CA2558811A1 - Dimeric fusion proteins and materials and methods for producing them - Google Patents

Abstract

Polypeptide fusions, dimeric fusion proteins, and materials and methods for making them are disclosed. One of the polypeptide fusions consists of a non-immunoglobulin polypeptide, a polypeptide linker, a dimerizing domain, and, optionally, a linking polypeptide. Another of the polypeptide fusions consists of a non-immunoglobulin polypeptide, a polypeptide linker, and a second dimerizing domain.

Description

DEMANDES OU BREVETS VOLUMINEUX
LA PRESENTE PARTIE I)E CETTE DEMANDE OU CE BREVETS
COMPRI~:ND PLUS D'UN TOME.
CECI EST ~.E TOME 1 DE 2 NOTE: Pour les tomes additionels, veillez contacter le Bureau Canadien des Brevets.
JUMBO APPLICATIONS / PATENTS
THIS SECTION OF THE APPLICATION / PATENT CONTAINS MORE
THAN ONE VOLUME.

NOTE: For additional vohxmes please contact the Canadian Patent Oi~ice.

Description DIIVVIERIC FUSION PROTEINS AND MATERIALS AND METHODS FOR
PRODUCING THEM
BACKGROUND OF THE INVENTION
A variety of proteins, including both cell-surface receptors and soluble ligands, require dimerization for proper biological activity. Although dimerization may readily occur in the native biological milieu, it is often problematic for recombinant forms of many proteins, including soluble forms of receptors that are embedded in the cell surface in their native state.
Cell-surface receptors that must dimerize in order to transduce a signal include members of the receptor tyrosine kinase family, the tumor necrosis factor (TNF) receptor family, and the class 1 and class 2 cytokine receptor families. Platelet derived growth factor '(PDGF) receptors, for example, dimerize upon ligand binding, resulting in autophosphorylation of tyrosine residues and initiation of intracellular signal transduction. Other examples of receptors that dimerize upon ligand binding include growth hormone receptors, interleukin-2 (IL-2) receptors, IL-3 receptors, IL-5 receptors, IL-6 receptors, granulocyte-macrophage colony stimulating factor (GM-CSF) receptors, oncostatin M (OSM) receptors, leukemia inhibitory factor (LIF) receptors, and ciliary neurotrophic factor (CNTF) receptors.
Dimerized receptors include both homodimers and heterodimers. Some receptor polypeptides function as subunits within a plurality of cytokine receptor dimers, See, for example, Cosman, Cytokine 5:95-106, 1993.
Naturally occuring soluble forms of many cell-surface receptors are known.
These soluble receptors correspond to the ligand-binding domains of their cell-surface counterparts. Soluble cytokine receptors inhibit cytokine responses and act as transport proteins.
See, for example, Aggarwal and Puri, "Common and Uncommon Features of Cytokines and Cytokine Receptors: An Overview," in Aggarwal and Puri, eds., Hunzazz Cytoki.ozes: Their- Role izz Disease and Therapy, Blackwell Science, 1995, 3-24. It has been found that dimerization of soluble receptor polypeptides through the use of fusion proteins may enhance the binding properties of these soluble receptors so that they become therapeutically useful antagonists of their cognate ligands. Typical of such dimeric fusions are immunoglobulin fusions. See, for example, Sledziewslci et al., U.S. Patents Nos. 5,155,027 and 5,567,584;
Jacobs et al., U.S.
Patent No. 5,605,690; Wanner et al., U.S. Patent No. 5,914,111; and Ashkenazi and Chamow, Curr. Opin. Iznfzzuuoi. 9:195-200, 1997.

To date, immunoglobulin fusion technology has not provided a commercially viable means to produce heterodimeric proteins. Using currently available technology, co-expression of two different fusion polypeptides in a recombinant cell generally results in a mixture of both homodimers and heterodimers. The costs associated with recovery and purification of heterodimers from the mixture has limited the commercial application of this technology. Thus, there remains a need in the art for an efficient method of producing soluble, dimeric proteins, including soluble receptor dimers and soluble heterodimers.
DESCRIPTION OF THE INVENTION
Within one aspect of the present invention there is provided a first polypeptide fusion consisting of, from amino terminus to carboxyl terminus, P1-L1-D1-(P2)n, wherein Pl is a non-immunoglobulin polypeptide; L1 is a polypeptide linker consisting of from 18 to 32 amino acid residues, wherein x of the residues are cysteine residues and x is an integer from 1 to 8; D1 is a dimerizing domain selected from the group consisting of an immunoglobulin CHl domain, a T-cell receptor Ca domain, a T-cell receptor C(3 domain, a major histocompatibility complex class I a3 domain, (32-microglobulin, a major histocompatibility complex class II cc2 domain, and a major histocompatibility complex class II (32 domain; P2 is a linking polypeptide consisting of from 1 to 29 amino acid residues, wherein at least one of the residues is a cysteine residue; and n is 0 or 1. Within one embodiment, P1 is an extracellular domain of a cell-surface receptor, such as a human receptor. Within another embodiment, Pl is not a member of the immunoglobulin superfamily. Within other embodiments, P1 is selected from the group consisting of 1L-17R, IL-20RA, IL-20RB, IL,-21R, IL,-28RA, IL.-31RA, CRF2-4, and ~yC. Within a further embodiment, L1 contains exactly two cysteine residues. Within another embodiment, L1 comprises an immunoglobulin hinge or fragment or variant thereof, such as an immunoglobulin hinge variant wherein the cysteine residue corresponding to Eu residue 220 is replaced by serine. Within another embodiment, L1 comprises a human ~yl hinge or fragment or variant thereof. Witlun an additional embodiment, L1 consists of 18 amino acid residues.
Within yet other embodiments, L1 comprises a plurality of glycine residues or L1 comprises a plurality of serine residues. Within related embodiments, L1 comprises [Gly-Ser-Gly-Ser]a (SEQ ID N0:2), wherein a is 1 or 2; or [Gly-Gly-Gly-Ser]b (SEQ ID N0:3), wherein b is 1 or 2.
Within an additional embodiment, L1 comprises a proteolytic cleavage site.
Within a further embodiment, D1 is an immunoglobulin CH1 domain, such as a human immunoglobulin CHl domain. Within yet other embodiments, the irnmunoglobulin CH1 domain is a 'y1 CHl domain, such as a human yl CH1 domain. Within another embodiment, P2 is a portion of an immunoglobulin hinge comprising a cysteine residue. Within a further embodiment, n is 1.
Within another embodiment, P2 consists of from 5 to 15 amino acid residues.
Within an additional embodiment, P2 contains exactly one cysteine residue.

Within a second aspect of the invention there is provided a second polypeptide fusion consisting of, from amino terminus to carboxyl terminus, P3-L2-D2, wherein P3 is a non-immunoglobulin polypeptide; L2 is a polypeptide linker consisting of from 18 to 32 amino acid residues, wherein y of the residues are cysteine residues and y is an integer from 1 to 8; and D2 is a dimerizing domain selected from the group consisting of an immunoglobulin light chain constant domain, a T-cell receptor Coc domain, a T-cell receptor C(3 domain, a major histocompatibility complex class I a3 domain, (32-microglobulin, a major histocompatibility complex class II ct2 domain, and a major histocompatibility complex class II
(32 domain. Within one embodiment of the invention, P3 is an extracellular domain of a cell-surface receptor, such as a human receptor. Within another embodiment, P3 is not a member of the immunoglobulin superfamily. Within other embodiments, P3 is selected from the group consisting of IL-17R, IL-20RA, IL.-20RB, IL-21R, IL-28RA, IL-31RA, CRF2-4, and ~yC. Within further embodiments, L2 contains exactly two cysteine residues, L2 comprises an immunoglobulin hinge or fragment or variant thereof, L2 comprises an immunoglobulin hinge variant wherein the cysteine residue corresponding to Eu residue 220 is replaced by serine, L2 comprises a human ~yl hinge or fragment or variant thereof, or L2 consists of 18 amino acid residues. Within other embodiments, L2 comprises a plurality of glycine residues or L2 comprises a plurality of serine residues. Within related embodiments, L2 comprises [Gly-Ser-Gly-Ser]a (SEQ ID
N0:2), wherein a is 1 or 2; or [Gly-Gly-Gly-Ser]b (SEQ ID NO:3), wherein b is 1 or 2.
Within a further embodiment, L2 comprises a proteolytic cleavage site. Within yet other embodiments, D2 is an immunoglobulin x light chain constant domain or immunoglobulin ~, light chain constant domain.
Within a third aspect of the invention there is provided a polynucleotide encoding the first polypeptide fusion disclosed above. Within one embodiment the polynucleotide is DNA.
Within a fourth aspect of the invention there is provided a polynucleotide encoding the second polypeptide fusion disclosed above. Within one embodiment the polynucleotide is DNA.
Within a fifth aspect of the invention there is provided an expression vector comprising the following operably linked elements: a transcription promoter; a DNA segment encoding the first polypeptide fusion disclosed above; and a transcription terminator.
Within a sixth aspect of the invention there is provided an expression vector comprising the following operably linked elements: a transcription promoter; a DNA segment encoding the second polypeptide fusion disclosed above; and a transcription terminator.
Within a seventh aspect of the invention there is provided a cultured cell into which has been introduced one or both of the expression vectors disclosed above, wherein the cell expresses the DNA segment(s).

Within an eighth aspect of the invention there is provided a dimeric protein consisting of a first polypeptide fusion as disclosed above disulfide bonded to a second polypeptide fusion as disclosed above, wherein each of x and y is an integer from 1 to 8 and x =
y. Within one embodiment, P1 and P3 are different. Within another embodiment, n=1. Within a further embodiment, x=2 and y=2. Within other embodiments, one of P1 and P3 is a zcytor7 extracellular domain and the other of P 1 and P3 is a DIRS 1 extracellular domain; one of P 1 and P3 is a zcytorll extracellular domain and the other of P1 and P3 is a DIRSl extracellular domain; one of P1 and P3 is a zalphal l extracellular domain and the other of P1 and P3 is an IL,-2 receptor y common extracellular domain; or one P1 and P3 is a PDGF a, receptor extracellular domain and the other of P1 and P3 is a PDGF (3 receptor extracellular domain.
Within a ninth aspect of the invention there is provided a method of making the dimeric protein disclosed above comprising the steps of (a) culturing a cell comprising first and second expression units, wherein the first expression unit comprises a transcription promoter, a first DNA segment encoding the first polypeptide fusion disclosed above, and a transcription terminator; and wherein the second expression unit comprises a transcription promoter, a second DNA segment encoding the second polypeptide fusion disclosed above, and a transcription terminator, whereby the first and second DNA segments are expressed and the encoded polypeptide fusions are produced as a dimeric protein; and (b) recovering the dimeric protein.
Within a tenth aspect of the invention there is provided a homodimeric protein consisting of two second polypeptide fusions as disclosed above, wherein D2 is an immunoglobulin light chain constant domain and wherein said fusions are joined to each other by at least one disulfide bond.
Within an eleventh aspect of the invention there is provided a method of making the homodimeric protein disclosed above comprising the steps of (a) culturing a cell comprising an expression vector comprising the following operably linked elements: a transcription promoter; a DNA segment encoding the second polypeptide fusion disclosed above wherein D2 is an immunoglobulin light chain constant domain; and a transcription terminator, whereby the DNA segment is expressed and the encoded polypeptide fusion is produced as a homodimeric protein; and (b) recovering the homodimeric protein.
These and other aspects of the invention are illustrated by the following detailed description and the attached drawings. Within the drawings:
Figs. lA-1B illustrate the amino acid sequence of a portion of a representative human immunoglobulin ~1 heavy chain (SEQ >D NO:1) (based on Ellison et al., Nucl. Acids Res.
10:4071, 1982). Amino acid sequence numbers are based on the Eu index (Edelman et al., Proc.
Natl. Acad. Sci. USA 63:78-85, 1969; Kabat et al., Sequences of Proteins of Imf~aunological Interest, US Department of Health and Human Services, NIH, Bethesda, MD, 1991). The Cys residues normally involved in disulfide bonding to the light chain constant region (LC) and heavy chain constant region (HC) are indicated. Boundaries of the CH1, hinge, CH2, and CH3 domains are shown.
Figs. 2A-2B show an alignment of representative human CH1 domain sequences and a consensus sequence derived from the alignment. Sequence gaps are indicated by "---".

