Derivatives of growth hormone and related proteins   

This application is a continuation of U.S. application Ser. No. 10/773,530, filed Feb. 5, 2004, which is a continuation of U.S. application Ser. No. 10/400,377, filed Mar. 26, 2003, now U.S. Pat. No. 7,148,333 which is a divisional of U.S. application Ser. No. 09/462,941, filed Jan. 14, 2000, now U.S. Pat. No. 6,608,183, which is a national stage application under 35 U.S.C. §371 of PCT Application Serial No. PCT/US98/14497, filed Jul. 13, 1998, which claims the benefit of priority from U.S. Provisional Application Ser. No. 60/052,516, filed Jul. 14, 1997. Each of the above-identified applications and patent is incorporated herein by reference in its entirety.

This invention was made in part with government support under Grant Nos. 1R43 CA84850 and 2R44CA84850, each awarded by the National Institutes of Health. The government has certain rights in the invention.

This application contains a Sequence Listing submitted as an electronic text file named “Sequence_Listing.txt”, having a size in bytes of 102 kb, and created on 8 Mar. 2001. The information contained in this electronic file is hereby incorporated by reference in its entirety pursuant to 37 CFR §1.52(e)(5).

The present invention relates to genetically engineered therapeutic proteins. More specifically, the engineered proteins include growth hormone and related proteins.

The following proteins are encoded by genes of the growth hormone (GH) supergene family (Bazan (1990); Mott and Campbell (1995); Silvennoinen and Ihle (1996)): growth hormone, prolactin, placental lactogen, erythropoietin (EPO), thrombopoietin (TPO), interleukin-2 (IL-2), IL-3, IL-4, IL-5, IL-6, IL-7, IL-9, IL-10, IL-11, IL-12 (p35 subunit), IL-13, IL-15, oncostatin M, ciliary neurotrophic factor, leukemia inhibitory factor, alpha interferon, beta interferon, gamma interferon, omega interferon, tau interferon, granulocyte-colony stimulating factor (G-CSF), granulocyte-macrophage colony stimulating factor (GM-CSF), macrophage colony stimulating factor (M-CSF) and cardiotrophin-1 (CT-1) (“the GH supergene family”). It is anticipated that additional members of this gene family will be identified in the future through gene cloning and sequencing. Members of the GH supergene family have similar secondary and tertiary structures, despite the fact that they generally have limited amino acid or DNA sequence identity. The shared structural features allow new members of the gene family to be readily identified.

There is considerable interest on the part of patients and healthcare providers in the development of long acting, “user-friendly” protein therapeutics. Proteins are expensive to manufacture and, unlike conventional small molecule drugs, are not readily absorbed by the body. Moreover, they are digested if taken orally. Therefore, natural proteins must be administered by injection. After injection, most proteins are cleared rapidly from the body, necessitating frequent, often daily, injections. Patients dislike injections, which leads to reduced compliance and reduced drug efficacy. Some proteins, such as erythropoietin (EPO), are effective when administered less often (three times per week for EPO) because they are glycosylated. However, glycosylated proteins are produced using expensive mammalian cell expression systems.

The length of time an injected protein remains in the body is finite and is determined by, e.g., the protein's size and whether or not the protein contains covalent modifications such as glycosylation. Circulating concentrations of injected proteins change constantly, often by several orders of magnitude, over a 24-hour period. Rapidly changing concentrations of protein agonists can have dramatic downstream consequences, at times under-stimulating and at other times over-stimulating target cells. Similar problems plague protein antagonists. These fluctuations can lead to decreased efficacy and increased frequency of adverse side effects for protein therapeutics. The rapid clearance of recombinant proteins from the body significantly increases the amount of protein required per patient and dramatically increases the cost of treatment. The cost of human protein pharmaceuticals is expected to increase dramatically in the years ahead as new and existing drugs are approved for more disease indications.

Thus, there is a need to develop protein delivery technologies that lower the costs of protein therapeutics to patients and healthcare providers. The present invention provides a solution to this problem by providing methods to prolong the circulating half-lives of protein therapeutics in the body so that the proteins do not have to be injected frequently. This solution also satisfies the needs and desires of patients for protein therapeutics that are “user-friendly”, i.e., protein therapeutics that do not require frequent injections. The present invention solves these and other problems by providing biologically active, cysteine-added variants of members of the growth hormone supergene family. The invention also provides for the chemical modification of these variants with cysteine-reactive polymers or other types of cysteine-reactive moieties to produce derivatives thereof and the molecules so produced.

The present invention provides cysteine variants of members of the GH supergene family. The variants comprise a cysteine residue substituted for a nonessential amino acid of the proteins. Preferably, the variants comprise a cysteine residue substituted for an amino acid selected from amino acids in the loop regions, the ends of the alpha helices, proximal to the first amphipathic helix, and distal to the final amphipathic helix or wherein the cysteine residue is added at the N-terminus or C-terminus of the proteins. Preferred sites for substitution are the N- and O-linked glycosylation sites.

Also provided are cysteine variants wherein the amino acid substituted for is in the A-B loop, B-C loop, the C-D loop or D-E loop of interferon/interferon-10-like members of the GH supergene family.

Also provided are cysteine variants of members of the GH supergene family wherein the cysteine residue is introduced between two amino acids in the natural protein. In particular, the cysteine residue is introduced into the loop regions, the ends of the alpha helices, proximal to the first amphipathic helix, or distal to the final amphipathic helix. Even more particularly, the cysteine variant is introduced between two amino acids in an N—O-linked glycosylation site or adjacent to an amino acid in an N-linked or O-linked glycosylation site.

More particularly are provided cysteine variants wherein the loop region where the cysteine is introduced is the A-B loop, the B-C loop, the C-D loop or D-E loop of interferon/interferon-10-like members of the GH supergene family.

Such cysteine substitutions or insertion mutations also can include the insertion of one or more additional amino acids amino acids at the amino-terminal or carboxy-terminal to the cysteine substitution or insertion.

Also provided are cysteine variants that are further derivatised by PEGylating the cysteine variants and including the derivatised proteins produced thereby.

As set forth in the examples, specific cysteine variants of the members of the GH supergene family also are provided, including for example, variants of GH. The GH cysteine variants can have the substituted-for amino acid or inserted cysteine located at the N-terminal end of the A-B loop, the B-C loop, the C-D loop, the first three or last three amino acids in the A, B, C and D helices and the amino acids proximal to helix A and distal to helix D.

More particularly, the cysteine can be substituted for the following amino acids: F1, T3, P5, E33, A34, K38, E39, Q40, S43, Q46, N47, P48, Q49, T50, S51, S55, T60, A98, N99, S100, G104, A105, S106, E129, D130, G131, S132, P133, T135, G136, Q137, K140, Q141, T142, S144, K145, D147, T148, N149, S150, H151, N152, D153, S184, E186, G187, S188, and G190.

Other examples of cysteine variants according to the invention include erythropoietin variants. Erythropoietin variants include those wherein the substituted for amino acid is located in the A-B loop, the B-C loop, the C-D loop, the amino acids proximal to helix A and distal to helix D and the N- or C-terminus. Even more specifically, the EPO cysteine variants include molecules wherein the amino acids indicated below have a cysteine substituted therefor: serine-126, N24, 125, T26, N38, 139, T40, N83, S84, A1, P2, P3, R4, D8, S9, T27, G28, A30, E31, H32, S34, N36, D43, T44, K45, N47, A50, K52, E55, G57, Q58, G77, Q78, A79, Q86, W88, E89, T107, R110, A111, G113, A114, Q115, K116, E117, A118, S120, P121, P122, D123, A124, A125, A127, A128, T132, K154, T157, G158, E159, A160, T163, G164, D165, R166 and S85.

The members of the GH supergene family include growth hormone, prolactin, placental lactogen, erythropoietin, thrombopoietin, interleukin-2, interleukin-3, interleukin-4, interleukin-5, interleukin-6, interleukin-7, interleukin-9, interleukin-10, interleukin-11, interleukin-12 (p35 subunit), interleukin-13, interleukin-15, oncostatin M, ciliary neurotrophic factor, leukemia inhibitory factor, alpha interferon, beta interferon, gamma interferon, omega interferon, tau interferon, granulocyte-colony stimulating factor, granulocyte-macrophage colony stimulating factor, macrophage colony stimulating factor, cardiotrophin-1 and other proteins identified and classified as members of the family. The proteins can be derived from any animal species including human, companion animals and farm animals.

Other variations and modifications to the invention will be obvious to those skilled in the art based on the specification and the “rules” set forth herein. All of these are considered as part of the invention.

The present invention relates to cysteine variants and, among other things, the site-specific conjugation of such proteins with polyethylene glycol (PEG) or other such moieties. PEG is a non-antigenic, inert polymer that significantly prolongs the length of time a protein circulates in the body. This allows the protein to be effective for a longer period of time. Covalent modification of proteins with PEG has proven to be a useful method to extend the circulating half-lives of proteins in the body (Abuchowski et al., 1984; Hershfield, 1987; Meyers et al., 1991). Covalent attachment of PEG to a protein increases the protein's effective size and reduces its rate of clearance rate from the body. PEGs are commercially available in several sizes, allowing the circulating half-lives of PEG-modified proteins to be tailored for individual indications through use of different size PEGs. Other benefits of PEG modification include an increase in protein solubility, an increase in in vivo protein stability and a decrease in protein immunogenicity (Katre et al., 1987; Katre, 1990).

The preferred method for PEGylating proteins is to covalently attach PEG to cysteine residues using cysteine-reactive PEGs. A number of highly specific, cysteine-reactive PEGs with different reactive groups (e.g., maleimide, vinylsulfone) and different size PEGs (2-20 kDa) are commercially available (e.g., from Shearwater, Polymers, Inc., Huntsville, Ala.). At neutral pH, these PEG reagents selectively attach to “free” cysteine residues, i.e., cysteine residues not involved in disulfide bonds. The conjugates are hydrolytically stable. Use of cysteine-reactive PEGs allows the development of homogeneous PEG-protein conjugates of defined structure.

Considerable progress has been made in recent years in determining the structures of commercially important protein therapeutics and understanding how they interact with their protein targets, e.g., cell-surface receptors, proteases, etc. This structural information can be used to design PEG-protein conjugates using cysteine-reactive PEGs. Cysteine residues in most proteins participate in disulfide bonds and are not available for PEGylation using cysteine-reactive PEGs. Through in vitro mutagenesis using recombinant DNA techniques, additional cysteine residues can be introduced anywhere into the protein. The added cysteines can be introduced at the beginning of the protein, at the end of the protein, between two amino acids in the protein sequence or, preferably, substituted for an existing amino acid in the protein sequence. The newly added “free” cysteines can serve as sites for the specific attachment of a PEG molecule using cysteine-reactive PEGs. The added cysteine must be exposed on the protein's surface and accessible for PEGylation for this method to be successful. If the site used to introduce an added cysteine site is non-essential for biological activity, then the PEGylated protein will display essentially wild type (normal) in vitro bioactivity. The major technical challenge in PEGylating proteins with cysteine-reactive PEGs is the identification of surface exposed, non-essential regions in the target protein where cysteine residues can be added or substituted for existing amino acids without loss of bioactivity.

Cysteine-added variants of a few human proteins and PEG-polymer conjugates of these proteins have been described. U.S. Pat. No. 5,206,344 describes cysteine-added variants of IL-2. These cysteine-added variants are located within the first 20 amino acids from the amino terminus of the mature IL-2 polypeptide chain. The preferred cysteine variant is at position 3 of the mature polypeptide chain, which corresponds to a threonine residue that is O-glycosylated in the naturally occurring protein. Substitution of cysteine for threonine at position 3 yields an IL-2 variant that can be PEGylated with a cysteine-reactive PEG and retain full in vitro bioactivity (Goodson and Katre, 1990). In contrast, natural IL-2 PEGylated with lysine-reactive PEGs displays reduced in vitro bioactivity (Goodson and Katre, 1990). The effects of cysteine substitutions at other positions in IL-2 were not reported.

U.S. Pat. No. 5,166,322 teaches cysteine-added variants of IL-3. These variants are located within the first 14 amino acids from the N-terminus of the mature protein sequence. The patent teaches expression of the proteins in bacteria and covalent modification of the proteins with cysteine-reactive PEGs. No information is provided as to whether the cysteine-added variants and PEG-conjugates of IL-3 are biologically active. Cysteine-added variants at other positions in the polypeptide chain were not reported.

World patent application WO9412219 and PCT application US95/06540 teach cysteine-added variants of insulin-like growth factor-I (IGF-I). IGF-I has a very different structure from GH and is not a member of the GH supergene family (Mott and Campbell, 1995). Cysteine substitutions at many positions in the IGF-I protein are described. Only certain of the cysteine-added variants are biologically active. The preferred site for the cysteine added variant is at amino acid position 69 in the mature protein chain. Cysteine substitutions at positions near the N-terminus of the protein (residues 1-3) yielded IGF-I variants with reduced biological activities and improper disulfide bonds.

World patent application WO9422466 teaches two cysteine-added variants of insulin-like growth factor (IGF) binding protein-1, which has a very different structure than GH and is not a member of the GH supergene family. The two cysteine-added IGF binding protein-1 variants disclosed are located at positions 98 and 101 in the mature protein chain and correspond to serine residues that are phosphorylated in the naturally-occurring protein.

U.S. patent application Ser. No. 07/822,296 teaches cysteine added variants of tumor necrosis factor binding protein, which is a soluble, truncated form of the tumor necrosis factor cellular receptor. Tumor necrosis factor binding protein has a very different structure than GH and is not a member of the GH supergene family.

IGF-I, IGF binding protein-1 and tumor necrosis factor binding protein have secondary and tertiary structures that are very different from GH and the proteins are not members of the GH supergene family. Because of this, it is difficult to use the information gained from studies of IGF-I, IGF binding protein-I and tumor necrosis factor binding protein to create cysteine-added variants of members of the GH supergene family. The studies with IL-2 and IL-3 were carried out before the structures of IL-2 and IL-3 were known (McKay 1992; Bazan, 1992) and before it was known that these proteins are members of the GH supergene family. Previous experiments aimed at identifying preferred sites for adding cysteine residues to IL-2 and IL-3 were largely empirical and were performed prior to experiments indicating that members of the GH supergene family possessed similar secondary and tertiary structures.

Based on the structural information now available for members of the GH supergene family, the present invention provides “rules” for determining a priori which regions and amino acid residues in members of the GH supergene family can be used to introduce or substitute cysteine residues without significant loss of biological activity. In contrast to the naturally occurring proteins, these cysteine-added variants of members of the GH supergene family will possess novel properties such as the ability to be covalently modified at defined sites within the polypeptide chain with cysteine-reactive polymers or other types of cysteine-reactive moieties. The covalently modified proteins will be biologically active.

GH is the best-studied member of the GH supergene family. GH is a 22 kDa protein secreted by the pituitary gland. GH stimulates metabolism of bone, cartilage and muscle and is the body's primary hormone for stimulating somatic growth during childhood. Recombinant human GH (rhGH) is used to treat short stature resulting from GH inadequacy and renal failure in children. GH is not glycosylated and can be produced in a fully active form in bacteria. The protein has a short in vivo half-life and must be administered by daily subcutaneous injection for maximum effectiveness (MacGillivray et al., 1996). Recombinant human GH (rhGH) was approved recently for treating cachexia in AIDS patients and is under study for treating cachexia associated with other diseases.

The sequence of human GH is well known (see, e.g., Martial et al. 1979; Goeddel et al. 1979 which are incorporated herein by reference; SEQ ID NO: 1). GH is closely related in sequence to prolactin and placental lactogen and these three proteins were considered originally to comprise a small gene family. The primary sequence of GH is highly conserved among animal species (Abdel-Meguid et al., 1987), consistent with the protein's broad species cross-reactivity. The three dimensional folding pattern of porcine GH has been solved by X-ray crystallography (Abdel-Meguid et al., 1987). The protein has a compact globular structure, comprising four amphipathic alpha helical bundles joined by loops. Human GH has a similar structure (de Vos et al., 1992). The four alpha helical regions are termed A-D beginning from the N-terminus of the protein. The loop regions are referred to by the helical regions they join, e.g., the A-B loop joins helical bundles A and B. The A-B and C-D loops are long, whereas the B-C loop is short. GH contains four cysteine residues, all of which participate in disulfide bonds. The disulfide assignments are cysteine-53 joined to cysteine165 and cysteine182 joined to cysteine189.

The crystal structure of GH bound to its receptor revealed that GH has two receptor binding sites and binds two receptor molecules (Cunningham et al., 1991; de Vos et al., 1992). The two receptor binding sites are referred to as site I and site II. Site I encompasses the Carboxy (C)-terminal end of helix D and parts of helix A and the A-B loop, whereas site II encompasses the Amino (N)-terminal region of helix A and a portion of helix C. Binding of GH to its receptor occurs sequentially, with site I always binding first. Site II then engages a second GH receptor, resulting in receptor dimerization and activation of the intracellular signaling pathways that lead to cellular responses to GH. A GH mutein in which site II has been mutated (a glycine to arginine mutation at amino acid 120) is able to bind a single GH receptor, but is unable to dimerize GH receptors; this mutein acts as a GH antagonist in vitro, presumably by occupying GH receptor sites without activating intracellular signaling pathways (Fuh et al., 1992).

The roles of particular regions and amino acids in GH receptor binding and intracellular signaling also have been studied using techniques such as mutagenesis, monoclonal antibodies and proteolytic digestion. The first mutagenesis experiments entailed replacing entire domains of GH with similar regions of the closely related protein, prolactin (Cunningham et al., 1989). One finding was that replacement of the B-C loop of GH with that of prolactin did not affect binding of the hybrid GH protein to a soluble form of the human GH receptor, implying that the B-C loop was non-essential for receptor binding. Alanine scanning mutagenesis (replacement of individual amino acids with alanine) identified 14 amino acids that are critical for GH bioactivity (Cunningham and Wells, 1989). These amino acids are located in the helices A, B, C, and D and the A-B loop and correspond to sites I and II identified from the structural studies. Two lysine residues at amino acid positions 41 and 172, K41 and K172, were determined to be critical components of the site I receptor binding site, which explains the decrease in bioactivity observed when K172 is acetylated (Teh and Chapman, 1988). Modification of K168 also significantly reduced GH receptor binding and bioactivity (de la Llosa et al., 1985; Martal et al., 1985; Teh and Chapman, 1988). Regions of GH responsible for binding the GH receptor have also been studied using monoclonal antibodies (Cunningham et al., 1989). A series of eight monoclonal antibodies was generated to human GH and analyzed for the ability to neutralize GH activity and prevent binding of GH to its recombinant soluble receptor. The latter studies allowed the putative binding site for each monoclonal antibody to be localized within the GH three-dimensional structure. Of interest was that monoclonal antibodies 1 and 8 were unable to displace GH from binding its receptor. The binding sites for these monoclonal antibodies were localized to the B-C loop (monoclonal number 1) and the N-terminal end of the A-B loop (monoclonal number 8). No monoclonals were studied that bound the C-D loop specifically. The monoclonal antibody studies suggest that the B-C loop and N-terminal end of the A-B loop are non-essential for receptor binding. Finally, limited cleavage of GH with trypsin was found to produce a two chain derivative that retained full activity (Mills et al., 1980; Li, 1982). Mapping studies indicated that trypsin cleaved and/or deleted amino acids between positions 134 and 149, which corresponds to the C-D loop. These studies suggest the C-D loop is not involved in receptor binding or GH bioactivity.

