 wherein at least 30% to at least 50% of all amino acid chemistries are represented at each amino acid position.
34. The method of claim 33 wherein said side-chain diversity is introduced by using combinatorial degenerate oligonucleotide synthesis.
35-41. (canceled)
Type: Application
Filed: Dec 9, 2011
Publication Date: Aug 9, 2012
Inventors: Lawrence Horowitz (Atherton, CA), Ramesh Bhastt (Belmont, CA), Aaron L. Kurtzman (San Carlos, CA)
Application Number: 13/374,065
International Classification: C40B 50/02 (20060101);