5 Fig. 3 shows an alignment of representative human light chain constant domain sequences and a consensus sequence derived from the alignment. Sequence gaps are indicated by "-_-"_ As used herein, the phrase "a cultured cell into which has been introduced an expression vector" includes Bells that have been physically manipulated to contain the vector, as well as progeny of the manipulated cells when the progeny also contain the vector.
The terms "amino-terminal" (or "N-terminal") and "carboxyl-terminal" (or "C-terminal") are used herein to denote positions within polypeptides. Where the context allows, these terms are used with reference to a particular sequence or portion of a polypeptide to denote proximity or relative position. For example, a certain sequence positioned carboxyl-terminal to a reference sequence within a polypeptide is located proximal to the carboxyl terminus of the reference sequence, but is not necessarily at the carboxyl terminus of the complete polypeptide.
The term "corresponding to", when applied to positions of amino acid residues in sequences, means corresponding positions in a plurality of sequences when the sequences are optimally aligned.
The term "expression vector" is used to denote a DNA molecule, linear or circular, that comprises a segment encoding a polypeptide of interest operably linked to additional segments that provide for its transcription. Such additional segments include promoter and terminator sequences, and may also include one or more origins of replication, one or more selectable markers, an enhancer, a polyadenylation signal, etc.
Expression vectors are generally derived from plasmid or viral DNA, or may contain elements of both.
An "immunoglobulin" is a serum protein which functions as an antibody in a vertebrate organism. Five classes of "immunoglobulin," or antibody, protein (IgG, IgA, IgM, IgD, and IgE) have been identified in higher vertebrates. IgG comprises the major class; it normally exists as the second most abundant protein found in plasma. In humans, IgG consists of four subclasses, designated IgGl, IgG2, IgG3, and IgG4. The heavy chain constant regions of the IgG class are identified with the Greek symbol ~(. For example, immunoglobulins of the IgGl subclass contain a 'y1 heavy chain constant region. Each immunoglobulin heavy chain possesses a constant region that consists of constant region protein domains (CH1, hinge, CH2, and CH3) that are essentially invariant for a given subclass in a species. DNA
sequences encoding human and non-human immunoglobulin chains are known in the art. See, for example, Ellison -et al., DNA 1:11-18, 1981; Ellison et al., Nucleic Acids Res. 10:4071-4079, 1982; Kenten et al.; Proc. Natl. Acad. Sci. USA 79:6661-6665, 1982; Seno et al., Nuc. Acids Res. 11:719-726, 1983; Riechmann -et al., Nature 332:323-327, 1988; Amster et al., Nuc. Acids Res. 8:2055-2065, 1980; Rusconi and Kohler; Natuz-e 314:330-334, 1985; Boss et al., Nuc. Acids Res. 12:3791-3806, 1984; Bothwell et al., Nature 298:380-382, 1982; van der Loo et al., Immunogenetics 42:333-341, 1995; Karlin -et al., J. Mol. Evol. 22:195-208, 1985; Kindsvogel et al., DNA 1:335 343, 1982; Breiner et al.; Gene 18:165-174, 1982; Kondo et al., Eur. J.
Iznrzzunol. 23:245-249, 1993; and GenBank Accession No. J00228. For a review of immunoglobulin structure and function see Putnam, The Plasma Proteins, Vol V, Academic Press, Inc., 49-140, 1987; and Padlan, Mol. Imrazunol. 31:169-217, 1994.
The term "immunoglobulin CH1 domain" denotes a wild-type immunoglobulin heavy chain CH1 constant domain or a variant thereof, wherein the variant folds into the higher order structure characteristic of native immunoglobulin heavy chain constant domains (two twisted (3 sheets stabilized by a single disulfide bond; see, for example, Amzel and Poljak, Arzrzu.
Rev. Imrnunol. 48:961-997, 1979) and is capable of dimerizing with an immunoglobulin light chain constant domain. A representative wild-type human ~yl heavy chain is shown in Figs. 1A
1B (SEQ ID NO:1). The CHl domain extends from Eu residue 118 to Eu residue 215 (residues 1 to 98 of SEQ ID NO:1) according to the domain boundaries disclosed by Edelman et al., Proc.
Natl. Acad. Sci. USA 63:78-85, 1969. Those skilled in the art will recognize that domain boundaries are approximate and may vary by ~ 5 residues depending upon the criteria used to identify them. Immunoglobulin CHl domains include, for example, the CH1 domains of human gammal (SEQ ID N0:48), gamma2 (SEQ ID N0:49), gamma3 (SEQ >D NO:50), gamma4 (SEQ ID N0:51), alphal (SEQ ID N0:52), alpha2 (SEQ ID N0:53), delta (SEQ ID
N0:54), epsilon (SEQ ID N0:55), and mu (SEQ ID N0:56) immunoglobulin chains, as well as a consensus human CHl domain (SEQ ID N0:57) based on an alignment of the wild-type human CHl domains as shown in SEQ ~ NOS:48-56 (Figs. 2A-2B). SEQ ID NOS:48-57 are extended at their carboxyl termini relative the the CHl domain as shown in Fig. 1.
Variants of native Ig heavy chain constant domains include, without limitation, those disclosed by Lesk and Chothia, Nature 335:188-190, 1988 and Carter et al., U.S. Patent No. 5,807,706.
Variants of native Ig CH1 domains will exhibit a pair-wise level of sequence identity to wild-type CH1 domains at least as great as the minimum pair-wise identity exhibited in the alignment shown in Figs. 2A-2B. Representative mouse epsilon, gamma2a, and gamma3 CH1 domain sequences are shown in SEQ >D NOS:58, 59, and 60, respectively.
An "immunoglobulin hinge" is that portion of an immunglobulin heavy chain connecting the variable and CHl domains. Within SEQ ID NO:1, the hinge is approximately residues 99 to 113 (Eu residues 216-230 as shown in Fig. lA).
The term "immunoglobulin light chain x or 7~ constant region" denotes a native immunoglobulin light chain constant domain of the ~ (e.g., SEQ ll~ N0:61) or ~, (e.g., SEQ ID
N0:62) isotype, or a variant thereof, wherein the variant folds into the higher order structure characteristic of native immunoglobulin light chains constant domains and is capable of dimerizing with an immunoglobulin CH1 domain. Variants include a consensus light chain constant region as shown in SEQ ID N0:63, as well as other variants that exhibit a pair-wise level of sequence identity to wild-type light chain constant domains at least as great as the minimum pair-wise identity exhibited in the alignment shown in Fig. 3.
The "immunoglobulin superfamily" is a functionally diverse family of proteins having structural similarities to immunoglobulins. These proteins have structural features, including disulfide bonds, characteristic of immunoglobulin constant and/or variable region domains. Many members of the immunoglobulin superfamily are cell-surface proteins that perform a recognition role.
"Non-covalent associations" between polypeptides or proteins include hydrogen bonding, steric interactions, hydrophobic interactions, and ionic interactions.
A "non-immunoglobulin polypeptide" is a polypeptide that is not an immunoglobulin or fragment of an immunoglobulin. However, the term "non-immunoglobulin polypeptide" does not exclude polypeptides that are members (or fragments of members) of the immunoglobulin superfamily or that contain immunoglobulin-like domains, so long as they are not themselves immunoglobulins. Non-immunoglobulin polypeptides that are members of the immunoglobulin superfamily include, without limitation, PDGF alpha receptor, PDGF beta receptor, CTLA-4, CSF-1 receptor, stem cell factor receptor, and the like.
0 "Operably linked" means that two or more entities are joined together such that they function in concert for their intended purposes. When referring to DNA
segments, the phrase indicates, for example, that coding sequences are joined in the correct reading frame, and transcription initiates in the promoter and proceeds through the coding segments) to the terminator. When referring to polypeptides, "operably linked" includes both covalently (e.g., by disulfide bonding) and non-covalently (e.g., by hydrogen bonding, hydrophobic interactions, or salt-bridge interactions) linked sequences, wherein the desired functions) of the sequences are retained.
A "polynucleotide" is a single- or double-stranded polymer of deoxyribonucleotide or ribonucleotide bases read from the 5' to the 3' end.
Polynucleotides include RNA and DNA, and may be isolated from natural sources, synthesized in vitro, or prepared from a combination of natural and synthetic molecules. Sizes of polynucleotides are expressed as base pairs (abbreviated "bp"), nucleotides ("nt"), or kilobases ("kb"). Where the context allows, the latter two terms may describe polynucleotides that are single-stranded or double-stranded. When the term is applied to double-stranded molecules it is used to denote overall length and will be understood to be equivalent to the term "base pairs". It will be recognized by those skilled in the art that the two strands of a double-stranded polynucleotide may differ slightly in length and that the ends thereof may be staggered as a result of enzymatic cleavage; thus all nucleotides within a double-stranded polynucleotide molecule may not be paired. Such unpaired ends will in general not exceed 20 nt in length.
A "polypeptide" is a polymer of amino acid residues joined by peptide bonds, whether produced naturally or synthetically. Polypeptides of less than about 10 amino acid residues are commonly referred to as "peptides".
The term "promoter" is used herein for its art-recognized meaning to denote a portion of a gene containing DNA sequences that provide for the binding of RNA
polymera.se and initiation of transcription. Promoter sequences are commonly, but not always, found in the 5' non-coding regions of genes.
A "protein" is a macromolecule comprising one or more polypeptide chains. A
protein may also comprise non-peptidic components, such as carbohydrate groups.
Carbohydrates and other non-peptidic substituents may be added to a protein by the cell in which the protein is produced, and will vary with the type of cell. Proteins are defined herein in terms of their amino acid backbone structures; substituents such as carbohydrate groups are generally not specified, but may be present nonetheless.
A "secretory signal sequence" is a DNA sequence that encodes a polypeptide (a "secretory peptide") that, as a component of a larger polypeptide, directs the larger polypeptide through a secretory pathway of a cell in which it is synthesized. The larger polypeptide is commonly cleaved to remove the secretory peptide during transit through the secretory pathway.
A "segment" is a portion of a larger molecule (e.g., polynucleotide or polypeptide) having specified attributes. For example, a DNA segment encoding a specified polypeptide is a portion of a longer DNA molecule, such as a plasmid or plasmid fragment, that, when read from the 5' to the 3' direction, encodes the sequence of amino acids of the specified polypeptide.
The present invention provides dimeric fusion proteins, including homodimers and heterodimers. Within certain embodiments of the invention the dimers are heterodimers consisting of first and second polypeptide fusions. The first polypeptide fusion comprises a first dimerizing domain and the second polypeptide fusion comprises a second dimerizing domain.
When the first and second fusions are co-expressed in a recombinant host cell, the first and second dimerizing domains associate with each other to form a heterodimer. The dimer is stabilized by the formation of one or more interchain disulfide bonds. The invention is particularly useful for the production of heterodimeric binding partner pairs, including soluble forms of heterodimeric receptors. However, as shown in Example 6, a second polypeptide fusion of the invention comprising an immunoglobulin light chain constant domain as the dimerizing domain forms a disulfide-bonded homodimer when expressed in the absence of the first polypeptide fusion.