Structures of a number of cytokines, including G-CSF (Hill et al., 1993), GM-CSF (Diederichs et al., 1991; Walter et al., 1992), IL-2 (Bazan, 1992; McKay, 1992), IL-4 (Redfield et al., 1991; Powers et al., 1992), and IL-5 (Milburn et al., 1993) have been determined by X-ray diffraction and NMR studies and show striking conservation with the GH structure, despite a lack of significant primary sequence homology. EPO is considered to be a member of this family based upon modeling and mutagenesis studies (Boissel et al., 1993; Wen et al., 1994). A large number of additional cytokines and growth factors including ciliary neurotrophic factor (CNTF), leukemia inhibitory factor (LIF), thrombopoietin (TPO), oncostatin M, macrophage colony stimulating factor (M-CSF), IL-3, IL-6, IL-7, IL-9, IL-12, IL-13, IL-15, and alpha, beta, omega, tau and gamma interferon belong to this family (reviewed in Mott and Campbell, 1995; Silvennoinen and Ihle 1996). All of the above cytokines and growth factors are now considered to comprise one large gene family, of which GH is the prototype.

In addition to sharing similar secondary and tertiary structures, members of this family share the property that they must oligomerize cell surface receptors to activate intracellular signaling pathways. Some GH family members, e.g., GH and EPO, bind a single type of receptor and cause it to form homodimers. Other family members, e.g., IL-2, IL-4, and IL-6, bind more than one type of receptor and cause the receptors to form heterodimers or higher order aggregates (Davis et al., 1993; Paonessa et al., 1995; Mott and Campbell, 1995). Mutagenesis studies have shown that, like GH, these other cytokines and growth factors contain multiple receptor binding sites, typically two, and bind their cognate receptors sequentially (Mott and Campbell, 1995; Matthews et al., 1996). Like GH, the primary receptor binding sites for these other family members occur primarily in the four alpha helices and the A-B loop (reviewed in Mott and Campbell, 1995). The specific amino acids in the helical bundles that participate in receptor binding differ amongst the family members (Mott and Campbell, 1995). Most of the cell surface receptors that interact with members of the GH supergene family are structurally related and comprise a second large multi-gene family (Bazan, 1990; Mott and Campbell, 1995; Silvennoinen and Ihle 1996).

A general conclusion reached from mutational studies of various members of the GH supergene family is that the loops joining the alpha helices generally tend to not be involved in receptor binding. In particular the short B-C loop appears to be non-essential for receptor binding in most, if not all, family members. For this reason, the B-C loop is a preferred region for introducing cysteine substitutions in members of the GH supergene family. The A-B loop, the B-C loop, the C-D loop (and D-E loop of interferon/IL-10-like members of the GH superfamily) also are preferred sites for introducing cysteine mutations. Amino acids proximal to helix A and distal to the final helix also tend not to be involved in receptor binding and also are preferred sites for introducing cysteine substitutions. Certain members of the GH family, e.g., EPO, IL-2, IL-3, IL-4, IL-6, G-CSF, GM-CSF, TPO, IL-10, IL-12 p35, IL-13, IL-15 and beta-interferon contain N-linked and O-linked sugars. The glycosylation sites in the proteins occur almost exclusively in the loop regions and not in the alpha helical bundles. Because the loop regions generally are not involved in receptor binding and because they are sites for the covalent attachment of sugar groups, they are preferred sites for introducing cysteine substitutions into the proteins. Amino acids that comprise the N- and O-linked glycosylation sites in the proteins are preferred sites for cysteine substitutions because these amino acids are surface-exposed, the natural protein can tolerate bulky sugar groups attached to the proteins at these sites and the glycosylation sites tend to be located away from the receptor binding sites.

Many additional members of the GH gene family are likely to be discovered in the future. New members of the GH supergene family can be identified through computer-aided secondary and tertiary structure analyses of the predicted protein sequences. Members of the GH supergene family will possess four or five amphipathic helices joined by non-helical amino acids (the loop regions). The proteins may contain a hydrophobic signal sequence at their N-terminus to promote secretion from the cell. Such later discovered members of the GH supergen family also are included within this invention.

The present invention provides “rules” for creating biologically active cysteine-added variants of members of the GH supergene family. These “rules” can be applied to any existing or future member of the GH supergene family. The cysteine-added variants will posses novel properties not shared by the naturally occurring proteins. Most importantly, the cysteine added variants will possess the property that they can be covalently modified with cysteine-reactive polymers or other types of cysteine-reactive moieties to generate biologically active proteins with improved properties such as increased in vivo half-life, increased solubility and improved in vivo efficacy.

Specifically, the present invention provides biologically active cysteine variants of members of the GH supergene family by substituting cysteine residues for non-essential amino acids in the proteins. Preferably, the cysteine residues are substituted for amino acids that comprise the loop regions, for amino acids near the ends of the alpha helices and for amino acids proximal to the first amphipathic helix or distal to the final amphipathic helix of these proteins. Other preferred sites for adding cysteine residues are at the N-terminus or C-terminus of the proteins. Cysteine residues also can be introduced between two amino acids in the disclosed regions of the polypeptide chain. The present invention teaches that N- and O-linked glycosylation sites in the proteins are preferred sites for introducing cysteine substitutions either by substitution for amino acids that make up the sites or, in the case of N-linked sites, introduction of cysteines therein. The glycosylation sites can be serine or threonine residues that are O-glycosylated or asparagine residues that are N-glycosylated. N-linked glycosylation sites have the general structure asparagine-X-serine or threonine (N-X-S/T), where X can be any amino acid. The asparagine residue, the amino acid in the X position and the serine/threonine residue of the N-linked glycosylation site are preferred sites for creating biologically active cysteine-added variants of these proteins. Amino acids immediately surrounding or adjacent to the O-linked and N-linked glycosylation sites (within about 10 residues on either side of the glycosylation site) are preferred sites for introducing cysteine-substitutions.

More generally, certain of the “rules” for identifying preferred sites for creating biologically active cysteine-added protein variants can be applied to any protein, not just proteins that are members of the GH supergene family. Specifically, preferred sites for creating biologically active cysteine variants of proteins (other than IL-2) are O-linked glycosylation sites. Amino acids immediately surrounding the O-linked glycosylation site (within about 10 residues on either side of the glycosylation site) also are preferred sites. N-linked glycosylation sites, and the amino acid residues immediately adjacent on either side of the glycosylation site (within about 10 residues of the N-X-S/T site) also are preferred sites for creating cysteine added protein variants. Amino acids that can be replaced with cysteine without significant loss of biological activity also are preferred sites for creating cysteine-added protein variants. Such non-essential amino acids can be identified by performing cysteine-scanning mutagenesis on the target protein and measuring effects on biological activity. Cysteine-scanning mutagenesis entails adding or substituting cysteine residues for individual amino acids in the polypeptide chain and determining the effect of the cysteine substitution on biological activity. Cysteine scanning mutagenesis is similar to alanine-scanning mutagenesis (Cunningham et al., 1992), except that target amino acids are individually replaced with cysteine rather than alanine residues.

Application of the “rules” to create cysteine-added variants and conjugates of protein antagonists also is contemplated. Excess production of cytokines and growth factors has been implicated in the pathology of many inflammatory conditions such as rheumatoid arthritis, asthma, allergies and wound scarring. Excess production of GH has been implicated as a cause of acromegaly. Certain growth factors and cytokines, e.g., GH and IL-6, have been implicated in proliferation of particular cancers. Many of the growth factors and cytokines implicated in inflammation and cancer are members of the GH supergene family. There is considerable interest in developing protein antagonists of these molecules to treat these diseases. One strategy involves engineering the cytokines and growth factors so that they can bind to, but not oligomerize receptors. This is accomplished by mutagenizing the second receptor binding site (site II) on the molecules. The resulting muteins are able to bind and occupy receptor sites but are incapable of activating intracellular signaling pathways. This strategy has been successfully applied to GH to make a GH antagonist (Cunningham et al., 1992). Similar strategies are being pursued to develop antagonists of other members of the GH supergene family such as IL-2 (Zurawski et al., 1990; Zurawski and Zurawski, 1992), IL-4 (Kruse et al., 1992), IL-5 (Tavernier et al., 1995), GM-CSF (Hercus et al., 1994) and EPO (Matthews et al., 1996). Since the preferred sites for adding cysteine residues to members of the GH supergene family described here lie outside of the receptor binding sites in these proteins, and thus removed from any sites used to create protein antagonists, the cysteine-added variants described herein could be used to generate long-acting versions of protein antagonists. As an example, Cunningham et al. (1992) developed an in vitro GH antagonist by mutating a glycine residue (amino acid 120) to an arginine. This glycine residue is a critical component of the second receptor binding site in GH; when it is replaced with arginine, GH cannot dimerize receptors. The glycine to arginine mutation at position 120 can be introduced into DNA sequences encoding the cysteine-added variants of GH contemplated herein to create a cysteine-added GH antagonist that can be conjugated with cysteine-reactive PEGs or other types of cysteine-reactive moieties. Similarly, amino acid changes in other proteins that turn the proteins from agonists to antagonists could be incorporated into DNA sequences encoding cysteine-added protein variants described herein. Considerable effort is being spent to identify amino acid changes that convert protein agonists to antagonists. Hercus et al.(1994) reported that substituting arginine or lysine for glutamic acid at position 21 in the mature GM-CSF protein converts GM-CSF from an agonist to an antagonist. Tavernier et al.(1995) reported that substituting glutamine for glutamic acid at position 13 of mature IL-5 creates an IL-5 antagonist.

Experimental strategies similar to those described above can be used to create cysteine-added variants (both agonists and antagonists) of members of the GH supergene family derived from various animals. This is possible because the primary amino acid sequences and structures of cytokines and growth factors are largely conserved between human and animal species. For this reason, the “rules” disclosed herein for creating biologically active cysteine-added variants of members of the GH supergene family will be useful for creating biologically active cysteine-added variants of members of the GH supergene family of companion animals (e.g., dogs, cats, horses) and commercial animal (e.g., cow, sheep, pig) species. Conjugation of these cysteine-added variants with cysteine-reactive PEGs will create long-acting versions of these proteins that will benefit the companion animal and commercial farm animal markets.

Proteins that are members of the GH supergene family (hematopoietic cytokines) are provided in Silvennoimem and Ihle (1996). Silvennoimem and Ihle (1996) also provide information about the structure and expression of these proteins. DNA sequences, encoded amino acids and in vitro and in vivo bioassays for the proteins described herein are described in Aggarwal and Gutterman (1992; 1996), Aggarwal (1998), and Silvennoimem and Ihle (1996). Bioassays for the proteins also are provided in catalogues of various commercial suppliers of these proteins such as R&D Systems, Inc. and Endogen, Inc.

The following examples are provided to demonstrate how these “rules” can be used to create cysteine-added variants of GH, erythropoietin, alpha interferon, beta interferon, G-CSF, GM-CSF and other members of the GH supergene family. The examples are not intended to be limiting, but only exemplary of specific embodiments of the invention.

This example discloses certain amino acids in GH that are non-essential for biological activity and which, when mutated to cysteine residues, will not alter the normal disulfide binding pattern and overall conformation of the molecule. These amino acids are located at the N-terminal end of the A-B loop (amino acids 34-52 of the mature protein sequence; SEQ ID NO: 1; Martial et al 1979; Goeddel et al 1979), the B-C loop (amino acids 97-105 in the mature protein sequence), and the C-D loop (amino acids 130-153 in the mature protein sequence). Also identified as preferred sites for introducing cysteine residues are the first three or last three amino acids in the A, B, C and D helices and the amino acids proximal to helix A and distal to helix D.

DNA sequences encoding wild type GH can be amplified using the polymerase chain reaction technique from commercially available single-stranded cDNA prepared from human pituitaries (ClonTech, San Diego, Calif.) or assembled using overlapping oligonucleotides. Specific mutations can be introduced into the GH sequence using a variety of procedures such as phage techniques (Kunkel et al 1987), PCR mutagenesis techniques (Innis et al 1990; White 1993) mutagenesis kits such as those sold by Stratagene (“Quick-Change Mutagenesis” kit, San Diego, Calif.) or Promega (Gene Editor Kit, Madison Wis.).

Cysteine substitutions can be introduced into any of the amino acids comprising the B-C loop, C-D loop and N-terminal end of the A-B loop or into the first three amino acids of the alpha helical regions that adjoin these regions. or in the region proximal to helix A or distal to helix D. Preferred sites for introduction of cysteine residues are: F1, T3, P5, E33, A34, K38, E39, Q40, S43, Q46, N47, P48, Q49, T50, S51, S55, T60, A98, N99, S100, G104, A105, S106, E129, D130, G131, S132, P133, T135, G136, Q137, K140, Q141, T142, S144, K145, D147, T148, N149, S150, H151, N152, D153, S184, E186, G187, S188, and G190. Cysteine residues also can be introduced at the beginning of the mature protein, i.e., proximal to the F1 amino acid, or following the last amino acid in the mature protein, i.e., following F191. If desirable, two or more such mutations can be readily combined in the same protein either by in vitro DNA recombination of cloned mutant genes and/or sequential construction of individual desired mutations.

The human GH gene was amplified from human pituitary single-stranded cDNA (commercially available from CLONTECH, Inc., Palo Alto, Calif.) using the polymerase chain reaction (PCR) technique and primers BB1 and BB2. The sequence of BB1 is 5′-GGGGGTCGACCATATGTTCCCAACCATTCCCTTATCCAG-3′ (SEQ ID NO: 24). The sequence of BB2 is 5′-GGGGGATCCTCACTAGAAGCCACAGCTGCCCTC-3′ (SEQ ID NO: 25). Primer BB1 was designed to encode an initiator methionine preceding the first amino acid of mature GH, phenylalanine, and SalI and NdeI sites for cloning purposes. The reverse primer, BB2, contains a BamHI site for cloning purposes. The PCR 100 microliter reactions contained 20 pmoles of each oligonucleotide primer, 1×PCR buffer (Perkin-Elmer buffer containing MgCl2), 200 micromolar concentration of each of the four nucleotides dA, dC, dG and dT, 2 ng of single-stranded cDNA, 2.5 units of Taq polymerase (Perkin-Elmer) and 2.5 units of Pfu polymerase (Stratagene, Inc). The PCR reaction conditions were 96° C. for 3 minutes, 35 cycles of (95° C., 1 minute; 63° C. for 30 seconds; 72° C. for 1 minute), followed by 10 minutes at 72° C. The thermocycler employed was the Amplitron II Thermal Cycler (Thermolyne). The approximate 600 bp PCR product was digested with SalI and BamHI, gel purified and cloned into similarly digested plasmid pUC19 (commercially available from New England BioLabs, Beverly, Mass.). The ligation mixture was transformed into E. coli strain DH5alpha and transformants selected on LB plates containing ampicillin. Several colonies were grown overnight in LB media and plasmid DNA isolated using miniplasmid DNA isolation kits purchased from Qiagen, Inc (Valencia, Calif.). Clone LB6 was determined to have the correct DNA sequence.

For expression in E. coli, clone LB6 was digested with NdeI and EcoRI, the approximate 600 bp fragment gel-purified, and cloned into plasmid pCYB1 (commercially available from New England BioLabs, Beverly, Mass.) that had been digested with the same enzymes and phosphatased. The ligation mixture was transformed into E. coli DH5alpha and transformants selected on LB ampicillin plates. Plasmid DNA was isolated from several transformants and screened by digestion with NdeI and EcoRI. A correct clone was identified and named pCYB1: wtGH (pBBT120). This plasmid was transformed into E. coli strains JM109 or W3110 (available from New England BioLabs and the American Type Culture Collection).

Wild type GH clone LB6 (pUC19: wild type GH) was used as the template to construct a GH clone containing the E. coli STII signal sequence (Picken et al. 1983). Because of its length, the STII sequence was added in two sequential PCR reactions. The first reaction used forward primer BB12 and reverse primer BB10. BB10 has the sequence:

(SEQ ID NO: 28) 5′CGCGGATCCGATTAGAATCCACAGCTCCCCTC 3′. BB12 has the sequence:

(SEQ ID NO: 30) 5′ATCTATGTTCGTTTTCTCTATCGCTACCAACGCTTACGCATTCCCAAC CATTCCCTTATCCAG-3′.

The PCR reactions were as described for amplifying wild type GH except that approximately 4 ng of plasmid LB6 was used as the template rather than single-stranded cDNA and the PCR conditions were 96° C. for 3 minutes, 30 cycles of (95° C. for 1 minute; 63° C. for 30 seconds; 72° C. for 1 minute) followed by 72° C. for 10 minutes. The approximate 630 bp PCR product was gel-purified using the Qiaex II Gel Extraction Kit (Qiagen, Inc), diluted 50-fold in water and 2 microliters used as template for the second PCR reaction. The second PCR reaction used reverse primer BB10 and forward primer BB11. BB11 has the sequence:

(SEQ ID NO: 29) 5′CCCCCTCTAGACATATGAAGAAGAACATCGCATTCCTGCTGGCATCTA TGTTCGTTTTCTCTATCG-3′.

Primer BB11 contains XbaI and NdeI sites for cloning purposes. PCR conditions were as described for the first reaction. The approximate 660 bp PCR product was digested with XbaI and BamHI, gel-purified and cloned into similarly cut plasmid pcDNA3.1(+) (Invitrogen, Inc. Carlsbad, Calif.). Clone pcDNA3.1(+)::stII-GH(5C) or “5C” was determined to have the correct DNA sequence.

Clone “5C” was cleaved with NdeI and BamHI and cloned into similarly cut pBBT108 (a derivative of pUC19 which lacks a Pst I site, this plasmid is described below). A clone with the correct insert was identified following digestion with these enzymes. This clone, designated pBBT111, was digested with NdeI and SalI, the 660 bp fragment containing the stII-GH fusion gene, was gel-purified and cloned into the plasmid expression vector pCYB1 (New England BioLabs) that had been digested with the same enzymes and phosphatased. A recombinant plasmid containing the stII-GH insertion was identified by restriction endonuclease digestions. One such isolate was chosen for further studies and was designated pBBT114. This plasmid was transformed into E. coli strains JM109 or W3110 (available from New England BioLabs and the American Type Culture Collection).

3. Construction of ompA-GH

Wild type GH clone LB6 (pUC19: wild type GH) was used as the template to construct a GH clone containing the E. coli ompA signal sequence (Movva et al 1980). Because of its length, the ompA sequence was added in two sequential PCR reactions. The first reaction used forward primer BB7:

(SEQ ID NO: 31) 5′GCAGTGGCACTGGCTGGTTTCGCTACCGTAGCGCAGGCCTTCCCAACC ATTCCCTTATCCAG 3′, and reverse primer BB10:

(SEQ ID NO: 28) 5′ CGCGGATCCGATTAGAATCCACAGCTCCCCTC 3′.

The PCR reactions were as described for amplifying wild type GH except that approximately 4 ng of plasmid LB6 was used as the template rather than single-stranded cDNA and the PCR conditions were 96° C. for 3 minutes, 30 cycles of (95° C. for 1 minute; 63° C. for 30 seconds; 72° C. for 1 minute) followed by 72° C. for 10 minutes. The approximate 630 bp PCR product was gel-purified using the Qiaex II Gel Extraction Kit (Qiagen, Inc), diluted 50-fold in water and 2 microliters used as template for the second PCR reaction. The second PCR reaction used reverse primer BB10 and forward PrimerBB6:

(SEQ ID NO: 32) 5′CCCCGTCGACACATATGAAGAAGACAGCTATCGCGATTGCAGTGGCAC TGGCTGGTTTC 3′.

PCR conditions were as described for the first reaction. The approximate 660 bp PCR product was gel-purified, digested with Sal I and Bam H1 and cloned into pUC19 (New England BioLabs) which was cut with Sal I and Bam H1 or pcDNA3.1(+) (Invitrogen) which had been cut by Xho I and Bam H1 (Sal I and Xho I produce compatible single-stranded overhangs). When several clones were sequenced, it was discovered that all pUC19 clones (8/8) contained errors in the region of the ompA sequence. Only one pcDNA3.1 (+) clone was sequenced and it contained a sequence ambiguity in the ompA region. In order to generate a correct ompA-GH fusion gene segments of two sequenced clones which contained different errors separated by a convenient restriction site were recombined and cloned into the pUC19-derivative that lacks the Pst I site (see pBBT108 described below). The resulting plasmid, termed pBBT 112, carries the ompA-GH fusion gene cloned as an Nde I-Bam H1 fragment into these same sites in pBBT108. This plasmid is designated pBBT112 and is used in PCR-based, site-specific mutagenesis of GH as described below.

To facilitate mutagenesis of the cloned GH gene for construction of selected cysteine substitution and insertion mutations a derivative of the plasmid pUC19 (New England BioLabs) lacking a Pst I site was constructed as follows. pUC19 plasmid DNA was digested with Pst I and subsequently treated at 75 deg. C. with PFU DNA Polymerase (Stratagene) using the vendor-supplied reaction buffer supplemented with 200 uM dNTPs. Under these conditions the polymerase will digest the 3′ single-stranded overhang created by Pst I digestion but will not digest into the double-stranded region. The net result will be the deletion of the 4 single-stranded bases which comprise the middle four bases of the Pst I recognition site. The resulting molecule has double-stranded, i.e., “blunt”, ends. Following these enzymatic reactions, the linear monomer was gel-purified using the Qiaex II Gel Extraction Kit (Qiagen, Inc). This purified DNA was treated with T4 DNA Ligase (New England BioLabs) according to the vendor protocols, digested with Pst I, and used to transform E coli DH5alpha. Transformants were picked and analyzed by restriction digestion with Pst I and Bam H1. One of the transformants which was not cleaved by Pst I but was cleaved at the nearby Bam H1 site was picked and designated pBBT108.

GH muteins were generally constructed using site-directed PCR-based mutagenesis as described in PCR Protocols: Current Methods and Applications edited by B. A. White, 1993 Humana Press, Inc., Totowa, N.J. and PCR Protocols: A Guide to Methods and Applications edited by Innis, M. A. et al 1990 Academic Press Inc San Diego, Calif. Typically PCR primer oligonucleotides are designed to incorporate nucleotide changes to the coding sequence of GH that result in substitution of a cysteine residue for an amino acid at a specific position within the protein. Such mutagenic oligonucleotide primers can also be designed to incorporate an additional cysteine residue at the carboxy terminus or amino terminus of the coding sequence of GH. In this latter case one or more additional amino acid residues could also be incorporated at the amino terminal and/or carboxy terminal to the added cysteine residue if that were desirable. Moreover, oligonucleotides can be designed to incorporate cysteine residues as insertion mutations at specific positions within the GH coding sequence if that were desirable. Again, one or more additional amino acids could be inserted along with the cysteine residue and these amino acids could be positioned at the amino terminal and/or carboxy terminal to the cysteine residue.

The cysteine substitution mutation T135C was constructed as follows. The mutagenic reverse oligonucleotide BB28:

(SEQ ID NO: 33) 5′CTGCTTGAAGATCTGCCCACACCGGGGGCTGCCATC3′ was designed to change the codon ACT for threonine at amino acid residue 135 to a TGT codon encoding cysteine and to span the nearby Bgl II site. This oligonucleotide was used in PCR along with the forward oligonucleotide BB34 5′GTAGCGCAGGCCTTCCCAACCATT3′ (SEQ ID NO: 34) which anneals to the junction region of the ompA-GH fusion gene and is not mutagenic. The PCR was performed in a 50 ul reaction in 1×PCR buffer (Perkin-Elmer buffer containing 1.5 mM MgCl2), 200 micromolar concentration of each of the four nucleotides dA, dC, dG and dT, with each oligonucleotide primer present at 0.5 μM, 5 pg of pBBT112 (described above) as template and 1.25 units of Amplitac DNA Polymerase (Perkin-Elmer) and 0.125 units of PFU DNA Polymerase (Stratagene). Reactions were performed in a Robocycler Gradient 96 thermal cycler (Stratagene). The program used entailed: 95 deg C. for 3 minutes followed by 25 cycles of 95 deg C. for 60 seconds, 45 deg C. or 50 deg C. or 55 deg C. for 75 seconds, 72 deg C. for 60 seconds followed by a hold at 6 deg C. The PCR reactions were analyzed by agarose gel electrophoresis to identify annealing temperatures that gave significant product of the expected size; ˜430 bp. The 45-deg C reaction was “cleaned up” using the QIAquick PCR Purification Kit (Qiagen), digested with Bgl II and Pst I. The resulting 278 bp Bgl II-Pst I fragment, which includes the putative T135C mutation, was gel-purified and ligated into pBBT111 the pUC19 derivative carrying the stII-GH fusion gene (described above) which had been digested with Bgl II and Pst I and gel-purified. Transformants from this ligation were initially screened by digestion with Bgl II and Pst I and subsequently one clone was sequenced to confirm the presence of the T135C mutation and the absence of any additional mutations that could potentially be introduced by the PCR reaction or by the synthetic oligonucleotides. The sequenced clone was found to have the correct sequence.

The substitution mutation S1132C was constructed using the protocol described above for T135C with the following differences: mutagenic reverse oligonucleotide BB29 5′CTGCTTGAAGATCTGCCCAGTCCGGGGGCAGCCATCTTC3′ (SEQ ID NO: 35) was used instead of BB28 and the PCR reaction with annealing temperature of 50 deg C. was used for cloning. One of two clones sequenced was found to have the correct sequence.

The substitution mutation T148C was constructed using an analogous protocol but employing a different cloning strategy. The mutagenic forward oligonucleotide BB30 5′GGGCAGATCTTCAAGCAGACCTACAGCAAGTTCGACTGCAACTCACACAAC3′ (SEQ ID NO: 36) was used in PCR with the non-mutagenic reverse primer BB33 5′CGCGGTACCCGGGATCCGATTAGAATCCACAGCT3′ (SEQ ID NO: 37) which anneals to the most 3′ end of the GH coding sequence and spans the Bam H1 site immediately downstream. PCR was performed as described above with the exception that the annealing temperatures used were 46, 51 and 56 deg C. Following PCR and gel analysis as described above the 46 and 51 deg C. reactions were pooled for cloning. These were digested with Bam H1 and Bgl II, gel-purified and cloned into pBBT111 which had been digested with Bam H1 and Bgl II, treated with Calf intestinal Alkaline Phosphatase (Promega) according to the vendor protocols, and gel-purified. Transformants from this ligation were analyzed by digestion with Bam H1 and Bgl II to identify clones in which the 188 bp Bam H1-Bgl II mutagenic PCR fragment was cloned in the proper orientation. Because Bam H1 and Bgl II generate compatible ends, this cloning step is not orientation specific. Five of six clones tested were shown to be correctly oriented. One of these was sequenced and was shown to contain the desired T148C mutation. The sequence of the remainder of the 188 bp Bam H1-Bgl II mutagenic PCR fragment in this clone was confirmed as correct.

The construction of the substitution mutation S144C was identical to the construction of T148C with the following exceptions. Mutagenic forward oligonucleotide BB31 5′GGGCAGATCTTCAAGCAGACCTACTGCAAGTTCGAC3′ (SEQ ID NO: 38) was used instead of BB30. Two of six clones tested were shown to be correctly oriented. One of these was sequenced and was shown to contain the desired S144C mutation. The sequence of the remainder of the 188 bp Bam H1-Bgl II mutagenic PCR fragment in this clone was confirmed as correct.

A mutation was also constructed that added a cysteine residue to the natural carboxy terminus of GH. The construction of this mutation, termed stp192C, was similar to that of T148C, but employed different oligonucleotide primers. The reverse mutagenic

oligonucleotide BB32

(SEQ ID NO: 39) 5′CGCGGTACCGGATCCTTAGCAGAAGCCACAGCTGCCCTCCAC3′ which inserts a TGC codon for cysteine between the codon for the carboxy terminal phe residue of GH and the TAA translational stop codon and spans the nearby Bam H1 site was used along with BB34 5′GTAGCGCAGGCCTTCCCAACCATT3′ (SEQ ID NO: 40) which is described above. Following PCR and gel analysis as described above, the 46 deg C. reaction was used for cloning. Three of six clones tested were shown to be correctly oriented. One of these was sequenced and was shown to contain the desired stp192C mutation. The sequence of the remainder of the 188 bp Bam H1H-Bgl II mutagenic PCR fragment in this clone was confirmed as correct.

Analogous PCR mutagenesis procedures can be used to generate other cysteine mutations. The choice of sequences for mutagenic oligonucleotides will be dictated by the position where the desired cysteine residue is to be placed and the propinquity of useful restriction endonuclease sites. Generally, it is desirable to place the mutation, i.e., the mismatched segment near the middle of the oligonucleotide to enhance the annealing of the oligonucleotide to the template. Appropriate annealing temperatures for any oligonucleotide can be determined empirically. It is also desirable for the mutagenic oligonucleotide to span a unique restriction site so that the PCR product can be cleaved to generate a fragment that can be readily cloned into a suitable vector, e.g., one that can be used to express the mutein or provides convenient restriction sites for excising the mutated gene and readily cloning it into such an expression vector. Sometimes mutation sites and restriction sites are separated by distances that are greater than that which is desirable for synthesis of synthetic oligonucleotides: it is generally desirable to keep such oligonucleotides under 80 bases in length and lengths of 30-40 bases are more preferable.

In instances where this is not possible, genes targeted for mutagenesis could be re-engineered or re-synthesized to incorporate restriction sites at appropriate positions. Alternatively, variations of PCR mutagenesis protocols employed above, such as the so-called “Megaprimer Method” (Barik, S., pp. 277-286 in Methods in Molecular Biology, Vol. 15: PCR Protocols: Current Methods and Applications edited by B. A. White, 1993, Humana Press, Inc., Totowa, N.J.) or “Gene Splicing by Overlap Extension” (Horton, R. M., pp. 251-261, in Methods in Molecular Biology, Vol. 15: PCR Protocols: Current Methods and Applications, edited by B. A. White, 1993, Humana Press, Inc., Totowa, N.J.) can also be employed to construct such mutations.

6. Expression of GH in pCYB1

To express GH in E. coli, pBBT120 (GH gene with no leader sequence cloned into the tac expression vector pCYB1) and pBBT114 (GH gene with stII leader sequence cloned into the tac expression vector pCYB 1) were transformed into E. coli strains JM109 and W3110. The parental vector pCYB1 was also transformed into JM109 and W3110. These strains were given the following designations:

BOB119: JM109 (pCYB1) BOB130: W3110 (pCYB1) BOB129: JM109 (pBBT120) BOB133: W3110 (pBBT120) BOB121: JM109 (pBBT114) BOB132: W3110 (pBBT114)

For expression, strains were grown overnight at 37° C. in Luria Broth (LB) (Sambrook, et al 1989) containing 100 μg/ml ampicillin. These saturated overnight cultures were diluted to ˜0.03 OD at A600 in LB containing 100 μg/ml ampicillin and incubated at 37° C. in shake flasks in rotary shaker, typically at 250-300 rpm. ODs were monitored and IPTG was added to a final concentration of 0.5 mM when culture ODs reached ˜0.25-0.5, typically between 0.3 and 0.4. Cultures were sampled typically at 1, 3, 5 and ˜16 h post-induction. The “˜16 h” time points represented overnight incubation of the cultures and exact times varied from ˜15-20 h. Samples of induced and uninduced cultures were pelleted by centrifugation, resuspended in 1× sample buffer (50 mM Tris-HCl (pH 6.8), 2% sodium lauryl sulfate, 10% glycerol, 0.1% bromphenol blue) with the addition of 1% β-mercaptoethanol when desirable. Samples were boiled for ˜10 minutes or heated to 95° C. for ˜10 minutes. Samples were cooled to room temperature before being loaded onto SDS polyacrylamide gels or were stored at −20° C. if not run immediately. Samples were run on precast 15% polyacrylamide “Ready Gels” (Bio-Rad, Hercules Calif.) using a Ready Gel Cell electrophoresis apparatus (Bio-Rad) according to the vendor protocols. Typically gels were run at 200 volts for ˜35-45 minutes. Gels were stained with Coomassie Blue or were analyzed by Western Blot following electro-blotting. Coomassie staining of whole cell lysates from strains BOB 129, BOB 133, BOB 121 and BOB 132 showed a band of ˜22 kD that co-migrated with purified recombinant human GH standard purchased from Research Diagnostics Inc (Flanders, N.J.). That band was most prominent in induced cultures following overnight induction. However, a band was also observed at that molecular weight in uninduced cultures of these same strains and could also be observed with and without induction in the BOB 119 and BOB 130 control strains that carried the expression vector pCYB1 lacking the GH gene. To clarify this observation, Western Blot analyses were performed on whole cell lysates of induced cultures of strains BOB119, BOB130, BOB129, BOB133, BOB121, and BOB132. Western blots were performed with polyclonal rabbit anti-human GH antiserum purchased from United States Biological; catalogue # G 9000-11 (Swampscott, Mass.) This primary antibody was used at a 1:5000 dilution and its binding was detected with goat anti-rabbit IgG Fc conjugated to alkaline phosphatase, (product # 31341) purchased from Pierce (Rockford, Ill.). This secondary antibody was used at a 1:10,0000 dilution. Alkaline phosphatase activity was detected using the ImmunoPure® Fast Red TR/AS-MX Substrate Kit (Pierce, Rockford Ill.) according to the vendor protocols. The Western Blots clearly demonstrated presence of GH in lysates of induced cultures of BOB 129, BOB 133, BOB 121, and BOB 132 at both 3 and 16 h post-induction. In the induced culture of control strains, BOB119 and BOB130, no GH was detected by Western blot at 3 or 16 h post-induction time points.

In these preliminary experiments, the highest yields of GH were obtained from BOB132 W3110(pBBT114) in which the GH gene is fused downstream of the stII secretion signal sequence. This strain was tested further to determine if the GH protein was secreted to the periplasm as would be expected. An induced culture of BOB132 was prepared as described above and subjected to osmotic shock according to the procedure of Koshland and Botstein (Cell 20 (1980) pp. 749-760). This procedure ruptures the outer membrane and releases the contents of the periplasm into the surrounding medium. Subsequent centrifugation separates the periplasmic contents, present in the supernatant from the remainder of the cell-associated components. In this experiment, the bulk of the GH synthesized by BOB132 was found to be localized to the periplasm. This result is consistent with the finding that the bulk of the total GH is also indistinguishable in size from the purified GH standard, which indicated that the stII signal sequence had been removed. This is indicative of secretion. A larger scale (500 ml) culture of BOB132 was also induced, cultured overnight and subjected to osmotic shock according to the procedure described by Hsiung et al, 1986 (Bio/Technology 4, pp. 991-995). Gel analysis again demonstrated that the bulk of the GH produced was soluble, periplasmic, and indistinguishable in size from the GH standard. This material could also be quantitatively bound to, and eluted from, a Q-Sepharose column using conditions very similar to those described for recombinant human GH by Becker and Hsiung, 1986 (FEBS Lett 204 pp145-150).

The human GH receptor was cloned by PCR using forward primer BB3 and reverse primer BB4. BB3 has the sequence:

(SEQ ID NO: 26) 5′-CCCCGGATCCGCCACCATGGATCTCTGGCAGCTGCTGTT-3′. BB4 has the sequence:

(SEQ ID NO: 27) 5′CCCCGTCGACTCTAGAGCTATTAAATACGTAGCTCTTGGG-3′. The template was single-stranded cDNA prepared from human liver (commercially available from CLONTECH Laboratories). Primers BB3 and BB4 contain BamHI and SalI restriction sites, respectively, for cloning purposes. The 100 μl PCR reactions contained 2.5 ng of the single-stranded cDNA and 20 picomoles of each primer in 1×PCR buffer (Perkin-Elmer buffer containing MgCl2), 200 micromolar concentration of each of the four nucleotides dA, dC, dG and dT, 2.5 units of Taq polymerase (Perkin-Elmer) and 2.5 units of Pfu polymerase (Stratagene, Inc). The PCR reaction conditions were: 96° C. for 3 minutes, 35 cycles of (95° C., 1 minute; 58° C. for 30 seconds; 72° C. for 2 minutes), followed by 10 minutes at 72° C. The thermocycler employed was the Amplitron II Thermal Cycler (Thermolyne). The approximate 1.9 kb PCR product was digested with BamHI and SalI and ligated with similarly cut plasmid pUC19 (New England BioLabs). However, none of the transformants obtained from this ligation reaction contained the 1.9 kb PCR fragment. Leung et al (Nature 1987 330 pp 537-543) also failed to obtain full-length cDNA clones of the human GH receptor in pUC19. Subsequently the PCR fragment was cloned into a low copy number vector, pACYC184 (New England BioLabs) at the BamHI and SalI sites in this vector. Such clones were obtained at reasonable frequencies but E coli strains carrying the cloned PCR fragment grew poorly, forming tiny and heterogeneous looking colonies, in the presence of chloramphenicol, which is used to select for maintenance of pACYC184.