The first polypeptide fusion consists of, from amino terminus to carboxyl terminus: a first non-immunoglobulin polypeptide; a polypeptide linker consisting of from 18 to 32 amino acid residues, wherein from 1 to 8 of the residues are cysteine residues; a first dimerizing domain; and, optionally, a linking polypeptide of from 1 to 29 amino acid residues including at least one cysteine residue. Within certain embodiments of the invention, the first dimerizing domain is an immunoglobulin CH1 domain.
The second polypeptide fusion consists of, from amino terminus to carboxyl terminus: a second non-immunoglobulin polypeptide; a polypeptide linker consisting of from 18 to 32 amino acid residues, wherein from 1 to 8 of said residues are cysteine residues; and a second dimerizing domain. Within certain embodiments of the invention the second dimerizing domain is an immunoglobulin light chain constant domain selected from the group consisting of C~ and C~,.
Within certain embodiments of the invention, the first and second non-immunoglobulin polypeptides have functions that are to be combined in a single molecule. Such functions can be disparate or complementary. "Disparate functions" is used herein to indicate distinct biological functions or properties that reside in individual polypeptides that do not naturally function in concert. Examples of polypeptides having disparate functions include targeting proteins that bind to cell-surface structures, cytotoxins, growth factors, enzymes, hormones, and cytokines. Thus, proteins of the present invention can combine, for example, a targeting protein and a cytotoxin so that the cytotoxin can be delivered to a particular cell or set of cells. "Complementary functions" is used herein to indicate that the first and second polypeptides function together in the natural milieu in a coordinated manner, and includes functional combinations that, in their natural occurance, result from covalent interactions or non-covalent interactions between the first and second polypeptides, or from spatial effects involving additional molecules. Typical of such complementary functions are the dimerization or other multimerization of cell-surface receptors and their ligands. As discussed above, receptors can exist in the cell membrane as pre-formed dimers or may form non-covalent dimers upon ligand binding. Higher order clustering of receptors has also been reported (e.g., Schwartz et al., Nature Immunol. 3:427-434, 2002).
Within certain embodiments of the invention the first and second non-immunoglobulin polypeptides are first and second binding partner polypeptides that functionally interact with each other to non-covalently associate with one or more additional polypeptides or proteins to perform a biological function. In general, the biological function is at least in part dependent, qualitatively or quantitatively, upon the association between the associated binding partner polypeptides and the additional polypeptide(s) or protein(s). Ligand-receptor binding is representative of such associations. Many ligands and receptors are multimeric. Cell-surface receptors may dimerize upon binding to dimeric ligands (e.g., PDGF receptors), receptors may exist as preformed dimers (e.g., insulin receptor), or ligand binding may result in the formation of higher order arrays of pre-existing dimeric receptors (e.g., CTLA-4).
Receptor monomers rnay associate covalently or non-covalently. If the first and second binding partner polypeptides are soluble receptor polypeptides, their binding of ligand may be enhanced by their association 5 (dimerization). If the first and second binding partner polypeptides are soluble ligands, dimerization may be required for binding to and/or activating cell surface receptors. Examples of binding partner polypeptides that can be used within the present invention include, without limitation, PDGF A, PDGF B, PDGF C, PDGF D, IL-12 p35 subunit, IL-12 p40 subunit, IL-23 p19 subunit, IL-27 p28 subunit, EBI3, zlutl (GPHB5) (US Patent Application Publication No.
10 20020160953A), zsig5l (GPHA2) (US Pat. No. 6,573,363), PDGF receptor ci (PDGF-Roc), PDGF-R(3, zalphal l (IL-21R) (US Patent Application Publication No.
20030148447A), zcytor7 (IL-20RA) (US Pat. No. 5,945,511), DIRSl (IL-20RB) (WIPO Publication No. WO
99/27103), zcytorll (IL-22RA1) (US Pat. No. 5,965,704), zcytorl7 (IL,-31RA) (US Patent Application Publication No. 20030096339A), zcytorl9 (IL-28RA) (US Patent Application Publication No.
20030027253A), CRF2-4 (1L10-R(3), TACI (US Pat. No. 5,969,102), BCMA, BAFF
receptor, TNF receptors, growth hormone receptor, KH97 (human common beta subunit), gp130, LIF
receptor (LIF-R(3), oncostatin M receptor (OSM-R(3) IL-2R(3, IL-2 receptor 'y common subunit (~yC), IL-3Ra,, IL-3R[3, IL,-4Ra,, IL-6Ra, IL-llRa, IL-12R[31, IL,-12R(32, 1L-13Ra1, IL13-Ra2, IL-17R, IL-23R, OPG, Fas, WSX-1 (TCCR) (US 5,792,850), CNTF-Rot,, and GM-CSF-Ra,.
Exemplary combinations of first and second non-immunoglobulin polypeptides are shown in Table 1.
TART.R 1 First non-I of a tideSecond non-I of a tide PDGF A PDGF B

IL-12 40 IL,-12 40 lL-12 40 IL-23 19 PDGF A PDGF A

PDGF B PDGF B

PDGF C PDGF C

PDGF D PDGF D

tinued PDGF-Ra PDGF-R(3 zc tor7 DIRS 1 zc torll DIRS1 zcytor 17 OSM-R(3 zc torl9 CRF2-4 1L-4Ra YC

IL-23R IL-l2R~i 1 1L-4Ra IL-13Ra1 IL-3R(3 IL-5Ra IL-3R(3 GM-CSF-Ra IL-3R(3 1L-3Ra IL-12R(31 IL-12R[32 1L-21R 'YC

IL-6Ra 130 OSM-R(3 130 LIF-R(3 130 PDGF B PDGF A

IL-12 p40 IL-12 35 PDGF-R(3 PDGF-Ra DIRS 1 zcytor7 DIRS1 zc torll OSM-R(3 zc torl7 CRF2-4 zc torl9 ~yC IL-4Ra I

Table ntinued 1, co IL-13Ra1 IL,-4Ra IL-5Rcc II,-3R(3 GM-CSF-Ra IL-3R~i IL-3Ra IL-3R(3 IL-12R[32 IL-12R(31 130 IL-6Ra 130 OSM-R(3 130 LIF-R(3 The polypeptide linker consists of from 18 to 32 amino acid residues, wherein from 1 to 8 of said residues are cysteine residues. Within a preferred embodiment of the invention, each linker contains exactly two cysteine residues. Within another embodiment, the linker is 18 amino acid residues in length. The linker is designed to provide sufficient space and flexibility between the dimerizing domain and the non-immunoglobulin polypeptide witlun each of the first and second polypeptide fusions to allow the first and second non-immunoglobulin polypeptides to perform their intended functions within the dimeric protein.
Those skilled in the art will recognize that the precise spacing, and hence the precise length of each polypeptide linker, will depend upon the choice of each of the first and second non-immunoglobulin polypeptides and its function. For example, binding partner polypeptides will generally associate in a dimeric structure, while polypeptides having disparate functions may exist in spaced relationship to each other. In any event, the linker length and composition are selected to provide the desired spacing and degree of flexibility, while also providing for one or more interchain disulfide bonds to stabilize the desired conformation.
Required linker lengths can be determined through molecular modeling of the first and second non-immunoglobulin polypeptides, such as by analysis of crystal structure data for native dimers. Such methods can also be used to determine the distance between the termini of the immunoglobulin domain components of the fusion protein. One skilled in the art can then calculate the required minimum linker length based on these determinations. In general, a maximum length of approximately 3 ~ per amino acid residue is used as a basis for predicting linleer length. To ensure sufficient length and flexibility in the linker it is often desirable to exceed the predicted minimum required length. Calculation of the effective length of a polypeptide in solution is routine in the art. See, for example, Creighton, Proteins: Structures arad Molecular Properties, 2"a edition, W.H. Freeman and Company, 1993, Chapter 5.
Within certain embodiments of the invention the polypeptide linker comprises an immunoglobulin hinge or a fragment or variant of an immunoglobulin hinge region. Within one embodiment of the invention the N-terminal most cysteine residue (Eu residue 220; residue 103 of SEQ ID NO:1), which in an assembled antibody forms a disulfide bond with an immunoglobulin light chain, is omitted from the hinge, either by replacement with another amino acid residue (e.g., Ser) or by deletion or truncation. Other changes in the hinge sequence can also be made. For example the Lys residue (Eu 218; residue 101 of SEQ ID
N0:1) can be changed to Arg. The polypeptide linker can thus comprise an immunoglobulin hinge region that contains two cysteine residues that form disulfide bonds with the polypeptide linker on the other chain. An immunoglobulin hinge region can be obtained from any immunoglobulin heavy chain. Gamma (IgG) hinge regions, such as the ~yl hinge, have been well characterized and are conveniently used within the present invention. Those skilled in the art will recognize that different hinge polypeptides will provide different angles within the dimeric proteins, thus specific hinge polypeptides may be selected to optimize the overall structure of the molecule.
Molecular modeling techniques, as disclosed above, can be used to select the optimal hinge to be used with any pair of non-immunoglobulin polypeptides.
Non-hinge residues within the linker polypeptide are selected to provide an overall hydrophilic character and to be non-immunogenic and flexible. As used herein, a "flexible" linker is one that lacks a substantially stable higher-order conformation in solution, although regions of local stability are permissible. In general, small, polar, and hydrophilic residues are preferred, and bulky and hydrophobic residues are undesirable.
Areas of local charge are to be avoided; if the linker polypeptide includes charged residues, they will ordinarily be positioned so as to provide a net neutral charge within a small region of the polypeptide. It is therefore preferred to place a charged residue adjacent to a residue of opposite charge. In general, preferred residues for inclusion within the linker polypeptide include Gly, Ser, Ala, Thr, Asn, and Gln; more preferred residues include Gly, Ser, Ala, and Thr; and the most preferred residues are Gly and Ser. In general, Phe, Tyr, Trp, Pro, Leu, Ile, Lys, and Arg residues will be avoided (unless present within an immunoglobulin hinge region of the linker), Pro residues due to their hydrophobicity and lack of flexibility, and Lys and Arg residues due to potential immunogenicity. However, these less desirable residues may be included to provide a specific proteolytic cleavage site as disclosed below. Cys residues will be included, as disclosed above, so as to provide for disulfide bonding. Exemplary linkers are those comprising the structure [Gly-Ser-Gly-Ser]~ (SEQ ID N0:2), wherein a is 1 or, 2, and [Gly-Gly-Gly-Ser]b (SEQ m N0:3), wherein b is 1 or 2. The sequence of the linker will also be designed to avoid unwanted proteolysis. However, within certain embodiments of the invention, the linker polypeptide comprises a proteolytic cleavage site to facilitate separation of the dimerizing domains from the remainder of the molecule, leaving intact the disulfide bonds) joining the first and second non-immunoglobulin polypeptides through their respective linker polypeptides.
Exemplary proteolytic cleavage sites include sequences cleaved by plasmin, thrombin, factor Xa, enterokinase, furin, rhinovirus 3C protease, renin, collagenase, and caspase-3. The use of these and other proteases to cleave fusion proteins is known in the art. See, for example, Rubinstein et al., WO 00/61768; van de Ven et al., U.S. Patent No. 5,935,815; Sgarlato, U.S.
Patent No.
5,935,824; and Fischer et al., U.S. Patent No. 6,010,844. Thrombin cleaves after the dipeptide sequence Pro-Arg. Enterokinase cleaves after the pentapeptide sequence Asp-Asp-Asp-Asp-Lys (SEQ ID NO:4). Factor Xa cleaves after the sequence Ile-Glu-Gly-Arg (SEQ ll~
NO:S).
Plasmin cleaves after the sequence Arg-Pro. The human rhinovirus 3C protease cleaves Gln-Gly peptide bonds, such as in the sequence Leu-Glu-Val-Leu-Phe-Gln-Gly-Pro (SEQ ID
N0:6).
Furin cleaves after Arg-Xaa-Lys/Arg-Arg (SEQ ID N0:7). Renin cleaves between adjacent leucine residues in the sequence Pro-Phe-His-Leu-Leu-Val-Tyr (SEQ ID N0:8).
Collagenase cleaves within the sequence Pro-Xaa-Gly-Pro-Xaa (SEQ m N0:9). Caspase-3 cleaves after the sequence Asp-Glu-Val-Asp (SEQ ID NO:10). Those skilled in the art will recognize that these proteolytic enzymes may also cleave at related sequences.
The first dimerizing domain is an immunoglobulin constant domain-like segment of an immunoglobulin or immunoglobulin superfamily member that is capable of dimerizing with a second dimerizing domain as disclosed below. First dimerizing domains within the present invention include immunoglobulin CH1 domains, T-cell receptor Ca and C(3 domains, major histocompatibility complex (MHC) class I oc3 domains, (32-microglobulin, and MHC class II ocz and (32 chains. These molecules are known in the art. See, for example, Orr et al., Biochemistry 18:5711-5720, 1979; Orr et al., Nature 282:266-270, 1979;
Kaufman et al., Cell 36:1-13, 1984; Kronenberg et al., An~2. Rev. Immunol. 4:529-591, 1986;
Lew et al., Immunology 57:3-18, 1986; Williams _and Barclay, Anfa. Rev. Immuuol. 6:381-405, 1988; and Suggs et al., Proc. Natl. Acad. Sci. USA 78:6613-6617, 1981.
The immunoglobulin CHl domain is capable of non-covalently pairing with an immunoglobulin light chain x or 7~ constant region. As disclosed above, the CHl domain can be a CH1 domain from the y, a,, ~, s, or 8 classes of immunoglobulins, or a variant of a wild-type domain. Within the ~y class, a CHl domain from any of the ~yl, ~Z, 'y3, or y4 subclasses can be used. Within one embodiment, a ~yl CHl domain is used. Both human and non-human sequences can be used. As disclosed in more detail below, Eu residue 150 (Phe) in a ~yl CH1 domain (residue 33 of SEQ ID NO:1) can be replaced with an Arg residue.
Other dimerizing domains can also be used within the invention. T-cell receptor (TCR) Cl domains of a, and (3 receptors (Ca and C(3 domains) are capable of non-covalently pairing with each other. MHC class I a3 domains non-covalently bind to (32-microglobulin.
MHC class II oc2 and (32 domains non-covalently bind to each other.
The optional linking polypeptide is a polypeptide consisting of from 1 to 29 amino acid residues wherein at least one of said residues is a cysteine residue. Within one 5 embodiment of the invention, the linking polypeptide is nine residues in length. Within other embodiments the linking polypeptide contains not more than 12 cysteine residues, not more than 2 cysteine residues, or exactly one cysteine residue. Within other embodiments, the linking polypeptide is a portion of an immunoglobulin hinge region comprising the cysteine residue that forms a disulfide bond with an immunoglobulin light chain (Eu residue 220).
10 The second dimerizing domain is an immunoglobulin constant domain-like segment of an immunoglobulin or immunoglobuliri superfamily member that is capable of dimerizing with the first dimerizing domain. Second dimerizing domains within the present invention include immunoglobulin light chain constant domains, T-cell receptor Ca and C(3 domains, major histocompatibility complex (MHC) class I a3 domains, X32-microglobulin, and 15 MHC class II oc2 and (32 domains. As disclosed above, Ig CHl domains dimerize with Ig light chain constant domains, T-cell receptor Ca and C(3 domains are capable of non-covalently pairing with each other, MHC class I oc3 domains non-covalently bind to (32-microglobulin, and MHC class II oc~, and (32 domains non-covalently bind to each other. Thus, first and second dimerizing domains are selected as shown in the following Table 2.
TART,F 2.
First Dimerizin Domain Second Dimerizing Domain I CH1 I li ht chain constant TCR Ca TCR C(3 TCR C(3 TCR Ca MHC class I a3 (32-micro lobulin (32-micro lobulin MHC class I a3 MHC class II a2 MHC class II (32 MHC class II (32 MHC class II a,2 When an immunoglobulin light chain constant domain is the second dimerizing domain, it is preferred to include within the domain the C-terminal cysteine residue (Eu residue 214; Edelman et al., ibid.) (residue 107 of SEQ 1D N0:61 or residue 104 of SEQ
ID N0:62).
Inclusion of this cysteine residue permits stabilization of the dimer structure by disulfide bonding to a cysteine residue within the optional linking polypeptide within the first polypeptide fusion.

Non-immunoglobulin polypeptides, hinge regions, and first and second dimerizing domains used within the present invention can be obtained from a variety of species.
If the dimeric protein is to be used therapeutically in humans, it is preferred that human polypeptide sequences be employed. However, non-human sequences can be used, as can variant sequences. For other uses, including in vitro diagnostic uses and veterinary uses, polypeptide sequences from humans or non-human animals can be employed, although sequences from the same species as the patient may be preferred for in vivo veterinary use or for in vitro uses where species specificity of intermolecular reactions is present. Thus, polypeptides for use within the present invention can be, without limitation, human, non-human primate, rodent, canine, feline, equine, bovine, ovine, porcine, lagomorph, and avian polypeptides, as well as variants thereof.
The present invention also provides polynucleotide molecules, including DNA
and RNA molecules, that encode the first and second fusion polypeptides disclosed above. The polynucleotides of the present invention include both single-stranded and double-stranded molecules. DNA sequences encoding immunoglobulins and non-immunoglobulin polypeptides, including receptor polypeptides and other binding partner polypeptides, are known in the art.
Additional DNA sequences encoding immunoglobulins and non-immunoglobulin polypeptides can be readily generated by those of ordinary skill in the art based on the genetic code.
Counterpart RNA sequences can be generated by substitution of U for T. Those skilled in the art will readily recognize that, in view of the degeneracy of the genetic code, considerable sequence variation is possible among polynucleotide molecules encoding a given polypeptide.
Methods for preparing DNA and RNA are well known in the art.
Complementary DNA (cDNA) clones are prepared from RNA that is isolated from a tissue or cell that produces large amounts of RNA encoding a polypeptide of interest.
Total RNA can be prepared using guanidine HCl extraction followed by isolation by centrifugation in a CsCI
gradient (Chirgwin et al., Bioclae»aistry 18:52-94, 1979). Poly (A)+ RNA is prepared from total RNA using the method of Aviv and Leder (Proc. Natl. Acid. Sci. USA 69:1408-1412, 1972).
Complementary DNA is prepared from poly(A)+ RNA using known methods. In the alternative, genomic DNA can be isolated. For some applications (e.g., expression in transgenic animals) it may be advantageous to use a genomic clone, or to modify a cDNA clone to include at least one genomic intron. Methods for identifying and isolating cDNA and genomic clones are well known and within the level of ordinary skill in the art, and include the use of the sequences disclosed herein, or parts thereof, for probing or priming a library.
Polynucleotides encoding polypeptides of interest are identified and isolated by, for example, hybridization or polymerise chain reaction ("PCR", Mullis, U.S. Patent 4,683,202). Expression libraries can be probed with antibodies to the polypeptide of interest, receptor fragments, or other specific binding partners.