The PCR fragment was simultaneously cloned into pcDNA3.1 (+) (Invitrogen). The approximate 1.9 kb PCR product was digested with BamHI and SalI and ligated into the BamHI and XhoI cloning sites of pcDNA3.1 (+). Only infrequent transformants from this ligation contained the cloned GH receptor cDNA and all of those were found to contain deletions of segments of the receptor coding sequence. One of these clones was sequenced and found to contain a deletion of 135 bp within the GH receptor coding sequence: the sequence of the rest of the gene was in agreement with that reported by Leung et at (1987).

The rabbit GH receptor was cloned by PCR using forward primer BB3 (described above) and reverse primer BB36. BB36 has the sequence

(SEQ ID NO: 41) 5′CCCCGTCGACTCTAGAGCCATTAGATACAAAGCTCTTGGG3′ and contains XbaI and Sal I restriction sites for cloning purposes. Rabbit liver poly(A)+ mRNA was purchased from CLONTECH, Inc. and used as the substrate in first strand synthesis of single-stranded cDNA to produce template for PCR amplification. First strand synthesis of single-stranded cDNA was accomplished using a 1st Strand cDNA Synthesis Kit for RT-PCR (AMV) kit from Boehringer Mannheim Corp (Indianapolis, Ind.) according to the vendor protocols. Parallel first strand cDNA syntheses were performed using random hexamers or BB36 as the primer. Subsequent PCR reactions with the products of the first strand syntheses as templates, were carried out with primers BB3 and BB36 according to the 1st Strand cDNA Synthesis Kit for RT-PCR (AMV) kit protocol and using 2.5 units of Amplitac DNA Polymerase (Perkin-Elmer) and 0.625 units of Pfu DNA Polymerase (Stratagene). The PCR reaction conditions were 96° C. for 3 minutes, 35 cycles of (95° C., 1 minute; 58° C. for 30 seconds; 72° C. for 2 minutes), followed by 10 minutes at 72° C. The thermocycler employed was the Amplitron II Thermal Cycler (Thermolyne). The expected ˜1.9 kb PCR product was observed in PCR reactions using random hexamer-primed or BB36 primed cDNA as template. The random hexamer-primed cDNA was used in subsequent cloning experiments. It was digested with Bam H1 and XbaI and run out over a 1.2% agarose gel. This digest generates two fragments (˜365 bp and ˜1600 bp) because the rabbit GH receptor gene contains an internal Bam H1 site. Both fragments were gel-purified. Initially the 1600 bp Bam H1-XbaI fragment was cloned into pcDNA3.1(+) which had been digested with these same two enzymes. These clones were readily obtained at reasonable frequencies and showed no evidence of deletions as determined by restriction digests and subsequent sequencing. To generate a full length clone, one of the plasmids containing the 1600 bp Bam H1-Xba I fragment (pcDNA3.1(+):: rab-ghr-2A) was digested with Bam H1, treated with Calf Intestinal Alkaline Phosphatase (Promega) according to the vendor protocols, gel-purified and ligated with the gel purified ˜365 bp Bam H1 fragment that contains the 5′ portion of the rabbit GH receptor gene. Transformants from this ligation were picked and analyzed by restriction digestion and PCR to confirm the presence of the ˜365 bp fragment and to determine its orientation relative to the distal segment of the rabbit GH receptor gene. Three out of four clones analyzed were found to contain the ˜365 bp fragment cloned in the correct orientation for reconstitution of the rabbit GH receptor gene. The lack of complications in the cloning in E coli of the rabbit gene, in contrast to the human gene, is consistent with the results of Leung et al (1987) who also readily obtained full length cDNA clones for the rabbit GH receptor gene but were unable to clone a full length cDNA of the human gene in E coli. The rabbit GH receptor can be employed in assays with human GH as a ligand as it has been shown that the human GH binds the rabbit receptor with high affinity (Leung et al 1987). Plasmids containing the cloned rabbit GH receptor should be sequenced to identify a rabbit GH receptor cDNA with the correct sequence before use.

As an alternative to the rabbit receptor, a chimeric receptor could be constructed which combines the extracellular domain of the human receptor with the transmembrane and cytoplasmic domains of the rabbit receptor. Such a chimeric receptor could be constructed by recombining the human and rabbit genes at the unique Nco I site that is present in each (Leung et al 1987). Such a recombinant, containing the human gene segment located 5′ to, or “upstream” of the Nco I site and the rabbit gene segment 3′ to, or “downstream” of the Nco I site would encode a chimeric receptor of precisely the desired type, having the extracellular domain of the human receptor with the transmembrane and cytoplasmic domains of the rabbit receptor. This would allow analysis of the interaction of GH, GH muteins, and PEGylated GH muteins with the natural receptor binding site but could avoid the necessity of cloning the full length human GH receptor in E coli.

The GH muteins can be expressed in a variety of expression systems such as bacteria, yeast or insect cells. Vectors for expressing GH muteins in these systems are available commercially from a number of suppliers such as Novagen, Inc. (pET15b for expression in E. coli), New England Biolabs (pC4B1 for expression in E. coli) Invitrogen (pVL1392, pVL1393, and pMELBAC for expression in insect cells using Baculovirus vectors, Pichia vectors for expression in yeast cells, and pcDNA3 for expression in mammalian cells). GH has been successfully produced in E. coli as a cytoplasmic protein and as a secreted, periplasmic, protein using the E. coli OmpA or STII signal sequences to promote secretion of the protein into the periplasmic space (Chang et al., 1987; Hsiung et al., 1986). It is preferable that the GH muteins are expressed as secreted proteins so that they do not contain an N-terminal methionine residue, which is not present in the natural human protein. For expression in E. coli, DNA sequences encoding GH or GH muteins can be cloned into E. coli expression vectors such as pET15b that uses the strong T7 promoter or pCYB1 that uses the TAC promoter. Adding IPTG (isopropylthiogalactopyranoside, available form Sigma Chemical Company) to the growth media, can induce expression of the protein. Recombinant GH will be secreted into the periplasmic space from which it can be released by, and subsequently purified, following osmotic shock (Becker and Hsiung, 1986). The protein can be purified further using other chromatographic methods such as ion-exchange, hydrophobic interaction, size-exclusion and reversed phase chromatography, all of which are well known to those of skill in the art (e.g., see Becker and Hsiung, 1986). Protein concentrations can be determined using commercially available protein assay kits such as those sold by BioRad Laboratories (Richmond, Calif.). If the GH proteins are insoluble when expressed in E. coli they can be refolded using procedures well known to those skilled in the art (see Cox et al., 1994 and World patent applications WO9422466 and WO9412219).

Alternatively, the proteins can be expressed in insect cells as secreted proteins. The expression plasmid can be modified to contain the GH signal sequence to promote secretion of the protein into the medium. The cDNAs can be cloned into commercially available vectors, e.g., pVL1392 from Invitrogen, Inc., and used to infect insect cells. The GH and GH muteins can be purified from conditioned media using conventional chromatography procedures. Antibodies to rhGH can be used in conjunction with Western blots to localize fractions containing the GH proteins during chromatography. Alternatively, fractions containing GH can be identified using ELISA assays.

The cysteine-added GH variants also can be expressed as intracellular or secreted proteins in eukaryotic cells such as yeast, insect cells or mammalian cells. Vectors for expressing the proteins and methods for performing such experiments are described in catalogues from various commercial supply companies such as Invitrogen, Inc., Stratagene, Inc. and ClonTech, Inc. The GH and GH muteins can be purified using conventional chromatography procedures.

Biological activity of the GH muteins can be measured using a cell line that proliferates in response to GH. Fuh et al. (1992) created a GH-responsive cell line by stably transforming a myeloid leukemia cell line, FDC-P1, with a chimeric receptor comprising the extracellular domain of the rabbit GH receptor fused to the mouse G-CSF receptor. This cell line proliferates in response to GH with a half maximal effective concentration (EC50) of 20 picomolar. A similar cell line can be constructed using the published sequences of these receptors and standard molecular biology techniques (Fuh et al., 1992). Alternatively, the extracellular domain of the human GH receptor can be fused to the mouse G-CSF receptor using the published sequences of these receptors and standard molecular biology techniques. Transformed cells expressing the chimeric receptor can be identified by flow cytometry using labeled GH, by the ability of transformed cells to bind radiolabeled GH, or by the ability of transformed cells to proliferate in response to added GH. Purified GH and GH muteins can be tested in cell proliferation assays using cells expressing the chimeric receptor to measure specific activities of the proteins. Cells can be plated in 96-well dishes with various concentrations of GH or GH muteins. After 18 h, cells are treated for 4 hours with 3H-thymidine and harvested for determination of incorporated radioactivity. The EC50 can be determined for each mutein. Assays should be performed at least three times for each mutein using triplicate wells for each data point. GH muteins displaying similar optimal levels of stimulation and EC50 values comparable to or greater than wild type GH are preferable.

GH muteins that retain in vitro activity can be PEGylated using a cysteine-reactive 8 kDa PEG-maleimide (or PEG-vinylsulfone) commercially available from Shearwater, Inc. Generally, methods for PEGylating the proteins with these reagents will be similar to those described in world patent applications WO 9412219 and WO 9422466 and PCT application US95/06540, with minor modifications. The recombinant proteins must be partially reduced with dithiothreitol (DTT) in order to achieve optimal PEGylation of the free cysteine. Although the free cysteine is not involved in a disulfide bond, it is relatively unreactive to cysteine-reactive PEGs unless this partial reduction step is performed. The amount of DTT required to partially reduce each mutein can be determined empirically, using a range of DTT concentrations. Typically, a 5-10 fold molar excess of DTT for 30 min at room temperature is sufficient. Partial reduction can be detected by a slight shift in the elution profile of the protein from a reversed-phase column. Care must be taken not to over-reduce the protein and expose additional cysteine residues. Over-reduction can be detected by reversed phase-HPLC (the protein will have a retention time similar to the fully reduced and denatured protein) and by the appearance of GH molecules containing two PEGs (detectable by a molecular weight change on SDS-PAGE). Wild type GH can serve as a control since it should not PEGylate under similar conditions. Excess DTT can be removed by size exclusion chromatography using spin columns. The partially reduced protein can be reacted with various concentrations of PEG-maleimide (PEG: protein molar ratios of 1:1, 5:1, 10:1 and 50:1) to determine the optimum ratio of the two reagents. PEGylation of the protein can be monitored by a molecular weight shift using sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE). The lowest amount of PEG that gives significant quantities of mono-pegylated product without giving di-pegylated product will be considered optimum (80% conversion to mono-pegylated product is considered good). Generally, mono-PEGylated protein can be purified from non-PEGylated protein and unreacted PEG by size-exclusion or ion exchange chromatography. The purified PEGylated protein can be tested in the cell proliferation assay described above to determine its specific activity.

The above experiments will allow identification of amino acids in the B-C loop, C-D loop or N-terminal end of the A-B loop in GH that can be changed to cysteine residues, PEGylated and retain in vitro biological activity. These muteins can be tested in animal disease models well known in the art.

Experiments can be performed to confirm that the PEG molecule is attached to the protein at the proper site. This can be accomplished by proteolytic digestion of the protein, purification of the PEG peptide (which will have a large molecular weight) by size exclusion, ion exchange or reversed phase chromatography, followed by amino acid sequencing or mass spectroscopy. The PEG-coupled amino acid will appear as a blank in the amino acid sequencing run.

The pharmacokinetic properties of the PEG-GH proteins can be determined as follows, or as described in world patent application WO9422466. Pairs of rats or mice can receive an intravenous bolus injection of the test proteins. Circulating levels of the proteins are measured over the course of 24 h by removing a small sample of blood from the animals at desired time points. Circulating levels of the test proteins can be quantitated using ELISA assays. Additional experiments can be performed using the subcutaneous route to administer the proteins. Similar experiments should be performed with the non-PEGylated protein to serve as a control. These experiments will reveal whether attachment of a PEG reagent to the protein alters its pharmacokinetic properties. Covalent modification of the protein with PEG should increase the protein's circulating half-life relative to the unPEGylated protein. Larger PEG molecules and/or attachment of multiple PEG molecules should lengthen the circulating half-life longer than smaller PEG molecules.

PEG-GH proteins can be tested in rodent models of growth hormone deficiency (Cox et al., 1994) and cachexia (Tomas et al., 1992; Read et al., 1992) to determine optimum dosing schedules and demonstrate efficacy. These studies can explore different size PEG molecules, e.g., 8 and 20 kDa, and dosing schedules to determine the optimum PEG size and dosing schedule. It is expected that the larger PEG molecule will increase the circulating half-life greater than the smaller PEG molecule and will require less frequent dosing. However, large proteins potentially may have reduced volumes of distribution in vivo; thus, it is possible a 20 kDa PEG attached to GH will limit bioavailability, reducing its efficacy. Rodent models will allow determination of whether this is the case. Once the optimum dosing schedules and PEG sizes are determined, the efficacy of PEG-GH to GH can be compared in the animal models. While all PEG-GH proteins having GH activity are included in the invention, the preferred PEG-GH proteins are those that enhance growth equal or superior to GH, but which can be given less frequently. PEG-GH should be more efficacious than GH when both are administered using the less frequent dosing schedules.

One GH deficiency model that can be used is a hypophysectomized rat. GH stimulates body weight gain and bone and cartilage growth in this model (Cox et al., 1994). Hypophysectomized rats can be purchased from Charles River. Rats can be injected with GH, PEG-GH or placebo and weight gain measured daily over a 10-14 day period. At time of sacrifice, tibial epiphysis width can be determined as a measure of bone growth. Experimental methods for performing these studies are described in Cox et al. (1994).

The efficacy of PEG-GH in rodent cachexia models can be tested in a similar manner. Daily administration of dexamethasone, via osmotic pumps or subcutaneous injection, can be used to induce weight loss (Tomas et al., 1992; Read et al., 1992; PCT patent application US95/06540).

Cysteine-Added Variants of Erythropoietin

This example relates to cysteine-added variants of erythropoietin (EPO). EPO is the hormone primarily responsible for stimulating erythropoiesis or red blood cell formation. EPO acts on immature red blood cell precursors to stimulate their further proliferation and differentiation into mature red blood cells. A commercial pharmaceutical version is available from Amgen, Inc. Human EPO is a 35-39 kDa glycoprotein secreted by the adult kidney. The mature human protein contains 166 amino acids and is heavily glycosylated. The sequence of human EPO (SEQ ID NO: 2) is shown in Lin et al 1985 and Jacobs et al. 1985, which are incorporated herein by reference. The primary sequence of EPO is highly conserved among species (greater than 80% identity; Wen et al., 1994). Sugar groups account for greater than 40% of the protein's mass. Human EPO contains three N-linked glycosylation sites and one O-linked glycosylation site. The N-linked glycosylation sites are conserved in different species whereas the O-linked glycosylation site is not. The extensive glycosylation of EPO has prevented the protein's crystallization, so the X-ray structure of the protein is not known. Human EPO contains four cysteine residues. The disulfide assignments are Cys7 to Cys161 and Cys29 to Cys33. Cys33 is not conserved in mouse EPO, suggesting that the Cys29-Cys33 disulfide bond is not critical to mouse EPO's structure or function. This conclusion also seems to hold for human EPO (Boissel et al., 1993).

The amino acid sequence of EPO is consistent with the protein being a member of the GH supergene family and mutational studies support this view of EPO's structure (Boissel et al., 1993; Wen et al., 1994). A model of the three dimensional structure of EPO, modeled after the GH structure has been proposed (Boissel et al., 1993; Wen et al., 1994). Amino acids in EPO important for receptor binding have been identified through mutagenesis experiments and reside primarily in the N-terminal half of presumptive helix A and the C-terminal half of presumptive helix D (Boissel et al., 1993; Wen et al., 1994; Matthews et al., 1996). Only a single cell surface receptor for EPO has been identified (D'Andrea et al., 1989). It is believed that EPO dimerizes its receptor in much the same way that GH dimerizes its receptor (Matthew's et al., 1996).

Human EPO contains three sites for N-linked glycosylation (asparagine-24, -38 and -83) and one site for O-linked glycosylation (serine-126). The N-linked glycosylation sites are located in the A-B and B-C loops and the O-glycosylation site is located in the C-D loop. The N-linked glycosylation sites are conserved among species whereas the O-linked glycosylation site is absent in rodent EPO (Wen et al., 1993). A non-O-linked glycosylated human variant containing methionine at position 126 has been described (U.S. Pat. No. 4,703,008). The N-linked sugar groups are heavily branched and contain terminal sialic acid residues (Sasaki et al., 1987; Takeuchi et al., 1988). N-38 and N-83 contain the most highly branched oligosaccharides (Sasaki et al., 1988).

The terminal sialic residues on EPO are critical for the protein's in vivo function because removal of these residues by digestion eliminates in vivo activity (Fukada et al., 1989; Spivak and Hogans, 1989). Loss of activity correlates with faster clearance of the asialated protein from the body. The circulating half-life of the asialated protein in rats is less than ten minutes, in contrast to that of the sialated protein, which is approximately 2 hr (Fukada et al., 1989; Spivak and Hogans, 1989). Thus, in vivo activity of EPO directly correlates with its circulating half-life.

The role of N-linked sugars in EPO's biological activities haS been better defined by mutating, individually and in combination, the three asparagine residues comprising the N-linked glycosylation sites. EPO muteins in which only a single N-linked glycosylation site was mutated, i.e., N24Q, N38Q and N83Q, were secreted from mammalian cells as efficiently as wild type EPO, indicating that N-linked glycosylation at all three sites is not required for protein secretion (N24Q indicates that the asparagine at position 24 is mutated to glutamine). In contrast, EPO muteins in which two or more N-linked glycosylation sites were mutated were secreted less efficiently than wild type EPO from mammalian cells (Yamaguchi et al., 1991; Delorme et al., 1992). Mutagenesis studies found that each of the single N-linked glycosylation site muteins had in vitro biological activities equal to or greater than wild type EPO. Thus, it was concluded that none of the N-linked glycosylation sites is essential for secretion or in vitro biological activity of EPO. In fact, removal of one of the glycosylation sites seemed to improve biological activity (Yamaguchi et al., 1991).

The in vivo biological activity of N-linked glycosylation muteins was studied by two groups. Yamaguchi et al. (1991) concluded that the N24Q and N83Q muteins had in vivo activities greater than wild type EPO, which correlated with their increased in vitro activities. These authors found that the N38Q mutein had decreased in vivo activity, about 60% of wild type EPO. N38 is the most heavily branched of the three N-linked glycosylation sites (Sasaki et al., 1988). Delorme et al. (1992) reported that mutating any of the N-linked glycosylation sites reduced in vivo biological activity by about 50%. Muteins in which two or more glycosylation sites were mutated had decreased in vivo activities in both studies.