The polynucleotides of the present invention can also be prepared by automated synthesis. The production of short, double-stranded segments (GO to 80 bp) is technically straightforward and can be accomplished by synthesizing the complementary strands and then annealing them. Longer segments (typically >300 bp) are assembled in modular form from single-stranded fragments that are from 20 to 100 nucleotides in length.
Automated synthesis of polynucleotides is within the level of ordinary skill in the art, and suitable equipment and reagents are available from commercial suppliers. See, in general, Glick and Pasternak, Molecular Biotechnology, Principles & Applications of Recombinant DNA, ASM
Press, Washington, D.C., 1994; Itakura et al., Ann. Rev. Bioclzem. 53: 323-356, 1984;
and Climie et al., Proc. Natl. Acad. Sci. USA 87:633-637, 1990.
The polypeptide fusions of the present invention can be produced in genetically engineered host cells according to conventional techniques. Suitable host cells are those cell types that can be transformed or transfected with exogenous DNA and grown in culture, and include bacteria, fungal cells, and cultured higher eukaryotic cells (including cultured cells of multicellular organisms), particularly cultured mammalian cells. Techniques for manipulating cloned DNA molecules and introducing exogenous DNA into a variety of host cells are disclosed by Sambrook et al., Molecular Cloning: A Laboratory Manual, 2nd ed., Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, 1989, and Ausubel et al., eds., Current Protocols in Molecular Biology, Green and Wiley and Sons, NY, 1993.
In general, a DNA sequence encoding a polypeptide fusion is operably linked to other genetic elements required for its expression, generally including a transcription promoter and terminator, within an expression vector. The vector will also commonly contain one or more selectable markers and one or more origins of replication, although those skilled in the art will recognize that within certain systems selectable markers may be provided on separate vectors, and replication of the exogenous DNA may be provided by integration into the host cell genome.
Selection of promoters, terminators, selectable markers, vectors and other elements is a matter of routine design within the level of ordinary skill in the art. Many such elements are described in the literature and are available through commercial suppliers.
To direct a polypeptide fusion into the secretory pathway of a host cell, a secretory signal sequence is provided in the expression vector. The secretory signal sequence may be that of the native non-immunoglobulin polypeptide, or may be derived from another secreted protein (e.g., t-PA; see, U.S. Patent No. 5,641,655) or synthesized de novo. An engineered cleavage site may be included at the junction between the secretory peptide and the remainder of the polypeptide fusion to optimize proteolytic processing in the host cell. The secretory signal sequence is operably linked to the DNA sequence encoding the polypeptide fusion, i.e., the two sequences are joined in the correct reading frame and positioned to direct the newly synthesized polypeptide fusion into the secretory pathway of the host cell. Secretory signal sequences are commonly positioned 5' to the DNA sequence encoding the polypeptide of interest, although certain signal sequences may be positioned elsewhere in the DNA sequence of interest (see, e.g., Welch et al., U.S. Patent No. 5,037,743; Holland et al., U.S. Patent No.
5,143,830).
Expression of polypeptide fusions via a host cell secretory pathway is expected to result in the production of dimeric proteins. Dimers may also be assembled iz2 vitro upon incubation of component polypeptides under suitable conditions. In general, in vitro assembly will include incubating the protein mixture under denaturing and reducing conditions followed by refolding and reoxidation of the polypeptides to form dimers. Recovery and assembly of proteins expressed in bacterial cells is disclosed below.
Cultured mammalian cells are suitable hosts for use within the present invention.
Methods for introducing exogenous DNA into mammalian host cells include calcium phosphate-mediated transfection (Wigler et al., Cell 14:725, 1978; Corsaro and Pearson, Sozzzatic Cell Genetics 7:603, 1981: Graham and Van der Eb, Virology 52:456, 1973), electroporation (Neumann et al., EMBO J. 1:841-845, 1982), DEAF-dextran mediated transfection (Ausubel et al., ibid.), and liposome-mediated transfection (Hawley-Nelson et al., Focus 15:73, 1993;
Ciccarone et al., Focus 15:80, 1993). The production of recombinant polypeptides in cultured mammalian cells is disclosed by, for example, Levinson et al., U.S. Patent No.
4,713,339; Hagen et al., U.S. Patent No. 4,784,950; Palmiter et al., U.S. Patent No. 4,579,821;
and Ringold, U.S.
Patent No. 4,656,134. Suitable cultured mammalian cells include the COS-1 (ATCC No. CRL
1650), COS-7 (ATCC No. CRL 1651), BHI~ (ATCC No. CRL 1632), BHK 570 (ATCC No.
CRL -10314), 293 (ATCC No. CRL 1573; Graham et al., J. Gez2. Virol. 36:59-72, 1977) and Chinese hamster ovary (e.g. CHO-I~1, ATCC No. CCL 61; CHO-DG44, Urlaub et al., Proc.
Natl. Acad. Sci. USA 77:4216-4220, 1980) cell lines. Additional suitable cell lines are known in the art and available from public depositories such as the American Type Culture Collection, Manassas, Virginia. Strong transcription promoters can be used, such as promoters from SV-40, cytomegalovirus, or myeloproliferative sarcoma virus. See, e.g., U.S. Patent No. 4,956,288 and U.S. Patent Application Publication No. 20030103986. Other suitable promoters include those from metallothionein genes (U.S. Patents Nos. 4,579,821 and 4,601,978) and the adenovirus major late promoter. Expression vectors for use in mammalian cells include pZP-1, pZP-9, and pZMP2l, which have been deposited with the American Type Culture Collection, University Blvd., Manassas, VA USA under accession numbers 98669, 98668, and PTA-5266, respectively, and derivatives of these vectors.
Drug selection is generally used to select for cultured mammalian cells into which foreign DNA has been inserted. Such cells are commonly referred to as "transfectants".
Cells that have been cultured in the presence of the selective agent and are able to pass the gene of interest to their progeny are referred to as "stable transfectants." An exemplary selectable marker is a gene encoding resistance to the antibiotic neomycin. Selection is carried out in the presence of a neomycin-type drug, such as G-418 or the like. Selection systems can also be used to increase the expression level of the gene of interest, a process referred to as "amplification."
Amplification is carned out by culturing transfectants in the presence of a low level of the selective agent and then increasing the amount of selective agent to select for cells that produce high levels of the products of the introduced genes. An exemplary amplifiable selectable marker is dihydrofolate reductase, which confers resistance to methotrexate. Other drug resistance genes (e.g. hygromycin resistance, mufti-drug resistance, puromycin acetyltransferase) can also be used. Cell-surface markers and other phenotypic selection markers can be used to facilitate identification of transfected cells (e.g., by fluorescence-activated cell sorting), and include, for example, CDB, CD4, nerve growth factor receptor, green fluorescent protein, and the like.
Other higher eukaryotic cells can also be used as hosts, including insect cells, plant cells and avian cells. The use of Agrobacterium rlii4ogenes as a vector for expressing genes in plant cells has been reviewed by Sinkar et al., J. Biosci.
(Bazzgalore) 11:47-58, 1987.
Transformation of insect cells and production of foreign polypeptides therein is disclosed by Guarino et al., U.S. Patent No. 5,162,222 and WIPO publication WO 94106463.
Insect cells can be infected with recombinant baculovirus, commonly derived from Autograplza californica nuclear polyhedrosis virus (AcNPV). See, King and Possee, The Baculovirus Expression. System: A Laboratory Guide, Chapman & Hall, London;
O'Reilly et al., Baculovirus Expression Vectors: A Laboratory Mayzual, Oxford University Press., New York, 1994; and Richardson, Ed., Baculovirus Expression Protocols. Methods izz Molecular Biology, Humana Press, Totowa, NJ, 1995. Recombinant baculovirus can also be produced through the use of a transposon-based system described by Luckow et al. (J. Virol. 67:4566-4579, 1993).
This system, which utilizes transfer vectors, is commercially available in kit form (BAC-TO-~ BAC kit; Life Technologies, Gaithersburg, MD). The transfer vector (e.g., PFASTBAC1; Life Technologies) contains a Tn7 transposon to move the DNA encoding the protein of interest into a baculovirus genome maintained in E. coli as a large plasmid called a "bacmid." See, Hill-Perkins -and Possee, J. Gezz. Virol. 71:971-976, 1990; Bonning et al., J.
Gefz. Virol. 75:1551-1556, 1994; and Chazenbalk and Rapoport, J. Biol. Chem. 270:1543-1549, 1995.
Using techniques known in the axt, a transfer vector encoding a polypeptide fusion is transformed into E. coli host cells, and the cells are screened for bacmids which contain an interrupted lacZ gene indicative of recombinant baculovirus. The bacmid DNA containing the recombinant baculovirus genome is isolated, using common techniques, and used to transfect Spodoptera frugiperda cells, such as Sf9 cells. Recombinant virus that expresses the polypeptide fusion is subsequently produced. Recombinant viral stocks are made by methods commonly used the art.
For protein production, the recombinant virus is used to infect host cells, typically a cell line derived from the fall armyworm, Spodoptera frugiperda (e.g., Sf9 or Sf21 cells) or Trichoplusia ni (e.g., HIGH FIVE cells; Invitrogen, Carlsbad, CA).
See, in general, Glick and Pasternak, ibid. See also, U.S. Patent No. 5,300,435. Serum-free media are used to grow and maintain the cells. Suitable media formulations are known in the art and can be obtained from commercial suppliers. The cells are grown up from an inoculation density of 5 approximately 2-5 x 105 cells to a density of 1-2 x 10~ cells, at which time a recombinant viral stock is added at a multiplicity of infection (MOI) of 0.1 to 10, more typically near 3.
Procedures used are generally described in available laboratory manuals (e.g., King and Possee, ibid.; O'Reilly et al., ibid.; Richardson, ibid.).
Fungal cells, including yeast cells, can also be used within the present invention.
10 Yeast species of particular interest in this regard include Sacclzaronzyces cerevisiae, Pichia pastoris, and Pichia nzetlaanolica. Methods for transforming S. cez-evisiae cells with exogenous DNA and producing recombinant polypeptides therefrom are disclosed by, for example, Kawasaki, U.S. Patent No. 4,599,311; Kawasaki et al., U.S. Patent No.
4,931,373; Brake, U.S.
Patent No. 4,870,008; Welch et al., U.S. Patent No. 5,037,743; and Murray et al., U.S. Patent 15 No. 4,845,075. Transformed cells are selected by phenotype determined by the selectable marker, commonly drug resistance or the ability to grow in the absence of a particular nutrient (e.g., leucine). An exemplary vector system for use in Saccharonzyces cez-evisiae is the POTI
vector system disclosed by Kawasaki et al. (U.S. Patent No. 4,931,373), which allows transformed cells to be selected by growth in glucose-containing media.
Suitable promoters and 20 terminators for use in yeast include those from glycolytic enzyme genes (see, e.g., Kawasaki, U.S. Patent No. 4,599,311; Kingsman et al., U.S. Patent No. 4,615,974; and Bitter, U.S. Patent No. 4,977,092) and alcohol dehydrogenase genes. See also U.S. Patents Nos.
4,990,446;
5,OG3,154; 5,139,936; and 4,661,454. Transformation systems for other yeasts, including Hansenula polymorpha, Schizosaccharomyces ponzbe, Kluyverozzzyces lactis, Kluyveroznyces fragilis, Ustilago maydis, Pi.chia pastoris, Piclzia methanolica, Pichia guillerrnondii, and Cazzdida nzaltosa are known in the art. See, for example, Gleeson et al., J.
Gen. Micz-obiol.
132:3459-3465, 1986; Cregg, U.S. Patent No. 4,882,279; and Raymond et al., Yeast 14:11-23, 1998. Aspergillus cells may be utilized according to the methods of McKnight et al., U.S. Patent No. 4,935,349. Methods for transforming Acremonium clazysogenunz are disclosed by Sumino et al., U.S. Patent No. 5,162,228. Methods for transforming Neurospora are disclosed by Lambowitz, U.S. Patent No. 4,486,533. Production of recombinant proteins in Pichia zzzetlzazzolica is disclosed in U.S. Patents Nos. 5,716,808; 5,736,383;
5,854,039; and 5,888,768.
Prokaryotic host cells, including strains of the bacteria Escherichia coli., Bacillus and other genera are also useful host cells within the present invention.
Techniques for transforming these hosts and expressing foreign DNA sequences cloned therein are well known in the art (see, e.g., Sambrook et al., ibid.). When expressing a polypeptide fusion in bacteria such as E. coli, the polypeptide may be retained in the cytoplasm, typically as insoluble granules, or may be directed to the periplasmic space by a bacterial secretion sequence.
In the former case, the cells are lysed, and the granules are recovered and denatured using, for example, guanidine HCl or urea. The denatured pohypeptide can then be refolded and dimerized by diluting the denaturant, such as by dialysis against a solution of urea and a combination of reduced and oxidized glutathione, followed by dialysis against a buffered saline solution. In the alternative, the protein may be recovered from the cytoplasm in soluble form and isolated without the use of denaturants. The protein is recovered from the cell as an aqueous extract in, for example, phosphate buffered saline. To capture the protein of interest, the extract is applied directly to a chromatographic medium, such as an immobilized antibody or heparin-Sepharose column. Secreted polypeptides can be recovered from the periplasmic space in a soluble and functional form by disrupting the cells (by, for example, sonication or osmotic shock) and recovering the protein, thereby obviating the need for denaturation and refolding. See, for example, Lu et al., J. Izzzzzzunol. Meth. 267:213-226, 2002.
Transformed or transfected host cells are cultured according to conventional procedures in a culture medium containing nutrients and other components required for the growth of the chosen host cells. A variety of suitable media, including defined media and complex media, are known in the art and generally include a carbon source, a nitrogen source, essential amino acids, vitamins and minerals. Media may also contain such components as growth factors or serum, as required. The growth medium will generally select for cells containing the exogenously added DNA by, for example, drug selection or deficiency in an essential nutrient which is complemented by the selectable marker carried on the expression vector or co-transfected into the host cell.
When the linker polypeptide segment comprises a proteolytic cleavage site, the fusion proteins of the present invention can be cleaved within the host cell to remove the dimerizing domain if the host cell produces a protease that cleaves at the cleavage site. If the host cell does not naturally produce the protease, it can be transfected to co-express the protease and the fusion protein. See, for example, U.S. Patents Nos. 5,648,254 and 5,935,815.
Proteins of the present invention that contain a cleavage site in the linker polypeptide can also be cleaved in vitro according to conventional methods.
The use of proteases for processing recombinant proteins is routine in the art and includes the use of immobilized proteases. See, for example, U.S. Patent No. 6,010,844. Specific reaction conditions are based on the protease to be used and will be adjusted to minimize unwanted proteolysis with the first polypeptide segment. In general, such parameters as reaction time and ratio of protease to substrate will be adjusted to obtain the desired result.
Proteins of the present invention are purified by conventional protein purification methods, typically by a combination of chromatographic techniques. See, in general, Affizzity Chz-onzatography: Principles & Methods, Pharmacia LKB
Biotechnology, Uppsala, Sweden, 1988; and Scopes, Proteizz Purificatio>2: Principles and Practice, Springer Verlag, New York, 1994. Proteins comprising an immunoglobulin heavy chain polypeptide can be purified by affinity clu-omatography on immobilized protein A. Additional purification steps, such as gel filtration, can be used to obtain the desired level of purity or to provide for desalting, buffer exchange, and the like.
The proteins of the present invention can be used for diagnosis, therapy, or research to provide one or more activities associated with the first and second non-immunoglobulin polypeptides. Such activities include, without limitation, receptor binding, receptor activation, and ligand binding. Those skilled in the art will readily envision a range of uses for the proteins. Therapeutic uses include, for example, use as cytokine antagonists, such as for the treatment of cancers or immunological disorders, and as growth factor agonists, such as to promote tissue growth or healing or to promote development of vasculature or other tissue.
Diagnostic uses include, for example, use as targeting agents for radioisotopes or other labels, for detecting the presence of molecules on cell surfaces or in biological fluids or extracts, or ~as controls in in vitro assays. Within research the proteins of the present invention can be used, for example, for labeling cells, assaying for the presence of cell-surface receptors or soluble molecules, and to study the biology of non-immunoglobulin polypeptides or their binding partners.
For pharmaceutical use, the proteins of the present invention are formulated for local or systemic (particularly intravenous or subcutaneous) delivery according to conventional methods. In general, pharmaceutical formulations will include a protein of the present invention in combination with a pharmaceutically acceptable delivery vehicle. Delivery vehicles can be aqueous, lipidic, semi-solid or solid as appropriate to the condition to be treated and mode of delivery. Pharmaceutically acceptable aqueous vehicles include, without limitation, saline, buffered saline, 5% dextrose in water, and the like, and may further include one or more excipients, preservatives, solubilizers, buffering agents, albumin to prevent protein loss on vial surfaces, etc. Other suitable delivery vehicles include biocompatible solid or semi-solid matrices, including powdered bone, ceramics, biodegradable and non-biodegradable synthetic polymers, and natural polymers; tissue adhesives (e.g., fibrin-based); aqueous polymeric gels;
liposomes; salves; creams; ointments; powders; and the like. These and other suitable vehicles are known in the art. Methods of formulation are well known in the art and are disclosed, for example, in Rezuington: The Science azzd Practice of Pharmacy, 20th ed., Gennaro et al., eds., Lippincott, Williams & Wilkins, Baltimore, 2000. In general, the composition is administered in an amount that produces a statistically significant beneficial effect, such as a statistically significant moderation or reversal of the progression or severity of a disease. The exact dose will be determined by the clinician according to accepted standards, taking into account the nature and severity of the condition to be treated, patient traits, etc.
Determination of dose is within the level of ordinary skill in the art. Depending upon the route and method of administration, the protein may be administered in a single dose, as a prolonged infusion, or intermittently over an extended period. Intravenous administration will be by bolus injection or infusion over a typical period of one to several hours. Sustained release formulations can be employed.
The invention is further illustrated by the following, non-limiting examples.
Example 1 A vector encoding a heavy chain fusion partner (Fuses) joined to an optimized t-PA leader (U.S. Patent No. 5,641,655) and the extracellular domain (ECD) of zcytor7 (IL-20RA) (U.S. Patent No. 5,945,511) was constructed. The DNA sequence encoding the fusion protein is shown in SEQ ID NO:11. The Fuses sequence included a restriction enzyme site (Spel), followed by sequences encoding the human Ig gammal hinge with a Cys to Ser mutation at Eu residue 220 (residue 103 of SEQ ID NO:1), two repeats of (gly3ser) linker (SEQ
ID N0:3), a thrombin cleavage motif, a human Ig gammal CH1 domain, and a 5-residue gammal hinge fragment. Component sequences were synthesized and assembled by polymerase chain reaction (PCR) using oligonucleotide primers and plasmid templates for zcytor7 and human Ig gammal for the protein domains. Primers for the zcytor7 sequence were oligonucleotides zc42398 (SEQ
ID N0:13) and zc42651 (SEQ ID N0:14). Primers for the Fuses domain were oligonucleotide zc42652 (SEQ ID N0:15) and two antisense oligonucleotides, zc42393 (SEQ ID
N0:16) and zc42399 (SEQ ID N0:17). The resulting PCR products included 40 by of flanking sequence to facilitate recombination in yeast as disclosed by Raymond et al., Biotechniques 26(1):134-138, 140-141, 1999; and U.S. Patent No. 6,207,442. The PCR reactions were run in a 100 ~.1 final volume containing 10 ~,l lOX Taq polymerase Reaction Buffer (Perkin Elmer), 8 ~.1 of 2.5 mM
dNTPs, 78 ~Cl dHzO, 2 ~.l each of 20 mM stock solutions of the two primers described above, and taq polymerase (2.5 units, Life Technologies). An equal volume of mineral oil was added, and the reaction mixture was heated to 94°C for 2 minutes, followed by 25 cycles at 94°C for 30 seconds, 45°C for 30 seconds, 72°C for 60 seconds followed by a 5-minute extension at 72°C.
Ten ~,1 of each of the 100 ~tl PCR reaction mixtures were run on a 1.0%
agarose gel with 1 x TBE buffer (45 mM tris-borate, 1 mM EDTA, pH8.3) for analysis. The remaining 90 ~,1 of each reaction mixture was precipitated with the addition of 5 ~,1 1 M NaCI and 250 ~,l of absolute ethanol.
The plasmid pZMP2lz, cut with BgIII, was used for recombination with the PCR fragment. Plasmid pZMP2lz was constructed from pZMP21 (deposited at the American Type Culture Collection, 10801 University Boulevard, Manassas, VA 20110-2209, and designated No. PTA-5266) by replacement of the DHFR gene with the zeomycin resistance gene. Recombination of the PCR product into the linearized pZMP2lz vector was performed essentially as disclosed by Raymond et al., ibi.d. One hundred microliters of competent S.
cerevisiae strain SF838-9D (Rothman et al., EMBO J. 8:2057-2065, 1989) were combined with ~,l of a 1:1 mixture of the PCR products for the zcytor7 and Fuses domains and 200 ng of pZMP2lz linearized with BgIII and transferred to a 0.2-cm electroporation cuvette. The 5 yeast/DNA mixtures were electropulsed at 0.75 kV (5 kVlcm), o0 ohms, 25 ~,F
in an electroporator (GENEPULSERII; BIO-RAD Laboratories, Hercules, CA). To each cuvette was added 600 ~tl of 1.2 M sorbitol, and the yeast was plated in two 300-~,l aliquots onto two URA-D
plates and incubated at 30°C. (URA-D contains 0.056% -Ura -Trp -Thr powder (made by combining 4.0 g L-adenine, 3.0 g L-arginine, 5.0 g L-aspartic acid, 2.0 g L-histidine, G.0 g L-10 isoleucine, 8.0 g L-leucine, 4.0 g L-lysine, 2.0 g L-methionine, 6.0 g L-phenylalanine, 5.0 g L-serine, 5.0 g L-tyrosine, and 6.0 g L-valine), 0.67% yeast nitrogen base without amino acids, 2%
D-glucose, 0.5% 200X tryptophan-threonine solution (3.0% L-threonine, 0.8% L-tryptophan in H20), 1.8% agar (BACTO, Difco Laboratories).) After about 48 hours, the Ura+
yeast transformants from a single plate were resuspended in 1 ml HZO and spun briefly to pellet the yeast cells. The cell pellet was resuspended in 1 ml of lysis buffer (2%
TRITON X-100, 1 %
SDS, 100 mM NaCI, 10 mM Tris, pH 8.0, 1 mM EDTA). Five hundred microliters of the lysis mixture was added to a microcentrifuge tube containing 300 ~,1 acid-washed glass beads and 200 ~.1 phenol-chloroform mixture, vortexed for 1-minute intervals two or three times, followed by a 5-minute spin in a microcentrifuge at maximum speed. Three hundred microliters of the aqueous phase was transferred to a fresh tube, and the DNA was precipitated with 600 ~.l ethanol (EtOH), followed by centrifugation for 10 minutes at 4°C. The DNA
pellet was resuspended in 10 ~,1 HzO.
Transformation of electrocompetent E. coli cells (DH10B, GibcoBRL) was done with 0.5-2 ~,1 yeast DNA prep and 40 ~1 of DHlOB cells. The cells were electropulsed at 1.7 kV, 25 ~,F and 400 oluns in an electroporator. Following electroporation, 1 ml SOC
(2% BACTO
Tryptone (Difco, Detroit, MI), 0.5% yeast extract (Difco), 10 mM NaCI, 2.5 mM
KCI, 10 mM
MgCh, 10 mM MgS04, 20 mM glucose) was plated in 250-~,1 aliquots on four LB
AMP plates (LB broth (Lennox), 1.8% BACTO Agar (Difco), 100 mg/L Ampicillin).
Individual E. coli clones harboring the correct construct were identified by restriction digest to verify the presence of the insert and to confirm that the various DNA
sequences had been joined correctly to one another. The inserts of positive clones were subjected to DNA sequence analysis. Larger scale plasmid DNA was isolated using a commercially available plasmid DNA purification kit (QIAGEN Maxi kit; QIAGEN
Inc., Valencia, CA) according to the manufacturer's instructions. The amino acid sequence of the fusion protein is shown in SEQ ID N0:12.
The zcytor7 ECD-Fuses sequence was modified by the addition of an eight amino acid Gly-Ser linker (SEQ ID N0:2) between the zcytor7 ECD and the gammal hinge, and an eight amino acid histidine tag at the carboxyl terminus to aid in purification. The gly ser linker primers were zc43099 (SEQ ID N0:18) and zc43098 (SEQ ID N0:19). The his tag primers were zc43107 (SEQ ID N0:20) and zc43097 (SEQ ID N0:21). Double-stranded synthetic linker sequences were made by extension of oligonucleotide primers and added to the 5 zcytor7 and Fuses sequences by overlap PCR. Ten ~.l of each PCR product was used as template for the further levels of assembly. PCR reactions, yeast and E. coli transformations, and restriction mapping were carried out essentially as described above. The sequence of this construct is shown in SEQ ID N0:22.
10 Example 2 The Fuses sequence was mutagenized to modify the "ball-and-socket joint" for flexibility in the VH -to CHl interface as disclosed by Lesk and Chothia, Nature 335:188-90, 1988. The CH1 Phe residue, Eu number 150 (labeled 149 in Lesk and Chothia, ibid.; residue 33 of SEQ ID NO:1) was changed to Arg as shown in SEQ ID N0:24 and N0:25. The 5' end of 15 the modified Fuses sequence was synthesized by PCR using the His-tagged Fuses sequence (Example 1) as a template with primer zc43107 (SEQ ID N0:20) and zc43146 (SEQ
ID N0:26), and the 3' end with zc43145 (SEQ ID N0:27) and zc43098 (SEQ ID N0:19). The resulting two Fuses fragments were assembled by overlap PCR with the primers zc43107 (SEQ ID
N0:20) and zc43098 (SEQ ID N0:19). The assembled sequence, designated Fuses F149R, was recombined 20 with the zcytor7 fragment made with zc42398 and zc43097 by recombination in yeast. The sequence of this fusion is shown in SEQ ID N0:24 and N0:25.
Example 3 A second Fuses fusion protein, comprising the extracellular domain of zalphall 25 (IL21R) (U.S. Patent Application Publication No. 20030148447), was made by insertion of the zalphall ECD and leader sequences into the vector pZMP2lz (Example 1) with the Fuses partner described in Example 2. Primers zc44173 (SEQ ID N0:28) and zc44819 (SEQ ll~
N0:29) were used to make the ECD coding sequence plus 40 by extra for recombination via PCR with a previously constructed zalphall Ig fusion vector as a template. The Fuses partner was modified for overlap with the zalphall ECD using primers zc44820 (SEQ ID
N0:30) and zc43098 (SEQ 177 N0:19), also with 40 by of flanking sequence for recombination. The two fragments were co-electroporated into S. cerevisiae SF838-9D with pZMP2lz that had been linearized with BgIII as described for the previous expression vector using the procedures for PCR, electroporation, recovery of DNA from yeast cultures and screening disclosed above (Example 1). The DNA and amino acid sequences of the zalphal l-Fuses polypeptide are shown in SEQ ID NOS:31 and 32.