The above studies indicate that some N-linked glycosylation is required for in vitro and in vivo activity of EPO. Individually, however, none of the three glycosylation sites is absolutely essential for activity. The N-linked sugars increase the apparent molecular weight of EPO and prolong its circulating half-life, which correlates with bioactivity. Natural EPO and EPO manufactured in mammalian cells have complex N-linked sugars containing galactose and terminal sialic acid residues. The galactose residues are recognized by specific receptors on hepatocytes and promote rapid clearance of EPO from the body unless the galactose residues are masked by the terminal sialic acid residues.

Mutagenesis studies concluded that O-linked glycosylation is not required for in vitro or in vivo function of EPO (Delorme et al., 1992). This is in keeping with the observation that rodent EPO is not O-glycosylated and with the existence of a naturally occurring human EPO variant in which serine-126 is replaced by methionine, with a corresponding lack of O-linked glycosylation. Mutagenesis of serine-126 revealed that certain amino acid changes at this site (to valine, histidine or glutamic acid) yielded EPO muteins with biological activities similar to wild type EPO, whereas other amino acid changes (to alanine or glycine) resulted in EPO molecules with severely reduced activities (Delorme et al., 1992). The effect of changing serine-126 to cysteine was not studied. The in vivo bioactivity of S126V EPO was found to be similar to wild type EPO (Delorme et al., 1992).

The requirement for complex, N-linked carbohydrates containing terminal sialic acid residues for in vivo activity of EPO has limited commercial manufacture of the protein to mammalian cells. The important functions of the sialated N-linked sugars are to prevent protein aggregation, increase protein stability and prolong the circulating half-life of the protein. The terminal sialic acid residues prolong EPO's circulating half-life by masking the underlying galactose residues, which are recognized by specific receptors on hepatocytes and promote clearance of the asialated protein. EPO can be produced in insect cells and is N-glycosylated and fully active in vitro; its activity in vivo has not been reported (Wojchowski et al., 1987).

This example provides for the design of cysteine-added EPO variants and their use in preparing conjugates using cysteine-reactive PEGs and other cysteine-reactive moieties. Certain amino acids in EPO are non-essential for biological activity and can be mutated to cysteine residues without altering the normal disulfide binding pattern and overall conformation of the molecule. These amino acids are located in the A-B loop (amino acids 23-58 of the mature protein sequence), the B-C loop (amino acids 77-89 of the mature protein sequence), the C-D loop (amino acids 108-131 of the mature protein sequence), proximal to helix A (amino acids 1-8) and distal to helix D (amino acids 153-166 of the mature protein sequence). Also contemplated as preferred sites for adding cysteine residues are at the N-terminus or C-terminus of the protein sequence. Preferred sites for cysteine substitutions are the O—-linked glycosylation site (serine-126) and the amino acids comprising the three N-linked glycosylation sites (N24, 125, T26, N38, 139, T40, N83, S84, S85). Glycosylation sites are attractive sites for introducing cysteine substitutions and attaching PEG molecules to EPO because (1) these sites are surface exposed; (2) the natural protein can tolerate bulky sugar groups at these positions; (3) the glycosylation sites are located in the putative loop regions and away from the receptor binding site (Wen et al., 1994); and (4) mutagenesis studies indicate these sites (at least individually) are not essential for in vitro or in vivo activity (Yamaguchi et al., 1991; Delorme et al., 1992). As discussed above, the local conformation of the region encompassing the O-glycosylation site region seems to be important for biological activity. Whether a cysteine substitution at position 126 affects biological activity has not been studied. The cysteine-29 to cysteine-33 disulfide bond is not necessary for biological activity of EPO because changing both residues to tyrosine simultaneously yielded a biologically active EPO protein (Boissel et al., 1993; Wen et al., 1994). A “free” cysteine can be created by changing either cysteine-29 or cysteine-33 to another amino acid. Preferred amino acid changes would be to serine or alanine. The remaining “free” cysteine (cysteine-29 or cysteine-33) would be a preferred site for covalently modifying the protein with cysteine-reactive moieties.

Bill et al. (1995) individually substituted cysteine for N24, N38 and N83 and reported that the muteins had greatly reduced in vitro biological activities (less than 20% of wild type activity). Bill et al.(1995) expressed the EPO variants as fusion proteins (fused to glutathionine-S-transferase) in bacteria. One aspect of the present invention is to provide expression systems in which the N24C, N38C and N83C EPO variants will have in vitro biological activities more similar to wild type EPO.

U.S. Pat. No. 4,703,008 contemplates naturally occurring variants of EPO as well as amino acid substitutions that are present in EPO proteins of mammals. Ovine EPO contains a cysteine residue at position 88 of the polypeptide chain. The inventor is unaware of any other naturally occurring human or animal cysteine variants of EPO in which the cysteine residue occurs in the polypeptide regions disclosed herein as being useful for generating cysteine-added EPO variants. U.S. Pat. No. 4,703,008 specifically teaches away from cysteine-added EPO variants by suggesting that expression of EPO might be improved by deleting cysteine residues or substituting naturally occurring cysteine residues with serine or histidine residues.

The mature protein form of EPO can contain 165 or 166 amino acids because of post-translational removal of the C-terminal arginine. Asp-165 is the C-terminus of the 165 amino acid form and Agr-166 is the C-terminal amino acid of the 166 amino acid form. The cysteine substitution and insertion mutations described herein can comprise either the 165 or 166 amino acid forms of mature EPO.

A cDNA encoding EPO can be cloned using the polymerase chain reaction (PCR) technique from the human HepG2 or Hep3B cell lines, which are known to express EPO when treated with hypoxia or cobalt chloride (Wen et al., 1993) and are available from the American Type Culture Collection (ATCC). Cysteine mutations can be introduced into the cDNA by standard phage, plasmid or PCR mutagenesis procedures as described for GH. As described above, the preferred sites for introduction of cysteine substitution mutations are in the A-B loop, the B-C loop, the C-D loop and the region proximal to helix A and distal to helix D. The most preferred sites in these regions are the N- and O-linked glycosylation sites: S126C; N24C; 125C, T26C; N38C; 139C, T40C; N83C, S 84C and S85C. Other preferred sites for cysteine substitution mutagenesis are in the A-B loop, the B-C loop and the C-D loop, amino acids surrounding the glycosylation sites and the region of the protein proximal to helix A and distal to helix D (Boissel et al., 1993; Wen et al., 1994). Other preferred sites for cysteine substitutions in these regions are: A1, P2, P3, R4, D8, S9, T27, G28, A30, E31, H32, S34, N36, D43, T44, K45, N47, A50, K52, E55, G57, Q58, G77, Q78, A79, Q86, W88, E89, T107, R110, A111, G113, A114, Q115, K116, E117, A118, S120, P121, P122, D123, A124, A125, A127, A128, T132, K154, T157, G158, E159, A160, T163, G164, D165 and R166. Cysteine residues also can be introduced proximal to the first amino acid of the mature protein, i.e., proximal to A1, or distal to the final amino acid in the mature protein, i.e., distal to D165 or R166. Other variants in which cys-29 or cys-33 have been replaced with other amino acids, preferably serine or alanine, also are provided.

Wild type EPO and EPO muteins can be expressed using insect cells to determine whether the cysteine-added muteins are biologically active. DNAs encoding EPO/EPO muteins can be cloned into the Baculovirus expression vector pVL1392 (available from Invitrogen, Inc. and Sigma Corporation (St. Louis, Mo.) and used to infect insect cells. Recombinant Baculoviruses producing EPO can be identified by Western blots of infected insect cell conditioned media using polyclonal anti-human EPO antiserum (available from R&D Systems). The secreted EPO mutein proteins can be purified by conventional chromatographic procedures well known to those of skill in the art. Protein concentrations can be determined using commercially available protein assay kits or ELISA assay kits (available from R&D Systems and Bio-Rad Laboratories).

Purified EPO and EPO muteins can be tested in cell proliferation assays using EPO-responsive cell lines such as UT7-epo (Wen et al., 1994) or TF1 (available from the ATCC) to measure specific activities of the proteins. Cells can be plated in 96-well microtiter plates with various concentrations of EPO. Assays should be performed in triplicate. After 1-3 days in culture, cell proliferation can be measured by 3H-thymidine incorporation, as described above for GH. The concentration of protein giving half-maximal stimulation (EC50) can be determined for each mutein. Assays should be performed at least three times for each mutein, with triplicate wells for each data point. EC50 values can be used to compare the relative potencies of the muteins. Alternatively, cell proliferation in response to added EPO muteins can be analyzed using an MTT dye-exclusion assay (Komatsu et al. 1991). Proteins displaying similar optimal levels of stimulation and EC50 values comparable to or greater than wild type EPO are preferable.

The above studies confirm identification of amino acid residues in EPO that can be changed to cysteine residues and retain biological activity. Muteins that retain activity can be PEGylated using a cysteine-reactive 8 kDa PEG-maleimide as described above for GH muteins. Wild type EPO should be used as a control since it should not react with the cysteine-reactive PEG under identical partial reduction conditions. The lowest amount of PEG that gives significant quantities of mono-PEGylated product without giving di-PEGylated product should be considered optimum. Mono-PEGylated protein can be purified from non-PEGylated protein and unreacted PEG by size-exclusion or ion exchange chromatography. The purified PEGylated proteins should be tested in the cell proliferation assay described above to determine their bioactivities.

One or more of the PEGylated EPO muteins that retain in vitro bioactivity, are candidates for testing in animal disease models. PEGylation of the protein at the proper amino acid can be determined as described for GH.

In vivo testing of PEGylated EPO muteins expressed using insect cells may require that they be re-engineered for expression in mammalian cells to ensure proper glycosylation. PEG-EPO candidates produced using insect cells can be tested in the animal models described below to determine if they are active in vivo and whether they are as active as PEG-EPO produced using mammalian cell expression systems. For expression in mammalian cells, the EPO muteins can be subcloned into commercially available eukaryotic expression vectors and used to stably transform Chinese Hamster Ovary (CHO) cells (available from the ATCC). Sublines can be screened for EPO expression using ELISA assays. Sufficient quantities of the insect cell- and mammalian cell-produced EPO muteins can be prepared to compare their biological activities in animal anemia models.

In vivo bioactivities of the EPO muteins can be tested using the artificial polycythemia or starved rodent models (Cotes and Bangham., 1961; Goldwasser and Gross., 1975). In the starved rodent model, rats are deprived of food on day one and treated with test samples on days two and three. On day four, rats receive an injection of radioactive iron-59. Approximately 18 h later, rats are anesthetized and blood samples drawn. The percent conversion of labeled iron into red blood cells is then determined. In the artificial polycythemia model, mice are maintained in a closed tank and exposed for several days to hypobaric air. The animals are then brought to normal air pressure. Red blood cell formation is suppressed for several days. On day four or six after return to normal air pressure, mice are injected with erythropoietin or saline. Mice receive one injection per day for one to two days. One day later the animals receive an intravenous injection of labeled iron-59. The mice are euthanized 20 h later and the amount of labeled iron incorporated into red blood cells determined. EPO stimulates red blood cell formation in both models as measured by a dose-dependent increase in labeled iron incorporated into red blood cells. In both models different dosing regimens and different times of injections can be studied to determine if PEG-EPO is biologically active and/or more potent and produces longer acting effects than natural EPO.

Alpha interferon is produced by leukocytes and has antiviral, anti-tumor and immunomodulatory effects. There are at least 20 distinct alpha interferon genes that encode proteins that share 70% or greater amino acid identity. Amino acid sequences of the known alpha interferon species are given in Blatt et al. (1996). A “consensus” interferon that incorporates the most common amino acids into a single polypeptide chain has been described (Blatt et al., 1996). A hybrid alpha interferon protein may be produced by splicing different parts of alpha interferon proteins into a single protein (Horisberger and Di Marco, 1995). Some alpha interferons contain N-linked glycosylation sites in the region proximal to helix A and near the B-C loop (Blatt et al. 1966). The alpha 2 interferon protein (SEQ ID NO: 3) contains four cysteine residues that form two disulfide bonds. The cys1-cys98 disulfide bond (cys1-cys99 in some alpha interferon species such as alpha 1; SEQ ID NO: 4) is not essential for activity. The alpha 2-interferon protein does not contain any N-linked glycosylation sites. The crystal structure of alpha interferon has been determined (Radhakrishnan et al., 1996).

This example provides cysteine added variants in the region proximal to the A helix, distal to the E helix, in the A-B loop, in the B-C loop, in the C-D loop and in the D-E loop. Preferred sites for the introduction of cysteine residues in these regions of the alpha interferon-2 species are: D2, L3, P4, Q5, T6, S8, Q20, R22, K23, S25, F27, S28, K31, D32, R33, D35, G37, F38, Q40, E41, E42, F43, G44, N45, Q46, F47, Q48, K49, A50, N65, S68, T69, K70, D71, S72, S73, A74, A75, D77, E78, T79, Y89, Q90, Q91, N93, D94, E96, A97, Q100, G102, G104, T106, E107, T108, P109, K112, E113, D114, S115, K131, E132, K133, K134, Y135, S136, A139, S152, S154, T155, N156, L157, Q158, E159, S160, L161, R162, S163, K164, E165. Variants in which cysteine residues are introduced proximal to the first amino acid of the mature protein, i.e., proximal to C1, or distal to the final amino acid in the mature protein, i.e., distal to E165 are provided. Other variants in which cys-1 or cys-98 (cys-99 in some alpha interferon species) have been replaced with other amino acids, preferably serine or alanine, also are provided. Other variants in which Cys-1 has been deleted (des Cys-1) also are provided. The cysteine variants may be in the context of any naturally occurring or non-natural alpha interferon sequence, e.g., consensus interferon or interferon protein hybrids. Some naturally occurring alpha interferon species (e.g., alpha interferon-1) contain a naturally occurring “free” cysteine. In such interferon species the naturally occurring free cysteine can be changed to another amino acid, preferably serine or alanine.

This example also provides cysteine variants of other alpha interferon species, including consensus interferon, at equivalent sites in these proteins. The alignment of the alpha interferon-2 species with other known alpha interferon species and consensus interferon is given in Blatt et al. (1996). The crystal structure of alpha interferon-2 has been determined by Rhadhakrishnan et al. (1996). Lydon et al (1985) found that deletion of the first four amino acids from the N-terminus of alpha interferon did not affect biological activity. Valenzuela et al (1985) found that substitution of Phe-47 in alpha interferon-2 with Cys, Tyr or Ser did not alter biological activity of the protein. Cys-1 and Cys-98 have been changed individually to glycine and serine, respectively, without altering biological activity of the protein (DeChiara et al, 1986).

DNA sequences encoding alpha interferon-2 can be amplified from human genomic DNA, since alpha interferon genes do not contain introns (Pestka et al., 1987). The DNA sequence of alpha interferon-2 is given in Goeddel et al. (1980). Alternatively, a cDNA for alpha interferon-2 can be isolated from human lymphoblastoid cell lines that are known to express alpha interferon spontaneously or after exposure to viruses (Goeddell et al., 1980; Pickering et al., 1980). Many of these cell lines are available from the American Type Culture Collection (Rockville, Md.). Specific mutations can be introduced into the alpha interferon sequence using plasmid-based site-directed mutagenesis kits (e.g., Quick-Change Mutagenesis Kit, Stratagene, Inc.), phage mutagenesis strategies or employing PCR mutagenesis as described for GH.

Alpha interferon has been successfully produced in E. coli as an intracellular protein (Tarnowski et al., 1986; Thatcher and Panayotatos, 1986). Similar procedures can be used to express alpha interferon muteins. Plasmids encoding alpha interferon or alpha interferon muteins can be cloned into an E. coli expression vector such as pET15b (Novagene, Inc.) that uses the strong T7 promoter or pCYB1 (New England BioLabs, Beverly, Mass.) that uses the TAC promoter. Expression of the protein can be induced by adding IPTG to the growth media.

Recombinant alpha interferon expressed in E. coli is sometimes soluble and sometimes insoluble (Tarnowski et al., 1986; Thatcher and Panayotatos, 1986). Insolubility appears to be related to the degree of overexpression of the protein. Insoluble alpha interferon proteins can be recovered as inclusion bodies and renatured to a fully active conformation following standard oxidative refolding protocols (Thatcher and Panayotatos, 1986; Cox et al., 1994). The alpha interferon proteins can be purified further using other chromatographic methods such as ion-exchange, hydrophobic interaction, size-exclusion and reversed phase resins (Thatcher and Panayotaos, 1986). Protein concentrations can be determined using commercially available protein assay kits (Bio-Rad Laboratories).

If E. coli expression of alpha interferon muteins is not successful, one can express the proteins in insect cells as secreted proteins as described for GH. The proteins can be modified to contain the natural alpha interferon signal sequence (Goeddell et al., 1980) or the honeybee mellitin signal sequence (Invitrogen, Inc.) to promote secretion of the proteins. Alpha interferon and alpha interferon muteins can be purified from conditioned media using conventional chromatography procedures. Antibodies to alpha interferon can be used in conjunction with Western blots to localize fractions containing the alpha interferon proteins during chromatography. Alternatively, fractions containing alpha interferon proteins can be identified using ELISAs.

Bioactivities of alpha interferon and alpha interferon muteins can be measured using an in vitro viral plaque reduction assay (Ozes et al., 1992; Lewis, 1995). Human HeLa cells can be plated in 96-well plates and grown to near confluency at 37° C. The cells are then washed and treated for 24 hour with different concentrations of each alpha interferon preparation. Controls should include no alpha interferon and wild type alpha interferon (commercially available from Endogen, Inc., Woburn, Mass.). A virus such as Vesicular stomatitis virus (VSV) or encephalomyocarditis virus (EMCV) is added to the plates and the plates incubated for a further 24-48 hours at 37° C. Additional controls should include samples without virus. When 90% or more of the cells have been killed in the virus-treated, no alpha interferon control wells (determined by visual inspection of the wells), the cell monolayer are stained with crystal violet and absorbance of the wells read using a microplate reader. Alternatively, the cell monolayers can be stained with the dye MTT (Lewis, 1995). Samples should be analyzed in duplicate or triplicate. EC50 values (the amount of protein required to inhibit the cytopathic effect of the virus by 50%) can be used to compare the relative potencies of the proteins. Wild type alpha interferon-2 protects cells from the cytopathic effects of VSV and EMCV and has a specific activity of approximately 2×108 units/mg in this assay (Ozes et al., 1992). Alpha interferon muteins displaying EC50 values comparable to wild type Alpha Interferon are preferable.