Example 4 A vector encoding a light chain fusion partner (FusL) joined to the leader and extracellular domain of DIRSl (IL20RB) (WO 99/27103) was constructed. The FusL
sequence included a restriction enzyme site (SpeI) and sequences encoding the human Ig gammal hinge, an eight-residue linker, a thrombin cleavage motif, and the human Ig kappa constant domain.
The sequences were synthesized and assembled by PCR with oligonucleotide primers using existing plasmid templates for DIRS 1 and human Ig kappa for the protein domains. The DIRS 1 domain was made by PCR using oligonucleotide primers zc42409 (SEQ ID N0:33) and zc42449 (SEQ ID N0:34). The FusL domain was made by PCR with oligonucleotide primers zc42650 (SEQ ID N0:35), and zc42397 (SEQ ID N0:36). PCR, DNA analysis, transformation of yeast and E. coli, plasmid mapping, and sequencing were performed essentially as described for the construction of vectors in Example 1. The DNA and amino acid sequences for the fusion are shown in SEQ ID N0:37 and N0:38.
The plasmid pZMP3l, which had been cut with BgIII, was used for recombination with the PCR fragment. Plasmid pZMP31 was constructed from pZMP21 (deposited at the American Type Culture Collection, 10801 University Boulevard, Manassas, VA 20110-2209, and designated No. PTA-5266) by the removal of the region from the truncated human CD8 alpha cDNA through the SV40 promoter/enhancer, leaving a single dicistronic cassette containing the polylinker followed by polio IRES, DHFR cDNA and SV40 poly A
region. Following recombination in yeast, the resulting vector was designated DIRS 1 C(FusL) pZMP3l.
The DIRS 1 ECD was modified by adding an eight amino acid (gly ser) linker (SEQ ~ N0:2). The gly ser linker primers were zc43105 (SEQ ID N0:39) and zc43106 (SEQ
ID N0:40). Double-stranded synthetic linker sequences were made by extension of oligonucleotide primers by PCR. To enable the recombination of this double-stranded oligonucleotide at one of the two SpeI sites of the pZMP31 vector, it was necessary to prepare several micrograms of a vector fragment that extended across that SpeI site. A
1.1 Kb PvuII to BsaBI fragment was isolated from the parent vector pZMP3l. This fragment was isolated by agarose gel electrophoresis followed by purification using a commercially available gel purification kit (obtained from QIAGEN, Inc., Valencia, CA). The DNA fragments (three from the parental vector and the PCR product) were recombined in yeast to construct pDIRS 1-FusL.
The DNA and amino acid sequences of the DIRS1-FusL polypeptide are shown in SEQ ID
N0:41 and SEQ ID N0:42.
Example 5 A second FusL fusion protein was made with the extracellular domain of IL-2~yC
in the same format by insertion of the IL-2~yC ECD sequence into the vector pZMP31 (described in Example 4) along with the FusL partner sequence described above, resulting in a ~yC-FusL
fusion including an optimized t-PA leader. Primers zc40915 (SEQ ID N0:43) and zc44745 (SEQ ID N0:44) were used to make the ECD sequence with 40 by overlaps for recombination, and primers zc44744 (SEQ ID N0:45) and zc42397 (SEQ ID N0:36) were used to make the FusL sequence by PCR. The two PCR products were inserted into BglII-cut pZMP31 as described above using recombination in yeast. The DNA and amino acid sequences of the IL-2~yC-FusL polypeptide shown in SEQ ID N0:46 and N0:47.
Example 6 Serum-free, suspension-adapted CHO DXB 11 cells (Graf and Chasin, Mol. Cell.
Biol. 2:93-96, 1982) were electroporated with pDIRS 1-FusL plasmid (described in Example 4), the zcytor7-Fuses plasmid (described in Example 2), and with both plasmids (to make the zcytol0 receptor heterodimer). The plasmids were linearized by digestion with PvuI, precipitated with sodium ,acetate and ethanol, then rinsed with 70% ethanol and dried. The pellets were resuspended at a concentration of 200 ~.g/100 ~.1 per electroporation in serum- and protein-free medium (EX-CELLTM 325, JRH Biosciences, Lenexa, KN) supplemented with 4 mM L-Glutamine, 1 % Hypoxanthine/Thymidine, 1 % vitamins, and 1 % Na pyruvate (media supplements purchased from INVITROGEN, Carlsbad, CA). Cells, growing at log phase, were pelleted and resuspended at 5 x 106 cells/800 ~.l per electroporation reaction. The electroporation was performed in an electroporator at 300 v and 950 ~,F in 4-mm cuvettes. The cells were suspended in 25 ml of the medium described above in 125-mL shake flasks and put on shakers in cell culture incubators at 37°C, at 80 rpm for 24 hours to recover. The cells were then pelleted and resuspended at 2.5 x 105 cells/mL in selective medium (EX-CELLTM