Alpha interferon muteins that retain activity can be PEGylated using procedures similar to those described for GH. Wild type alpha interferon-2 can serve as a control since it should not PEGylate under similar conditions. The lowest amount of PEG that gives significant quantities of mono-pegylated product without giving di-pegylated product should be considered optimum. Mono-PEGylated protein can be purified from non-PEGylated protein and unreacted PEG by size-exclusion or ion exchange chromatography. The purified PEGylated proteins can be tested in the viral plaque reduction bioassay described above to determine their bioactivities. PEGylated alpha interferon proteins with bioactivities comparable to wild type alpha interferon are preferable. Mapping the PEG attachment site and determination of pharmacokinetic data for the PEGylated protein can be performed as described for GH.

In vivo bioactivities of the PEG-alpha interferon muteins can be tested using tumor xenograft models in nude mice and viral infection models (Balkwill, 1986; Fish et al., 1986). Since PEG-alpha interferon bioactivity may be species-specific, one should confirm activity of the PEGylated protein using appropriate animal cell lines in in vitro virus plaque reduction assays, similar to those described above. Next, one should explore the effects of different dosing regimens and different times of injections to determine if PEG-alpha interferon is more potent and produces longer lasting effects than non-PEGylated alpha interferon.

The novel alpha interferon-derived molecules of this example can be formulated and tested for activity essentially as set forth in Examples 1 and 2, substituting, however, the appropriate assays and other considerations known in the art related to the specific proteins of this example.

Beta interferon is produced by fibroblasts and exhibits antiviral, antitumor and immunomodulatory effects. The single-copy beta interferon gene encodes a preprotein that is cleaved to yield a mature protein of 166 amino acids (Taniguchi et al. 1980; SEQ ID NO: 5). The protein contains three cysteines, one of which, cysteine-17, is “free”, i.e., it does not participate in a disulfide bond. The protein contains one N-linked glycosylation site. The crystal structure of the protein has been determined (Karpusas et al. 1997)

This example provides cysteine-added variants at any of the three amino acids that comprise the N-linked glycosylation sites, i.e., N80C, E81C or T82C. This example also provides cysteine-added variants in the region proximal to the A helix, distal to the E helix, in the A-B loop, in the B-C loop, in the C-D loop and in the D-E loop. Preferred sites for introduction of cysteine residues in these regions are: M1, S2, Y3, N4, L5, Q23, N25, G26, R27, E29, Y30, K33, D34, R35, N37, D39, E42, E43, K45, Q46, L47, Q48, Q49, Q51, K52, E53, A68, F70, R71, Q72, D73, S74, S75, S76, T77, G78, E107, K108, E109, D110, F111, T112, R113, G114, K115, L116, A135, K136, E137, K138, S139, 1157, N158, R159, L160, T161, G162, Y163, L164, R165 and N166. Variants in which cysteine residues are introduced proximal to the first amino acid of the mature protein, i.e., proximal to M1, or distal to the final amino acid in the mature protein, i.e., distal to N166 also are provided.

These variants are produced in the context of the natural protein sequence or a variant protein in which the naturally occurring “free” cysteine residue (cysteine-17) has been changed to another amino acid, preferably serine or alanine.

The novel beta interferon-derived molecules of this example can be formulated and tested for activity essentially as set forth in Examples 1 and 2, substituting, however, the appropriate assays and other considerations known in the art related to the specific proteins of this example.

G-CSF is a pleuripotent cytokine that stimulates the proliferation, differentiation and function of granulocytes. The protein is produced by activated monocytes and macrophages. The amino acid sequence of G-CSF (SEQ ID NO: 6) is given in Souza et al. (1986), Nagata et al. (1986a, b) and U.S. Pat. No. 4,810,643 all incorporated herein by reference. The human protein is synthesized as a preprotein of 204 or 207 amino acids that is cleaved to yield mature proteins of 174 or 177 amino acids. The larger form has lower specific activity than the smaller form. The protein contains five cysteines, four of which are involved in disulfide bonds. Cysteine-17 is not involved in a disulfide bond. Substitution of cysteine-17 with serine yields a mutant G-CSF protein that is fully active (U.S. Pat. No. 4,810,643). The protein is O-glycosylated at threonine-133 of the mature protein.

This example provides a cysteine-added variant at threonine-133. This example provides other cysteine-added variants in the region proximal to helix A, distal to helix D, in the A-B loop, B-C loop and C-D loop. Preferred sites for introduction of cysteine substitutions in these regions are: T1, P2, L3, G4, P5, A6, S7, S8, L9, P10, Q11, S12, T38, K40, S53, G55, W58, A59, P60, S62, S63, P65, S66, Q67, A68, Q70, A72, Q90, A91, E93, G94, S96, E98, G100, G125, M126, A127, A129, Q131, T133, Q134, G135, A136, A139, A141, S142, A143, Q145, Q173 and P174. Variants in which cysteine residues are introduced proximal to the first amino acid of the mature protein, i.e., proximal to T1, or distal to the final amino acid in the mature protein, i.e., distal to P174 are provided. These variants are provided in the context of the natural protein sequence or a variant protein in which the naturally occurring “free” cysteine residue (cysteine-17) has been changed to another amino acid, preferably serine or alanine.

A cDNA encoding human G-CSF can be purchased from R&D Systems (Minneapolis, Minn.) or amplified using PCR from mRNA isolated from human carcinoma cell lines such as 5637 and U87-MG known to express G-CSF constitutively (Park et al., 1989; Nagata, 1994). These cell lines are available from the American Type Culture collection (Rockville, Md.). Specific mutations can be introduced into the G-CSF sequence using plasmid-based site-directed mutagenesis kits (e.g., Quick-Change Mutagenesis Kit, Stratagene, Inc.), phage mutagenesis methods or employing PCR mutagenesis as described for GH.

G-CSF has been successfully produced in E. coli as an intracellular protein (Souza et al., 1986). One can employ similar procedures to express G-CSF and G-CSF muteins. Plasmids encoding G-CSF or G-CSF muteins can be cloned into an E. coli expression vector such as pET15b (available from Novagen, Inc., Madison, Wis.) that uses the strong T7 promoter or pCYB1 (available from New England BioLabs, Beverly, Mass.) that uses the strong TAC promoter. Expression of the protein can be induced by adding IPTG to the growth media. Recombinant G-CSF expressed in E. coli is insoluble and can be recovered as inclusion bodies. The protein can be renatured to a fully active conformation following standard oxidative refolding protocols (Souza et al., 1986; Lu et al., 1992; Cox et al., 1994). Similar procedures can be used to refold cysteine muteins. The proteins can be purified further using other chromatographic methods such as ion exchange, hydrophobic interaction, size-exclusion and reversed phase resins (Souza et al., 1986; Kuga et al., 1989; Lu et al., 1992). Protein concentrations can be determined using commercially available protein assay kits (Bio-Rad Laboratories).

If E. coli expression of G-CSF or G-CSF muteins is not successful, one can express G-CSF and G-CSF muteins in insect cells as secreted proteins as described for GH. The proteins can be modified to contain the natural G-CSF signal sequence (Souza et al., 1986; Nagata et al., 1986a; Nagata et al, 1986b) or the honeybee mellitin signal sequence (Invitrogen, Inc., Carlsbad, Calif.) to promote secretion of the proteins. G-CSF and G-CSF muteins can be purified from conditioned media using conventional chromatography procedures. Antibodies to G-CSF can be used in conjunction with Western blots to localize fractions containing the G-CSF proteins during chromatography. Alternatively, fractions containing G-CSF proteins can be identified using ELISAs.

G-CSF muteins also can be expressed in mammalian cells as described for erythropoietin in Example 2.

Bioactivities of G-CSF and the G-CSF muteins can be measured using an in vitro cell proliferation assay. The mouse NFS-60 cell line and the human AML-193 cell line can be used to measure G-CSF bioactivity (Tsuchiya et al., 1986; Lange et al., 1987; Shirafuji et al., 1989). Both cell lines proliferate in response to human G-CSF. The AML-193 cell line is preferable since it is of human origin, which eliminates the possibility of a false conclusion resulting from species differences. The NFS60 cell lines proliferates in response to G-CSF with a half-maximal effective concentration (EC50) of 10-20 picomolar. Purified G-CSF and G-CSF muteins can be tested in cell proliferation assays using these cell lines to determine specific activities of the proteins, using published methods (Tsuchiya et al., 1986; Lange et al., 1987; Shirafuji et al., 1989). Cells can be plated in 96-well tissue culture dishes with different concentrations of G-CSF or G-CSF muteins. After 1-3 days at 37° C. in a humidified tissue culture incubator, proliferation can be measured by 3H-thymidine incorporation as described for GH. Assays should be performed at least three times for each mutein using triplicate wells for each data point. EC50 values can be used to compare the relative potencies of the muteins. G-CSF muteins displaying similar optimal levels of stimulation and EC50 values comparable to wild type G-CSF are preferable.

G-CSF muteins that retain activity can be PEGylated using procedures similar to those described for GH. Wild type G-CSF and ser-17 G-CSF can serve as controls since they should not PEGylate under similar conditions. The lowest amount of PEG that gives significant quantities of mono-PEGylated product without giving di-PEGylated product should be considered optimum. Mono-PEGylated protein can be purified from non-PEGylated protein and unreacted PEG by size-exclusion or ion exchange chromatography. The purified PEGylated proteins can be tested in the cell proliferation assay described above to determine their bioactivities.

The PEG site in the protein can be mapped using procedures similar to those described for GH. Pharmacokinetic data for the PEGylated proteins can be obtained using procedures similar to those described for GH.

Initial studies to demonstrate in vivo efficacy of PEG-G-CSF can be done in normal Sprague-Dawley rats, which can be purchased from Charles River. Groups of rats should receive single subcutaneous or intravenous injections of various doses of G-CSF, PEG-G-CSF or placebo. Animals should be sacrificed at daily intervals for up to a week for determination of neutrophil and total white blood cell counts in peripheral blood. Other blood cell types (platelets and red blood cells) can be measured to demonstrate cell specificity.

The efficacy of PEG-G-CSF can be tested in a rat neutropenia model. Neutropenia can be induced by treatment with cyclophosphamide, which is a commonly used chemotherapeutic agent that is myelosuppressive. G-CSF accelerates recovery of normal neutrophil levels in cyclophosphamide-treated animals (Kubota et al., 1990). Rats receive an injection of cyclophosphamide on day 0 to induce neutropenia. The animals are then be divided into different groups, which will receive subcutaneous injections of G-CSF, PEG-G-CSF or placebo. Neutrophil and total white blood cell counts in peripheral blood should be measured daily until they return to normal levels. Initially one should confirm that G-CSF accelerates recovery from neutropenia when injected daily. Next, one should explore the effects of different dosing regimens and different times of injections to determine if PEG-G-CSF is more potent and produces longer lasting effects than non-PEGylated G-CSF.

The novel GCSF-derived molecules of this example can be formulated and tested for activity essentially as set forth in Examples 1 and 2, substituting, however, the appropriate assays and other considerations known in the art related to the specific proteins of this example.

Thrombopoietin stimulates the development of megakaryocyte precursors of platelets. The amino acid sequence of TPO (SEQ ID NO: 7) is given in Bartley et al.(1994), Foster et al.(1994), de Sauvage et al.(1994), each incorporated herein by reference.

The protein is synthesized as a 353 amino acid precursor protein that is cleaved to yield a mature protein of 332 amino acids. The N-terminal 154 amino acids have homology with EPO and other members of the GH supergene family. The C-terminal 199 amino acids do not share homology with any other known proteins. The C-terminal region contains six N-linked glycosylation sites and multiple O-linked glycosylation sites (Hoffman et al., 1996). O-linked glycosylation sites also are found in the region proximal to the A helix, in the A-B loop, and at the C-terminus of Helix C (Hoffman et al., 1996). A truncated TPO protein containing only residues 1-195 of the mature protein is fully active in vitro (Bartley et al., 1994).

This example provides cysteine-added variants at any of the amino acids that comprise the N-linked glycosylation sites and the O-linked glycosylation sites. This example also provides cysteine-added variants in the region proximal to the A helix, distal to the D helix, in the A-B loop, in the B-C loop, and in the C-D loop.

Preferred sites for introduction of cysteine residues are: S1, P2, A3, P4, P5, A6, T37, A43, D45, S47, G49, E50, K52, T53, Q54, E56, E57, T58, A76, A77, R78, G79, Q80, G82, T84, S87, S88, G109, T110, Q111, P113, P114, Q115, G116, R117, T118, T119, A120, H121, K122, G146, G147, S148, T149, A155, T158, T159, A160, S163, T165, S166, T170, N176, R177, T178, S179, G180, E183, T184, N185, F186, T187, A188, S189, A190, T192, T193, G194, S195, N213, Q214, T215, S216, S218, N234, G235, T236, S244, T247, S254, S255, T257, S258, T260, S262, S272, S274, T276, T280, T291, T294, S307, T310, T312, T314, S315, N319, T320, S321, T323, S325, Q326, N327, L328, S329, Q330, E331 and G332. Variants in which cysteine residues are introduced proximal to the first amino acid of the mature protein, i.e., proximal to S1, or distal to the final amino acid in the mature protein, i.e., distal to G332 are provided. The cysteine-added variants are provided in the context of the natural human protein or a variant protein that is truncated between amino acids 147 and the C-terminus of the natural protein, G332. Variants in which cysteine residues are added distal to the final amino acid of a TPO protein that is truncated between amino acids 147 and 332 also are provided.

The novel TPO-derived molecules of this example can be formulated and tested for activity essentially as set forth in Examples 1 and 2, substituting, however, the appropriate assays and other considerations known in the art related to the specific proteins of this example.

GM-CSF stimulates the proliferation and differentiation of various hematopoietic cells, including neutrophil, monocyte, eosinophil, erythroid, and megakaryocyte cell lineages. The amino acid sequence of human GM-CSF (SEQ ID NO: 8) is given in Cantrell et al. (1985) and Lee et al (1985) both incorporated herein by reference.

GM-CSF is produced as a 144 amino acid preprotein that is cleaved to yield a mature 127 amino acid protein. The mature protein has two sites for N-linked glycosylation. One site is located at the C-terminal end of Helix A; the second site is in the A-B loop.

This example provides cysteine-added variants at any of the amino acids that comprise the N-linked glycosylation sites, i.e., N27C, L28C, S29C, N37C, E38C and T39C. This example also provides cysteine-added variants in the region proximal to the A helix, distal to the D helix, in the A-B loop, in the B-C loop, and in the C-D loop. Preferred sites for introduction of cysteine substitutions in these regions are: A1, P2, A3, R4, S5, P6, S7, P8, S9, T10, Q11, R30, D31, T32, A33, A34, E35, E41, S44, E45, D48, Q50, E51, T53, Q64, G65, R67, G68, S69, L70, T71, K72, K74, G75, T91, E93, T94, S95, A97, T98, T102, I117, D120, E123, V125, Q126 and E127. Variants in which cysteine residues are introduced proximal to the first amino acid of the mature protein, i.e., proximal to A1, or distal to the final amino acid in the mature protein, i.e., distal to E127 are provided.

The novel GM-CSF-derived molecules of this example can be formulated and tested for activity essentially as set forth in Examples 1 and 2, substituting, however, the appropriate assays and other considerations known in the art related to the specific proteins of this example.

IL-2 is a T cell growth factor that is synthesized by activated T cells. The protein stimulates clonal expansion of activated T cells. Human IL-2 is synthesized as a 153 amino acid precursor that is cleaved to yield a 133 amino acid mature protein (Tadatsugu et al., 1983; Devos et al., 1983; SEQ ID NO: 9).

The amino acid sequence of IL-2 is set forth in (Tadatsugu et al. 1983; Devos et al. 1983). The mature protein contains three cysteine residues, two of which form a disulfide bond. Cysteine-125 of the mature protein is not involved in a disulfide bond. Replacement of cysteine-125 with serine yields an IL-2 mutein with full biological activity (Wang et al., 1984). The protein is O-glycosylated at threonine-3 of the mature protein chain.

This example provides cysteine-added variants in the last four positions of the D helix, in the region distal to the D helix, in the A-B loop, in the B-C loop, and in the C-D loop. These variants are provided in the context of the natural protein sequence or a variant protein in which the naturally occurring “free” cysteine residue (cysteine-125) has been changed to another amino acid, preferably serine or alanine. Variants in which cysteine residues are introduced proximal to the first amino acid, i.e., A1, or distal to the final amino acid, i.e., T133, of the mature protein, also are provided.

The novel IL-2-derived molecules of this example can be formulated and tested for activity essentially as set forth in Examples 1 and 2, substituting, however, the appropriate assays and other considerations known in the art related to the specific proteins of this example.

IL-3 is produced by activated T cells and stimulates the proliferation and differentiation of pleuripotent hematopoietic stem cells. The amino acid sequence of human IL-3 (SEQ ID NO: 10) is given in Yang et al. (1986); Dorssers et al. (1987) and Otsuka et al. (1988) all incorporated herein by reference. The protein contains two cysteine residues and two N-linked glycosylation sites. Two alleles have been described, resulting in isoforms having serine or proline at amino acid position 8 or the mature protein.

This example provides cysteine-added variants at any of the amino acids that comprise the N-linked glycosylation sites. This example also provides cysteine-added variants in the region proximal to the A helix, distal to the D helix, in the A-B loop, in the B-C loop, and in the C-D loop. Variants in which cysteine residues are introduced proximal to the first amino acid or distal to the final amino acid in the mature protein, also are provided.

The novel IL-3-derived molecules of this example can be formulated and tested for activity essentially as set forth in Examples 1 and 2, substituting, however, the appropriate assays and other considerations known in the art related to the specific proteins of this example.

IL-4 is a pleiotropic cytokine that stimulates the proliferation and differentiation of monocytes and T and B cells. IL-4 has been implicated in the process that leads to B cells secreting IgE, which is believed to play a role in asthma and atopy. The bioactivity of IL-4 is species specific. IL-4 is synthesized as a 153 amino acid precursor protein that is cleaved to yield a mature protein of 129 amino acids. The amino acid sequence of human IL-4 (SEQ ID NO:11) is given in Yokota et al. (1986) which is incorporated herein by reference. The protein contains six cysteine residues and two N-linked glycosylation sites. The glycosylation sites are located in the A-B and C-D loops.

This example provides cysteine-added variants at any of the amino acids that comprise the N-linked glycosylation sites. This example also provides cysteine-added variants in the region proximal to the A helix, distal to the D helix, in the A-B loop, in the B-C loop, and in the C-D loop. Variants in which cysteine residues are introduced proximal to the first amino acid, H1, or distal to the final amino acid, S129, of the mature protein are provided.

The novel IL-4-derived molecules of this example can be formulated and tested for activity essentially as set forth in Examples 1 and 2, substituting, however, the appropriate assays and other considerations known in the art related to the specific proteins of this example.

IL-5 is a differentiation and activation factor for eosinophils. The amino acid sequence of human IL-5 (SEQ ID NO: 12) is given in Yokota et al. (1987) which is incorporated herein by reference. The mature protein contains 115 amino acids and exists in solution as a disulfide-linked homodimer. The protein contains both O-linked and N-linked glycosylation sites.

This example provides cysteine-added variants in the region proximal to the A helix, distal to the D helix, in the A-B loop, in the B-C loop, and in the C-D loop. Variants in which cysteine residues are added proximal to the first amino acid or distal to the final amino acid in the mature protein are provided.

The novel IL-5-derived molecules of this example can be formulated and tested for activity essentially as set forth in Examples 1 and 2, substituting, however, the appropriate assays and other considerations known in the art related to the specific proteins of this example.

IL-6 stimulates the proliferation and differentiation of many cell types. The amino acid sequence of human IL-6 (SEQ ID NO: 13) is given in Hirano et al. (1986) which is incorporated herein by reference. IL-6 is synthesized as a 212 amino acid preprotein that is cleaved to generate a 184 amino acid mature protein. The mature protein contains two sites for N-linked glycosylation and one site for O-glycosylation, at T137, T138, T142 or T143.

This example provides cysteine-added variants at any of the amino acids that comprise the N-linked glycosylation sites and at the O-linked glycosylation site. Also provided are cysteine-added variants in the region proximal to the A helix, distal to the D helix, in the A-B loop, in the B-C loop, and in the C-D loop. Variants in which cysteine residues are added proximal to the first amino acid or distal to the final amino acid of the mature protein also are provided.

The novel IL-6-derived molecules of this example can be formulated and tested for activity essentially as set forth in Examples 1 and 2, substituting, however, the appropriate assays and other considerations known in the art related to the specific proteins of this example.

IL-7 stimulates proliferation of immature B cells and acts on mature T cells. The amino acid sequence of human IL-7 (SEQ ID NO: 14) is given in Goodwin et al. (1989) which is incorporated herein by reference. The protein is synthesized as a 177 amino acid preprotein that is cleaved to yield a 152 amino acid mature protein that contains three sites for N-linked glycosylation.

This example provides cysteine-added variants at any of the amino acids that comprise the N-linked glycosylation sites. This example also provides cysteine-added variants in the region proximal to the A helix, distal to the D helix, in the A-B loop, in the B-C loop, and in the C-D loop.

The novel IL-7-derived molecules of this example can be formulated and tested for activity essentially as set forth in Examples 1 and 2, substituting, however, the appropriate assays and other considerations known in the art related to the specific proteins of this example. Variants in which cysteine residues are added proximal to the first amino acid or distal to the final amino acid of the mature protein also are provided.

IL-9 is a pleiotropic cytokine that acts on many cell types in the lymphoid, myeloid and mast cell lineages. IL-9 stimulates the proliferation of activated T cells and cytotoxic T lymphocytes, stimulates proliferation of mast cell precursors and synergizes with erythropoietin in stimulating immature red blood cell precursors. The amino acid sequence of human IL-9 (SEQ ID NO: 15) is given in Yang et al. (1989) which is incorporated herein by reference. IL-9 is synthesized as a precursor protein of 144 amino acids that is cleaved to yield a mature protein of 126 amino acids. The protein contains four potential N-linked glycosylation sites.

This example provides cysteine-added variants at any of the three amino acids that comprise the N-linked glycosylation sites. This example also provides cysteine-added variants in the region proximal to the A helix, distal to the D helix, in the A-B loop, in the B-C loop, and in the C-D loop. Variants in which cysteine residues are added proximal to the first amino acid or distal to the final amino acid of the mature protein also are provided.

The novel IL-9-derived molecules of this example can be formulated and tested for activity essentially as set forth in Examples 1 and 2, substituting, however, the appropriate assays and other considerations known in the art related to the specific proteins of this example.

The amino acid sequence of human IL-10 (SEQ ID NO: 16) is given in Vieira et al. (1991) which is incorporated herein by reference. IL-10 is synthesized as a 178 amino acid precursor protein that is cleaved to yield a mature protein of 160 amino acids. IL-10 can function to activate or suppress the immune system. The protein shares structural homology with the interferons, i.e., it contains five amphipathic helices. The protein contains one N-linked glycosylation site.

This example provides cysteine-added variants at any of the three amino acids comprising the N-linked glycosylation site. This example also provides cysteine-added variants in the region proximal to the A helix, distal to the E helix, in the A-B loop, in the B-C loop, in the C-D loop and in the D-E loop. Variants in which cysteine residues are added proximal to the first amino acid or distal to the final amino acid of the mature protein also are provided.

The novel IL-10-derived molecules of this example can be formulated and tested for activity essentially as set forth in Examples 1 and 2, substituting, however, the appropriate assays and other considerations known in the art related to the specific proteins of this example.

IL-11 is a pleiotropic cytokine that stimulates hematopoiesis, lymphopoeisis and acute phase responses. IL-11 shares many biological effects with IL-6. The amino acid sequence of human IL-11 (SEQ ID NO: 17) is given in Kawashima et al. (1991) and Paul et al. (1990) both incorporated herein by reference. IL-11 is synthesized as a precursor protein of 199 amino acids that is cleaved to yield a mature protein of 178 amino acids. There are no N-linked glycosylation sites in the protein.

This example provides cysteine-added variants in the region proximal to the A helix, distal to the D helix, in the A-B loop, in the B-C loop, and in the C-D loop. Variants in which cysteine residues are added proximal to the first amino acid or distal to the final amino acid of the mature protein also are provided.

The novel IL-11-derived molecules of this example can be formulated and tested for activity essentially as set forth in Examples 1 and 2, substituting, however, the appropriate assays and other considerations known in the art related to the specific proteins of this example.

IL-12 stimulates proliferation and differentiation of NK cells and cytotoxic T lymphocytes. IL-12 exists as a heterodimer of a p35 subunit and a p40 subunit. The p35 subunit is a member of the GH supergene family. The amino acid sequence of the p35 subunit (SEQ ID NO: 18) is given in Gubler et al. (1991) and Wolf et al. (1991) both incorporated herein by reference. p35 is synthesized as a precursor protein of 197 amino acids and is cleaved to yield a mature protein of 175 amino acids. The protein contains 7 cysteine residues and three potential N-linked glycosylation sites.

This example provides cysteine-added variants at any of the three amino acids that comprise the three N-linked glycosylation sites. This example also provides cysteine-added variants in the region proximal to the A helix, distal to the D helix, in the A-B loop, in the B-C loop, and in the C-D loop. These variants are provided in the context of the natural protein sequence or a variant protein in which the naturally occurring “free” cysteine residue has been changed to another amino acid, preferably serine or alanine. Variants in which cysteine residues are added proximal to the first amino acid or distal to the final amino acid of the mature protein also are provided.

The novel IL-12 p35-derived molecules of this example can be formulated and tested for activity essentially as set forth in Examples 1 and 2, substituting, however, the appropriate assays and other considerations known in the art related to the specific proteins of this example.

IL-113 shares many biological properties with IL-4. The amino acid sequence of human IL-13 (SEQ ID NO: 19) is given in McKenzie et al. (1993) and Minty et al. (1993) both incorporated herein by reference. The protein is synthesized as a 132 amino acid precursor protein that is cleaved to yield a mature protein of 112 amino acids. The mature protein contains 5 cysteine residues and multiple N-linked glycosylation sites. A variant in which glutamine at position 78 is deleted due to alternative mRNA splicing has been described (McKenzie et al 1993)

This example provides cysteine-added variants at any of the three amino acids comprising the N-linked glycosylation sites. This example also provides cysteine-added variants in the region proximal to the A helix, distal to the D helix, in the A-B loop, in the B-C loop, and in the C-D loop. These variants are provided in the context of the natural protein sequence or a variant sequence in which the pre-existing “free” cysteine has been changed to another amino acid, preferably to alanine or serine. Variants in which cysteine residues are added proximal to the first amino acid or distal to the final amino acid of the mature protein also are provided.

The novel IL-13-derived molecules of this example can be formulated and tested for activity essentially as set forth in Examples 1 and 2, substituting, however, the appropriate assays and other considerations known in the art related to the specific proteins of this example.

IL-15 stimulates the proliferation and differentiation of T cells, NK cells, LAK cells and Tumor Infiltrating Lymphocytes. IL-15 may be useful for treating cancer and viral infections. The sequence of IL-15 (SEQ ID NO: 20) is given in Anderson et al. (1995) which is incorporated herein by reference. IL-15 contains two N-linked glycosylation sites, which are located in the C-D loop and C-terminal end of the D helix. IL-15 encodes a 162 amino acid preprotein that is cleaved to generate a mature 114 amino acid protein.

This example provides cysteine-added variants at any of the three amino acids comprising the N-linked glycosylation sites in the C-D loop or C-terminal end of the D helix. This example also provides cysteine-added variants proximal to helix A, in the A-B loop, the B-C loop, the C-D loop or distal to helix D. Variants in which cysteine residues are added proximal to the first amino acid or distal to the final amino acid of the mature protein also are provided.

The novel IL-15-derived molecules of this example can be formulated and tested for activity essentially as set forth in Examples 1 and 2, substituting, however, the appropriate assays and other considerations known in the art related to the specific proteins of this example.

M-CSF regulates the growth, differentiation and function of monocytes. The protein is a disulfide-linked homodimer. Multiple molecular weight species of M-CSF, which arise from differential mRNA splicing, have been described. The amino acid sequence of human M-CSF and its various processed forms are given in Kawasaki et al (1985), Wong et al. (1987) and Cerretti et al (1988) which are incorporated herein by reference. Cysteine-added variants can be produced following the general teachings of this application and in accordance with the examples set forth herein.

The novel MCSF-derived molecules of this example can be formulated and tested for activity essentially as set forth in Examples 1 and 2, substituting, however, the appropriate assays and other considerations known in the art related to the specific proteins of this example.

Oncostatin M is a multifunctional cytokine that affects the growth and differentiation of many cell types. The amino acid sequence of human oncostatin M (SEQ ID NO: 21) is given in Malik et al. (1989) which is incorporated herein by reference. Oncostatin M is produced by activated monocytes and T lymphocytes. Oncostatin M is synthesized as a 252 amino acid preprotein that is cleaved sequentially to yield a 227 amino acid protein and then a 196 amino acid protein (Linsley et al., 1990). The mature protein contains O-linked glycosylation sites and two N-linked glycosylation sites. The protein is O-glycosylated at T160, T162 and S165. The mature protein contains five cysteine residues.

This example provides cysteine-added variants at either of the three amino acids comprising the N-linked glycosylation sites or at the amino acids that comprise the O-linked glycosylation sites. This example also provides cysteine-added variants proximal to helix A, in the A-B loop, in the B-C loop, in the C-D loop or distal to helix D. These variants are provided in the context of the natural protein sequence or a variant sequence in which the pre-existing “free” cysteine has been changed to another amino acid, preferably to alanine or serine. Variants in which cysteine residues are added proximal to the first amino acid or distal to the final amino acid of the mature protein also are provided.

The novel Oncostatin M-derived molecules of this example can be formulated and tested for activity essentially as set forth in Examples 1 and 2, substituting, however, the appropriate assays and other considerations known in the art related to the specific proteins of this example.

The amino acid sequence of human CNTF (SEQ ID NO: 22) is given in Lam et al.

(1991) which is incorporated herein by reference. CNTF is a 200 amino acid protein that contains no glycosylation sites or signal sequence for secretion. The protein contains one cysteine residue. CNTF functions as a survival factor for nerve cells.

This example provides cysteine-added variants proximal to helix A, in the A-B loop, the B-C loop, the C-D loop or distal to helix D. These variants are provided in the context of the natural protein sequence or a variant sequence in which the pre-existing “free” cysteine has been changed to another amino acid, preferably to alanine or serine. Variants in which cysteine residues are added proximal to the first amino acid or distal to the final amino acid of the mature protein also are provided.

The novel CNTF-derived molecules of this example can be formulated and tested for activity essentially as set forth in Examples 1 and 2, substituting, however, the appropriate assays and other considerations known in the art related to the specific proteins of this example.

The amino acid sequence of LIF (SEQ ID NO: 23) is given in Moreau et al. (1988) and Gough et al. (1988) both incorporated herein by reference. The human gene encodes a 202 amino acid precursor that is cleaved to yield a mature protein of 180 amino acids. The protein contains six cysteine residues, all of which participate in disulfide bonds. The protein contains multiple O- and N-linked glycosylation sites. The crystal structure of the protein was determined by Robinson et al. (1994). The protein affects the growth and differentiation of many cell types.

This example provides cysteine-added variants at any of the three amino acids comprising the N-linked glycosylation sites or the O-linked glycosylation site. Also provided are cysteine-added variants proximal to helix A, in the A-B loop, the B-C loop, the C-D loop or distal to helix D. Variants in which cysteine residues are added proximal to the first amino acid or distal to the final amino acid of the mature protein also are provided.

The novel LIF-derived molecules of this example can be formulated and tested for activity essentially as set forth in Examples 1 and 2, substituting, however, the appropriate assays and other considerations known in the art related to the specific proteins of this example.

All of the documents cited herein are incorporated herein by reference.

The protein analogues disclosed herein can be used for the known therapeutic uses of the native proteins in essentially the same forms and doses all well known in the art.

While the exemplary preferred embodiments of the present invention are described herein with particularity, those having ordinary skill in the art will recognize changes, modifications, additions, and applications other than those specifically described herein, and may adapt the preferred embodiments and methods without departing from the spirit of this invention.