supplemented with 4 mM L-Glutamine, 1% vitamins, 1% Na Pyruvate and 200 ~,g/mL
zeocin).
Cell lines containing the DHFR gene were further cultured in increasing concentrations of methotrexate up to 1 ~,M once the cultures were capable of growing in the absence of hypoxanthine/thymidine supplementation. Once the cultures were growing actively in their respective selection media and the viability had increased to over 95%, cultures were established for harvest and analysis of protein. Cultures were seeded at 1 x 106 cells/mL
at 25 mL in shake flasks, and allowed to grow for 48 hours, then harvested. The supernatants were filtered through 0.22 ~,m filters and purified by IMAC (Immobilized Metal Affinity Chromatography). The crude supernatants and the purified proteins were analyzed by SDS PAGE, Western blot, ELISA, and B cell activity assay.
Cells transformed with the Fuses or FusL constructs only or cotransformed with both were analyzed by SDS PAGE and Western blot for expression, secretion into the medium, and assembly. Western blots were performed on 30-~,l samples of conditioned medium and lysates from cells derived from an equivalent volume of culture. Duplicate blots were reacted with antibodies to each chain. The standard was zcytor7 Ig (the zcytor7 ECD
joined to a human Ig gammal Fc fusion partner). Zcytor7-Fuses protein was detected only in the lysate in the Fuses homodimer cells (transformed with Fuses construct only), indicating a lack of secretion of the Fuses homodimer alone. The heterodimeric protein was detected in the conditioned medium as well as in the lysate of the heterodimer cell line (transformed with both Fuses and FusL
constructs). FusL was detected only in the conditioned medium in both the FusL
homodimer and the heterodimer cell lines, indicating good secretion of the FusL subunit both alone and when assembled with the Fuses. Purified protein was analyzed for information about assembly of the two subunits. Coomassie blue stained SDS PAGE of purified zcytor7 ECD-FusH/DIRSl ECD-FusL heterodimer, both reduced and nonreduced, demonstrated the presence of the disulfide-bonded heterodimer.
Example 7 Serum-free, suspension-adapted CHO DXB 11 cells were electroporated with the IL-2~yC-FusL plasmid (described in Example 5), the zalphall-Fuses plasmid (described in Example 3), and both plasmids (to make the IL-21 receptor heterodimer).
The plasmids were linearized by digestion with PvuI, precipitated with sodium acetate and ethanol, then rinsed with 70% ethanol and dried. The pellets were resuspended at a concentration of 200 ~,g/100 ~.l per electroporation in serum- and protein-free medium (EX-CELLTM 325, JRH Biosciences, Lenexa, KN) supplemented with 4 mM L-Glutamine, 1%
Hypoxanthine/Thymidine, 1 % vitamins and 1 % Na pyruvate (supplements from INVITROGEN, " Carlsbad, CA). Cells, growing at log phase, were pelleted and resuspended at-5 x 106 cells/800 ,ul per electroporation reaction. The electroporation was performed in an electroporator at 300 v and 950 ~,F in 4-mm cuvettes. The cells were suspended in 25 ml of the medium described above in 125-mL shake flasks and put on shakers in cell culture incubators at 37°C, at 80 rpm for 24 hours to recover. The cells were then pelleted and resuspended at 2.5 x 105 cells/mL in selective medium (EX-CELLTM 325 supplemented with 4 mM L-Glutamine, 1%
vitamins, 1%
Na Pyruvate and 200 ~,g/mL zeocin). Cell lines containing the DHFR gene were further cultured in increasing concentrations of methotrexate up to 1 ~.M once the cultures were .capable of growing in the absence of hypoxanthine/thymidine supplementation. Once the cultures were growing actively in their respective selection media and the viability had increased to over 95%, cultures were established for harvest and analysis of protein. Cultures were seeded at 1 x 10~
cells/mL at 25 mL in shake flasks, and allowed to grow for 48 hours, then harvested. The supernatants were filtered through 0.22 ~,m filters and purified by IMAC. The crude supernatants and the purified proteins were analyzed by SDS PAGE, Western blot, ELISA, and B-cell activity assay.