Abdel-Meguid, S. S., Shieh, h.-S., Smith W. W., Dayringer, H. E., Violand, B. N. and Bentle, L. A. (1987) Proc. Natl. Acad. Sci. USA 84: 6434-6437. Abuchowski, A., Kazo, G. M., Verhoest, C. R., van Es, T., Kafkewitz, D., Nucci, M. L., Viau, A. T. and Davis, F. F. (1984) Cancer Biochem. Biophys. 7: 175-186. Aggarwal, B. B. (1998) Human Cytokines. Handbook for Basic and Clinical Research. Volume III. Blackwell Science, Malden, Mass. Aggarwal, B. B. and Gutterman, J. U. (1992) Human Cytokines. Handbook for Basic and Clinical Research. Volume I. Blackwell Scientific Publications, Cambridge, Mass. Aggarwal, B. B. and Gutterman, J. U. (1996) Human Cytokines. Handbook for Basic and Clinical Research. Volume II. Blackwell Science, Cambridge, Mass. Anderson, D. M., Johnson, L., Glaccum, M. B., Copeland, N. G., Gilbert, D. J., Jenkins, N. A., Valentine, V., Kirstein, M. N., Shapiro, D. N., Morris, S. W., Grabstein, K. and cosman, D. (1995) Genomics 25: 701-706. Balkwill, F. R. (1986) Methods Enzymology 119: 649-657. Bartley, T. D., Bogenberger, J. et al., (1994) Cell 77: 1117-1124. Bazan, F. (1991) Immunology Today 11: 350-354. Bazan, J. F. (1992) Science 257: 410-411. Becker, G. W. and Hsiung, H. M. (1986) FEBS Letters 204: 145-150. Bill, R. M., Winter, P. C., McHale, C. M., Hodges, V. M., Elder, G. E., Caley, J., Flitsch, S. L., Bicknell, R. Lappin, T. R. J. (1995) Biochem. Biophys. Acta 1261: 35-43. Bittorf, T., Jaster, R., Brock, J. (1993) FEBS Letters 336: 133-136. Blatt, L. M., Davis, J. M., Klein, S. B. and Taylor, M. W. (1996) Journal of Interferon and cytokine Research 16: 489-499. Boissel, J.-P., Lee, W.-R., Presnell, S. R., Cohen, F. E. and Bunn, H. F. (1993) J. Biol. Chem. 268: 15983-15993. Butler, C. A. (1996) Lehman Brothers Technical Report “A Paradigm for Success”. Cantrell, M. A., Anderson, D., Cerretti, D. P., Price, V., McKereghan, K., Tushinski, R. J., Mochizuki, D. Y., Larsen, A., Grabstein, K., Gillis, S, and Cosman, D. Proc. Natl. Acad. Sci. USA 82: 6250-6254. Cerretti, D. P. et al. (1988) Molecular Immunology 25: 761. Chang, C. N., Rey, B., Bochner, B., Heyneker, H. and Gray, G. (1987) Gene: 189-196. Cotes, P. M. and Bangham, D. R. (1961) Nature 191: 1065-1067 Cox, G. N., McDermott, M. J., Merkel, E., Stroh, C. A., Ko, S. C., Squires, C. H., Gleason, T. M. and Russell, D. (1994) Endocrinology 135: 1913-1920. Cunningham, B. C. and Wells, J. A. (1989) Science 244: 1081-1085. Cunningham, B. C., Jhurani, P., Ng, P. and Wells, J. A. (1989) Science 243: 1330-1336. Cunningham, B. C., Ultsch, M., de Vos, A. M., Mulkerrin, M. G., Clauser, K. R. and Wells, J. A. (1991) Science 254: 821-825. D'Andrea, A. D., Lodish, H. F. and Wong, G. G. (1989) Cell 57: 277-285. Davis, S., Aldrich, T. H., Stahl, N., Pan, L., Taga, T., Kishimoto, T., Ip, N. Y. and Yancopoulus, G. D. (1993) Science 260: 1805-1808. DeChiara, T. M., Erlitz, F. and Tarnowski,k, S. J. (1986) Methods Enzymology, 119: 403-15. de la Llosa, P., Chene, N. and Martal, J. (1985) FEBS Letts. 191: 211-215 Delorme, E., Lorenzini, T., Giffin, J., Martin, F., Jacobsen, F., Boone, T., Elliot, S. (1992) Biochemistry 31: 9871-9876. de Sauvage, F. J., Hass, P. E., Spencer, S. D. et al. (1994) Nature 369: 533-538. de Vos, A. M., Ultsch, M. and Kossiakoff, A. A. (1992) Science 255: 306-312. Devos, R., Plaetinck, G., Cheroutre, H., Simons, G., Degrave, W., Tavernier, J., Remaut, E. and Fiers, W. (1983) Nucleic Acids Research 11: 4307. Diederichs, K., Boone, T. and Karplus, A. (1991) Science 154: 1779-1782. Dorssers, L., Burger, H., Bot, F., Delwel, R., Van Kessel, A. H. M. G., Lowenberg, B. and Wagemaker, G. (1987) 55: 115-124. Dube, S., Fisher, J. W., Powell, J. S. (1988) J. Biol. Chem. 263: 17526-17521. Fish, E. N., Banerjee, K., Levine, H. L. and Stebbing, N. (1986) Antimicrob. Agents Chemother. 30: 52-56. Foster, D. C., Sprecher, C. A., Grant, F. J., Kramer, J. M. et al. (1994) Proc. Natl. Acad. Sci. USA 91: 13023-13027. Fukuda, M. N., Sasaki, H., Fukuda, M. (1989) Blood 73: 84-89. Goeddel, D. V., Heyneker, H. L., Hozumi, T. et al., (1979) Nature 281: 544-548. Goeddel, D. V., Yelverton, E., Ullrich, A., Heynecker, H. L., Miozzari, G., Holmes, W., Seeburg, P. H., Dull, T., Mat, L., Stebbing, N., Crea, R., Maeda, S., McCandliss, R., Sloma, A., Tabor, J. M., Gross, M., Familletti, P. C. and Pestka, S. (1980) Nature 287: 411-416. Goldwasser, E. and Gross, M. (1975) Methods Enzymology 37: 109-121. Goodson, R. J. and Katre, N. V. (1990) Biotechnology 8: 343-346. Goodwin, R. G., Lupton, S., Schmierer, A., Hjerrild, K. J., Jerzy, R., Clevenger, W., Gillis, S., Cosman, D. and Namen, A. E. (1989) Proc. Natl. Acad. Sci. USA 86: 302-306. Gough, N. M. et al., (1988) Proc. Natl. Acad. Sci. USA 85: 2623-2627. Gubler, U., Chua, A. O., Schoenhaut, D. S., Dwyer, C. M., McComas, W., Motyka, R., Nabavi, N., Wolitzky, A. G., Quinn, P. M., Familletti, P. C. and Gately, M. K. (1991) Proc. Natl. Acad. Sci. USA 88: 4143-4147. Hershfield, M. S., Buckley, R. H., Greenberg, M. L. et al., (1987) N. Engl. J. Medicine 316: 589-596. Hill, C. P., Osslund, T. D. and Eisenberg, D. (1993) Proc. Natl. Acad. Sci. USA 90:5167-5171. Hoffman, R. C., Andersen, H., Walker, K., Krakover, J. D., Patel, S., Stamm, M. R. and Osborn, S. G. (1996) Biochemistry 35: 14849-14861. Hsiung, H. M., Mayne, N. G. and Becker, G. W. (1986) Biotechnology 4: 991-995. Hercus, T. R., Bagley, C. J., Cambareri, B., Dottore, M., Woodcock, J. M., Vadas, M. A., Shannon, M. F., and Lopez, A. F. (1994) Proc. Natl. Acad. Sci. USA 91: 5838-5842. Hirano, T., Yasukawa, K., Harada, H., Taga, T., Watanabe, Y., Matsuda, T., Kashiwamura, S.-I., Nakajima, K., Koyama, K., Iwamatsu, A., Tsunasawa, S., Sakiyama, F., Matsui, H., Takahara, Y., Taniguchi, T. and Kishimoto, T. (1986) Nature 324: 73-76. Horisberger, M. A. and Di Marco, S. (1995) Pharmac. Ther. 66: 507-534. Innis, M. A., Gelfand, D. H., Sninsky, J. J. and White, T. J. (1990) PCR Protocols: A Guide to Methods and Applications. Academic Press, San Diego, Calif. Jacobs, K., Shoemaker, C., Rudersdorf, R. et al., (1985) Nature 313: 806-810. Karpusas, M., Nolte, M., Benton, C. B., Meier, W., Lipscomb, W. N. and Goelz, S. (1997) Proc. Natl. Acad. Sci. USA 94: 11813-11818. Katre, N. V. (1990) J. Immunology 144: 209-213. Katre, N. V., Knauf, M. J. and Laird, W. J. (1987) Proc. Natl. Acad. Sci. USA 84: 1487-1491. Kawasaki, E. S. et al., (1985) Science 230: 291. Kawashima, I., Ohsumi, J., Mita-Honjo, K., Shimoda-Takano, Ishikawa, H., Sakakibara, S., Miyadai, K. and Takiguchi, Y. (1991) FEBS Letts. 283: 199-202. Kingsley, D. M. (1994) Genes Dev. 8: 133-146. Komatsu, N., Nakauchi, H., Miwa, A., Ishihara, T., Eguguchi, M. et al., (1991) Cancer Research 51, 341-348. Kruse, N., Tony, H.-P. and Sebald, W. (1992) EMBO J. 11: 3237-3244. Lam, A., Fuller, F., Miller, J., Kloss, J., Manthorpe, M., Varon, S, and Cordell, B. (1991) Gene 102: 271-276. Kubota, N., Orita, T., Hattori, K., Oh-eda, M., Ochi, N. and Yamazaki, T. (1990) J. Biochem. 107: 486-492. Kuga, T., Komatsu, Y., Yamasaki, M., De4kine, S., Miyaji, H., Nishi, T., Sato, M., Yokoo, Y., Asano, M., Okabe, M., Morimoto, M. and Itoh, S. (1989) Bioch. Biophys. Res. Comm. 159: 103-111. Kunkel, T. A., Roberts, J. D. and Zakour, R. A. (1987) Methods in Enzymology 154: 367-382. Lange, B., Valtieri, M., Santoli, D., Caracciolo, D., Mavilio, F., Gemperlein, I., Griffin, C., Emanuel, B., Finan, J., Nowell, P. and Rovera, G. (1987) Blood 70: 192-199. Lee, F., Yokota, T., Otsuka, T., Gemmell, L., Larson, N., Luh, J., Arai, K.-I. and Rennick, D. (1985) Proc. Natl. Acad. Sci. USA 82: 4360-4364. Lewis, J. A. (1995) Chapter 9: Antiviral activity of cytokines. pp. 129-141. Li, C. H. (1982) Mol. Cell. Biochem. 46: 31-41. Lin, F.-K. (1996) U.S. Pat. No. 5,547,933. Lin, F.-K., Suggs, S., Lin, C.-H., et al., Proc. Natl. Acad. Sci. USA (1985) 82: 7580-7584. Linsley, P. S., Kallestad, J., Ochs, V. and Neubauer, M. (1990) 10: 1882-1890. Livnah, O., Stura, E. A., Johnson, D. L. et al., (1996) Science 273: 464-471. Lu, H. S., Clogston, C. L., Narhi, L. O., Merewether, L. A., Pearl, W. R. and Boone, T. C. (1992) J. Biol. Chem. 267: 8770-8777. Lydon, N. B., Favre, C., Bore, S., Neyret, O., Benureau, S., Levine, A. M., Seelig, G. F., Nagabhushan, T. L. and Trotta, P. P. (1985) Biochemistry 24: 4131-41. MacGillivray, M. H., Baptista, J. and Johnson, A. (1996) J. Clin. Endocrinol. Metab. 81: 1806-1809. Malik, N., Kallestad, J. C., Gunderson, N. L., Austin, S. D., Neubauer, M. G., Ochs, V., Marquardt, H., Zarling, J. M., Shoyab, M., Wei, C.-M., Linsley, P. S, and Rose, T. M. (1989) Molecular and Cellular Biology 9: 2847-2853. Martial, J., Chene, N. and de la Llosa, P. (1985) FEBS. Letts. 180: 295-299. Martial, J. A., Hallewell, R. A., Baxter, J. D. and Goodman, H. M. (1979) Science 205: 602-606. Matthews, D. J., Topping, R. S., Cass, R. T. and Giebel, L. B. (1996) Proc. Natl. Acad. Sci. USA 93: 9471-9476. McKay, D. B. (1992) Science 257: 412. McKenzie, A. N. J., Culpepper, J. A., Malefyt, R. de W., Briere, F., Punnonen, J., Aversa, G., Sato, A., Dang, W., Cocks, B. G., Menon, S., de Vries, J. E., Banchereau, J., and Zurawski, G. (1993) Proc. Natl. Acad. Sci. USA 90: 3735-3739. Meyers, F. J., Paradise, C., Scudder, S. A., Goodman, G. and Konrad, M. (1991) Clin. Pharmacol. Ther. 49: 307-313. Milburn, M. V., Hassell, A. M., Lambert, M. H., Jordan, S. R., Proudfoot, A. E., Graber, P. and Wells, T. N. C. (1993) Nature 363: 172-176. Mills, J. B., Kostyo, J. L., Reagan, C. R., Wagner, S. A., Moseley, M. H. and Wilhelm, A. E. (1980) Endocrinology 107: 391-399. Minty, A., Chalon, P., Derocq, J.-M., Dumont, X., Guilemot, J.-C., Kaghad, M., Labit, C., Leplatois, P., Liauzun, P., Miloux, B., Minty, C., Casellas, P., Loison. G., Lupker, J., Shire, D., Ferrara, P. and Caput, D. (1993) Nature 362: 248-250. Moreau, J.-F., Donaldson, D. D., Bennett, F., Witek-Giannotti, J., Clark, S. C. and Wong, G. G. (1988) Nature 336: 690-692. Mott, H. R. and Campbell, I. D. (1995) Current Opinion in Structural Biology 5: 114-121. Martial, J. A., Hallewell, R. A., Baxter, J. D. and Goodman, H. M. (1979) Science 205: 602-606. Nagata, S. (1994) in Cytokines and Their Receptors, N. A. Nicola ed., Oxford University Press, Oxford, pp. 158-160. Nagata, S., Tsuchiya, M., Asano, S., Kziro, Y., Yamazaki, T., Yamamoto, O., Hirata, Y., Kubota, N., Oh-eda, M., Nomura, H. and Ono, M. (1986a) Nature 319: 415-418. Nagata, S., Tsuchiya, M., Asano, S., Yamamoto, O., Hirata, Y., Kubota, N., Oh-eda, M., Nomura, H. and Yamazaki, T. (1986b) EMBO J. 5: 575-581. O'Reilly, D. R., Miller, L. K., Luckow, V. A. (1992) Caculovirus Expression Vectors, W.H. Freeman Co. publishers, New York. Otsuka, T., Miyajima, A., Brown, N., Otsu, K., Abrams, J., Saeland, S., Caux, C., De Waal-Malefyt, De Vries, J., Meyerson, P., Yokota, K., Gemmell, Rennick, D., Lee, F., Arai, N., Arai, K.-I. and Yokota, T. (1988) J. Immunology 140: 2288-2295. Owers-Narhi, L., Arakawa, T., Aoki, K. H., Elmore, R., Rohde, M. F., Boone, T., Strickland, T. W. (1991) J. Biol. Chem. 266: 23022-23026. Ozes, O. S., Reiter, Z., Klein, S., Blatt, L. M. and Taylor, M. W. (1992) J. Interferon Research 12: 55-59. Paonessa, G., Graziani, R., de Serio, A., Savino, R., Ciapponi, L., Lahm, A., Ssalvati, A. L., Toniatti, C. and Ciliberto, G. (1995) EMBO J. 14: 1942-1951. Park, L. S., Waldron, P. E., Friend, D., Sassenfeld, H. M., Price, V., Anderson, D., Cosman, D., Andrews, R. G., Bernstein, I. D. and Urdal, D. L. (1989) Blood 74: 56-65. Paul, S. R., Bennett, F., Calvetti, J. A., Kelleher, K., Wood, C. R., O'Hara, R. M., Leary, A. C., Sibley, B., Clark, S. C., William, D. A. and Yang, Y.-C. (1990) Proc. Natl. Acad. Sci. USA 87: 7512-7516. Pestka, S., Langer, J. A., Zoon, K. C. and Samuel, C. E. (1987) Ann. Rev. Biochem. 56: 727-777. Pickering, L. A., Kronenberg, L. H. and Stewart, W. E., II (1980) Proc. Natl. Acad. Scio. USA 77: 5938-5942. Powers, R., Garrett, D. S., March, C. J., Frieden, E. A., Gronenborn, A. M. and Clore, G. M. (1992) Science 256:1673-1677. Radhakrishnan, R., Walter, L. J., Hruka, A., Reichert, P., Trotta, P. P., Nagabhushan, T. L. and Walter, M. R. (1996) Structure 4: 1453-1463. Read, L. C., Tomas, F. M., Howarth, G. S., Martin, A. A., Edson, K. J., Gillespie, C. M., Owens, P. C. and Ballard, F. J. (1992) J. Endocrinology 133: 421-431. Redfield, C., Smith L. J., Boyd, J., Lawrence, G. M. P., Edwards, R. G., Smith, R. A. G., and Dobson, C. M. (1991) Biochemistry 30: 11029-11035. Robinson, R. C., Grey, L. M., Stauton, D., Vankelecom, H., Vernallis, A. B., Moreau, J.-F., Stuart, D. I., Heath, J. K. and Jones, E. Y. (1994) Cell 77: 1101-1116. Sasaki, H., Bothner, B., Dell, A. and Fukuda, M. (1987) J. Biol. Chem. 262: 12059-12076. Sasaki, H. Ochi, N., Dell, A. and Fukuda, M. (1988) Biochemistry 27: 8616-8626. Shaw, G., Veldman, G. and Wooters, J. L. (1992) U.S. Pat. No. 5,166,322. Shirafuji, N., Asano, S., Matsuda, S., Watari, K., Takaku, F. and Nagata, S. (1989) Exp Hematol. 17: 116-119. Silvennoinen, O. and Ihle, J. N. (1996) Signalling by the Hematopoietic Cytokine Receptors, R. G. Landes, Company, Austin, Tex. Souza, L. M., Boone, T. C., Gabrilove, J., Lai, P. H., Zsebo, K. M., Murdock, D. C., Chazin, V. R., Bruszewski, J., Lu, H., Chen, K. K., Barendt, J., Platzer, E., Moore, M. A. S., Mertelsmann, R. and Welte, K. (1986) Science 232: 61-65. Spivak, J. L., Hogans, B. B. (1989) Blood 73: 90-99. Takeuchi, M., Takasaki, S., Miyazaki, H., Takashi, D., Hoshi, S., Kochibe, N. and Kotaba, A. (1988) J. Biol. Chem. 263: 3657-3663. Tanaka, H., Satake-Ishikawa, R., Ishikawa, M., Matsuki, S, and Asano, K. (1991) Cancer Research 51: 3710-3714. Taniguchi, T., Matsui, H., Fujita, T., Takaoka, C., Kashima, N., Yoshimoto, R. and Hamuro, J. (1983) Nature 302: 305-310. Taniguchi, T., Ohno, S., Fujii-Kuriyama, Y. and Muramatsu, M. (1980) Gene 10: 11-15. Tarnowski, S. J., Roy, S. K., Liptak, R. A., Lee, D. K. and Ning, R. Y. (1986) Methods Enzymology 119:153-165. Tavernier, J., Tuypens, T., Verhee, A., Plaetinck, G., Devos, R., Van Der Heyden, J., Guisez, Y. and Oefner, C. (1995) Proc. Natl. Acad. Sci. USA 92: 5194-5198. Teh, L.-C. and Chapman, G. E. (1988) Biochem. Biophys. Res. Comm. 150: 391-398. Thatcher, D. R. and Panayotatos, N. (1986) Methods Enzymology 119: 166-177. Tomas, F. M., Knowles, S. E., Owens, P. C., Chandler, C. S., Francis, G. L., Read, L. C. and Ballard, F. J. (1992) Biochem. J. 282: 91-97. Tsuchiya, M., Asano, S., Kaziro, Y. and Nagata, S. (1986) Proc. Natl. Acad. Sci. USA 83: 7633-7637. Tsuda, E., Kawanishi, G., Ueda, M., Masuda, S, and Sasaki, R. (1990) Eur. J. Biochem. 188: 405-411. Vieira, P., de Waal-Malefyt, R., Dang, M. N., Johnson, K. E., Kastelein, R., Fiorentino, D. F., de Vries, J. E., Roncarolo, M.-G., Mosmann, T. R. and Moore, K. W. (1991) Proc. Natl. Acad. Sci. USA 88: 1172-1176. Wang, A., Lu, S.-D., and Mark, D. F. (1984) Science 224: 1431-1433. Walter, M. R., Cook, W. J., Ealick, S. E., Nagabhusan, T. L., Trotta, P. T. and Bugg, C. E. (1992) J. Mol. Biol. 224: 1075-1085. Wen, D., Boissel, J.-P. R., Tracy, T. E., Gruninger, R. H., Mulcahy, L. S., Czelusniak, J., Goodman, M. and Bunn, H. F. (1993) Blood: 1507-1516. Wen, D., Boissel, J. P., Showers, M., Ruch, B. C. and Bunn, H. F. (1994) J. Biol. Chem. 269: 22839-22846. White, B. A., (1993), Methods in Molecular Biology, volume 15: PCR Protocols: Current Methods and Applications. Humana Press, Totowa, N.J. Wojchowski, D. M., Orkin, S. H., Sytkowshi, A. J. (1987) Biochim. Biophys. Acta 910: 224-232. Wolf, S. F., Temple, P. A., Kobayashi, M., Young, D., Dicig, M., Lowe, L., Dzialo, R., Fitz, L., Ferenz, C., Hewick, R. M., Kelleher, K., Herrmann, S. H., Clark, S. C., Azzoni, L., Chan, S. H., Trinchieri, G. and Perussia, B. (1991) J. Immunology 146: 3074-3081. Wong, G. G. et al., (1987) Science 235: 1504. Wrighton, N. C., Farrell, F. X., Chang, R. et al., (1996) Science 273: 458-463. Yamaguchi, K., Akai, K., Kawanishi, G. Ueda, M., Masuda, S., Sasaki, R. (1991). J. Biol. Chem. 266: 20434-20439. Yang, Y.-C., Ciarletta, A. B., Temple, P. A., Chung, M. P., Kovacic, S., Witek-Giannotti, J. S., Leary, A. C., Kriz, R., Donahue, R. E., Wong, G. G. and Clark, S. C. (1986) Cell 47: 3-10. Yang, Y.-C., Ricciadi, S., Ciarletta, A., Calvetti, J., Kelleher, K. and Clark, S. C. (1989) Blood 74: 1880-1884. Yokota, T., Otsuka, T., Mosmann, T., Banchereau, J., DeFrance, T., Blanchard, D., De Vries, J. E., Lee F. and Arai, K.-I. (1986) Proc. Natl. Acad. Sci. USA 83: 5894-5898. Yokota, T., Coffman, R. L., Hagiwara, H., Rennick, D. M., Takebe, Y., Yokota, K., Gemmell, L., Shrader, B., Yang, G., Meyerson, P., Luh, J., Hoy, P., Pene, J., Briere, F., Spits, H., Banchereau, J., De Vries, J., Lee, F., Arai, N. and Arai, K.-I. (1987) Proc. Natl. Acad. Sci. USA 84: 7388-7392. Zurawski, S. M., Imler, J. L., and Zurawski, G. (1990) EMBO J. 9: 3899-3905. Zurawski, S. M. and Zurawski, G. (1992) EMBO J. 11: 3905-3910.