A B-cell proliferation assay was performed with marine B-cells (Parrish-Novak et al, Nature 408:57-63, 2000) comparing inhibition mediated by soluble IL-21 receptor (Fc fusion heterodimer ("IL-2IR Fc")) generated by purification with two epitope tags with the zalphall ECD FusH/~yC FusL heterodimer purified by I1VIAC chromatography alone. The assay measured inhibition of uptake of [3H]thymidine by marine CD43- B cells treated with 2 ~.glml anti-CD40, I00 ng/ml marine IL-21, and varying concentrations of IL,-21 receptor (Fc fusion heterodirner or FusH/FusL heterodimer). The activity of the FusH/FusL
heterodimer was shown to be approximately equivalent to that of the purified Fc fusion heterodimer.
A cell-based inhibition assay was performed to compare the activities of the soluble Fc fusion to the the FusH/FusL form of the heterodimer. Fixed amounts of 1L-21 were preincubated with a concentration sexies of each heterodimer. BaF3 cells (Palacios and Steinmetz, Cell 4I: 727-734, 1985; Mathey-Prevot et al., Mol. Cell. Biol. 6:
4133-4135, 1986) transformed with IL-21R heterodimer constructs were incubated in a 96-well plate (VIEWPLATE; PerkinElmer, Wellesley, MA) ~ for 48 hours at 37°C and 6%
COZ in 1L-3 depleted medium containing the test mixture of soluble receptor heterodimer and II,-21 at a series of ratios from 0.4:1 up to 100:1. Cell growth was measured with 20 ~.l indicator dye (ALAMARBLUE~; Trek Diagnostic Systems, Cleveland, OIL per well. Fluorescence was read in microplate spectrofluorometer (GEMINI EM; Molecular Devices Corporation, Sunnyvale, CA) at an excitation wavelength of 530 nm and an emission wavelength of 590 nm. Results were normalized for the activity of IL21 on Baf3/IL21R cells. The activities of the two heterodimers were the same on a molar basis.

DEMANDES OU BREVETS VOLUMINEUX
LA PRESENTE PARTIE DE CETTE DEMANDE OU CE BREVETS
COMPRI~:ND PLUS D'UN TOME.
CECI EST L,E TOME 1 DE 2 NOTE: Pour les tomes additionels, veillez contacter le Bureau Canadien des Brevets.
JUMBO APPLICATIONS / PATENTS
THIS SECTION OF THE APPLICATION / PATENT CONTAINS MORE
THAN ONE VOLUME.

NOTE: For additional valumes please contact the Canadian Patent Office.

Claims (36)

1. A dimeric protein consisting of a first polypeptide fusion disulfide bonded to a second polypeptide fusion, wherein said first polypeptide fusion consists of, from amino terminus to carboxyl terminus, P1-Ll-D1-(P2)n, wherein said second polypeptide fusion consists of, from amino terminus to carboxyl terminus, P3-L2-D2, and wherein:
P1 is a first non-immunoglobulin polypeptide;
L1 is a first polypeptide linker consisting of from 18 to 32 amino acid residues, wherein x of said residues are cysteine residues;
D1 is a first dimerizing domain selected from the group consisting of an immunoglobulin CH1 domain, a T-cell receptor C.alpha. domain, a T-cell receptor C.beta. domain, a major histocompatibility complex class I .alpha.3 domain, .beta.2-microglobulin, a major histocompatibility complex class II .alpha.2 domain, and a major histocompatibility complex class II .beta.2 domain;
P2 is a linking polypeptide consisting of from 1 to 29 amino acid residues, wherein at least one of said residues is a cysteine residue;
n is 0 or 1;
P3 is a second non-immunoglobulin polypeptide;
L2 is a second polypeptide linker consisting of from 18 to 32 amino acid residues, wherein y of said residues are cysteine residues;
D2 is a second dimerizing domain selected from the group consisting of an immunoglobulin light chain constant domain, a T-cell receptor C.alpha. domain, a T-cell receptor C.beta.
domain, a major histocompatibility complex class I .alpha.3 domain, .beta.2-microglobulin, a major histocompatibility complex class II .alpha.2 domain, and a major histocompatibility complex class II
.beta.2 domain;
each of x and y is an integer from 1 to 8; and x=y.

2. The dimeric protein of claim 1 wherein P1 and P3 are different.

3. The dimeric protein of claim 1 wherein n=1.

4. The dimeric protein of claim 1 wherein x=2 and y=2.

5. The dimeric protein of claim 1, wherein each of P1 and P3 is an extracellular domain of a cell-surface receptor.

6. The dimeric protein of claim 5, wherein each of P1 and P3 is an extracellular domain of a human cell-surface receptor.

7. The dimeric protein of claim 1, wherein each of P1 and P3 is not a member of the immunoglobulin superfamily.

8. The dimeric protein of claim 1 wherein each of P1 and P3 is individually selected from the group consisting of IL-17R, IL-20RA, IL-20RB, IL-21R, IL-28RA, IL-31RA, CRF2-4, and .gamma.C.

9. The dimeric protein of claim 1 wherein each of L1 and L2 contains exactly two cysteine residues.

10. The dimeric protein of claim 1 wherein each of L1 and L2 comprises an immunoglobulin hinge or fragment or variant thereof.

11. The dimeric protein of claim 10 wherein each of L1 and L2 comprises an immunoglobulin hinge variant wherein the cysteine residue corresponding to Eu residue 220 is replaced by serine.

12. The dimeric protein of claim 1 wherein each of L1 and L2 comprises a human .gamma.1 hinge or fragment or variant thereof.

13. The dimeric protein of claim 1 wherein each of L1 and L2 consists of 18 amino acid residues.

14. The dimeric protein of claim 1 wherein each of L1 and L2 comprises a plurality of glycine residues.

15. The dimeric protein of claim 1 wherein each of L1 and L2 comprises a plurality of serine residues.

16. The dimeric protein of claim 1 wherein each of L1 and L2 comprises [Gly-Ser-Gly-Ser]a (SEQ ID NO:2), wherein a is 1 or 2; or [Gly-Gly-Gly-Ser]b (SEQ
ID NO:3), wherein b is 1 or 2.

17. The dimeric protein of claim 1 wherein each of L1 and L2 comprises a proteolytic cleavage site.

18. The polypeptide fusion of claim 1 wherein:
D1 is an immunoglobulin CH1 domain; and D2 is an immunoglobulin .kappa. light chain constant domain or immunoglobulin .lambda.
light chain constant domain.

19. The dimeric protein of claim 18 wherein D2 is an immunoglobulin .kappa.
light chain constant domain.

20. The dimeric protein of claim 1 wherein:
a) one of P1 and P3 is a zcytor7 extracellular domain and the other of P1 and is a DIRS1 extracellular domain;
b) one of P1 and P3 is a zcytor11 extracellular domain and the other of P1 and P3 is a DIRS1 extracellular domain;
c) one of P1 and P3 is a zalpha11 extracellular domain and the other of P1 and P3 is an IL-2 receptor .lambda. common extracellular domain; or d) one P1 and P3 is a PDGF .alpha. receptor extracellular domain and the other of P1 and P3 is a PDGF .beta. receptor extracellular domain.

21. A polypeptide fusion consisting of, from amino terminus to carboxyl terminus, P1-L-D1-(P2)n, wherein:
P1 is a non-immunoglobulin polypeptide;
L is a polypeptide linker consisting of from 18 to 32 amino acid residues, wherein from 1 to 8 of said residues are cysteine residues;
D1 is a dimerizing domain selected from the group consisting of an immunoglobulin CH1 domain, a T-cell receptor C.alpha. domain, a T-cell receptor C.beta. domain, a major histocompatibility complex class I .alpha.3 domain, .beta.2-microglobulin, a major histocompatibility complex class II .alpha.2 domain, and a major histocompatibility complex class II .beta.2 domain;
P2 is a linking polypeptide consisting of from 1 to 29 amino acid residues, wherein at least one of said residues is a cysteine residue; and n is 0 or 1.

22. The polypeptide fusion of claim 21 wherein D1 is an immunoglobulin .lambda.1 CH1 domain.

23. The polypeptide fusion of claim 22 wherein the .gamma.1 CH1 domain is a human .gamma.1 CH1 domain.

24. The polypeptide fusion of claim 21 wherein P2 is a portion of an immunoglobulin hinge comprising a cysteine residue.

25. The polypeptide fusion of claim 21 wherein n is 1 and P2 consists of from to 15 amino acid residues.

26. The polypeptide fusion of claim 21 wherein n is 1 and P2 contains exactly one cysteine residue.

27. A polypeptide fusion consisting of, from amino terminus to carboxyl terminus, P3-L-D2, wherein:
P3 is a non-immunoglobulin polypeptide;
L is a polypeptide linker consisting of from 18 to 32 amino acid residues, wherein from 1 to 8 of said residues are cysteine residues; and D2 is a dimerizing domain selected from the group consisting of an immunoglobulin light chain constant domain, a T-cell receptor C.alpha. domain, a T-cell receptor C.beta.
domain, a major histocompatibility complex class I .alpha.3 domain, .beta.2-microglobulin, a major histocompatibility complex class II .alpha.2 domain, and a major histocompatibility complex class II
.beta.2 domain.

28. A polynucleotide encoding the polypeptide fusion of claim 21.

29. A polynucleotide encoding the polypeptide fusion of claim 27.

30. An expression vector comprising the following operably linked elements:
a transcription promoter;
a DNA segment encoding the polypeptide fusion of claim 21; and a transcription terminator.

31. A cultured cell into which has been introduced the expression vector of claim 30, wherein the cell expresses the DNA segment.

32. An expression vector comprising the following operably linked elements:
a transcription promoter;

a DNA segment encoding the polypeptide fusion of claim 27; and a transcription terminator.

33. A cultured cell into which has been introduced the expression vector of claim 32, wherein the cell expresses the DNA segment.

34. A method of making the dimeric protein of claim 1 comprising:
(a) culturing a cell comprising first and second expression units, wherein said first expression unit comprises:
a transcription promoter;
a first DNA segment encoding the first polypeptide fusion; and a transcription terminator, and wherein said second expression unit comprises:
a transcription promoter;
a second DNA segment encoding the second polypeptide fusion; and a transcription terminator, whereby said first and second DNA segments are expressed and the encoded polypeptide fusions are produced as a dimeric protein; and (b) recovering the dimeric protein.

35. A dimeric protein consisting of two polypeptide chains joined to each other by at least one disulfide bond, wherein each of said polypeptide chains is a polypeptide fusion consisting of, from amino terminus to carboxyl terminus, P3-L-D2, wherein:
P3 is a non-immunoglobulin polypeptide;
L is a polypeptide linker consisting of from 18 to 32 amino acid residues, wherein from 1 to 8 of said residues are cysteine residues; and D2 is an immunoglobulin light chain constant domain.

36. A method of making the dimeric protein of claim 35 comprising:
(a) culturing a cell comprising an expression unit, wherein said expression unit comprises:
a transcription promoter;
a DNA segment encoding the polypeptide fusion; and a transcription terminator, whereby said DNA segment is expressed and the encoded polypeptide fusion is produced as a homodimeric protein; and (b) recovering the homodimeric protein.