EP0746569A4 - Receptor-binding determinant from leukaemia inhibitory factor - Google Patents

Abstract

The present invention relates generally to molecules carrying one or more binding determinants for alpha -chain of human Leukaemia Inhibitory Factor binding receptor and to genetic sequences encoding same. More particularly, the present invention comtemplates a molecule which is non-naturally occurring and which comprises a carrier portion and an active portion and wherein the active portion comprises amino acid residues or chemical equivalents thereof which constitute a binding determinant for the alpha -chain of the human Leukaemia Inhibitory Factor (hLIF) binding receptor.

Description

RECEPTOR-BINDING DETERMINANT FROM LEUKAEMIA INHIBITORY FACTOR

The present invention relates generally to molecules carrying one or more bindin determin.ants for the α-chain of human Leukaemia Inhibitory Factor binding recepto and to genetic sequences encoding same.

Bibliographic details of the publications referred to in this specification are collecte at the end of the description. Sequence Identity Numbers (SEQ ID NOs.) for amin acid sequences referred to in the specification are defined following the bibliography

Throughout this specification, unless the context requires otherwise, the wor "comprise", or v.ariations such as "comprises" or "comprising", will be understood t imply the inclusion of a stated element or integer or group of elements or integer but not the exclusion of any other element or integer or group of elements o integers.

Certain aspects of the present invention relate to leukaemia inhibitory facto (hereinafter referred to as "LIF") of human (h) or murine (m) origin or t combinations thereof. Where reference is made to amino acid residues in thes molecules, the residues are numbered consecutively from the first serine residue o the mature, native protein (see Figure 4). In certain circumstances, the productio of recombinant LIF molecules may involve a fusion protein with glutathione-S transferase (GST) in an expression vector system. Agents such as thrombin may b used to cleave the GST portion of the LIF-GST molecule which can result in a additional glycine residue at position -1 (Figure 4). Murine and human LIF hybrid

(MH) referred to in the specification are defined in Table 1 and Figure 1. Th legend to Table 1 describes the nomenclature used to define the hybrid structures It should be noted that the numbering system of amino acid residues referred to i

Australian Provisional Application No. PL7102 filed on 3 February, 1993 and fro which the present application claims priority starts at the initiating methione of th immature, native protein. Accordingly, the serine at position + 1 described in th present specification corresponds to position +24 in the aforementioned provisiona application.

LIF is a glycoprotein that was originaUy purified and cloned on the basis of its abilit to induce terminal macrophage differentiation of the Ml myeloid leukaemic cell lin (Hilton et al 1988a). It has since been shown to have a variety of activities on wide range of cell types including megakaryocytes, osteoblasts, hepatocytes adipocytes, neurons, embryonal stem cells and primordial germ cells (Metcaff, 1991)

It is proposed that LIF transduces its biological signal via a multi-subunit membrane bound receptor. The receptor for LIF is a member of the hemopoietin or cytokin family of receptors, which generally have roles in cell growth and differentiatibn These receptors are characterised by their extracellular domain, which contains a least one copy of an approximately 200 amino acid hemopoietin domain (Cosman 1990). A number of prim.ai'y amino acid sequence motifs distinguish this domain including pairs of disulfide-bonded -cysteine residues .and the Trp-Ser-X-Trp-Ser moti (where X is any amino acid). The over.all second^ and tertiary fold of th hemopoietin domain is predicted to be similar in each member of the recepto f-amily (Bazan, 1990a). Like the extracellular domain of the growth hormon receptor (de Nos et al, 1992), the 200 amino acids of the haempoietin domain for 14 anti-parallel β-strands, folded into two barrel structures. The LLF recepto consists of two known subunits, both of which are members of the hemopoieti receptor family. The LLF receptor α-chain binds LIF with low-affinity and contain two hemopoietin domains (Gearing et al, 1991). The β-chain of the LLF receptor ha been identified as gpl30 (Gearing et al, 1992) which is also a component of th interleukin (IL)-6, IL-11, oncostatin M (OSM) and ciliary neurotrophic facto receptor (CΝTF) complexes.

The ligands for this family of receptors are unrelated at the primary amino aci sequence level, but secondary and tertiary structural predictions indicate that all known ligands for hemopoietin receptors have a similar overaU fold, suggestin evolution from a common ancestor (Bazan, 1990b). Members of this family o ligands form an anti-parallel, four α-helical bundle (the helices are designated A, B, C and D ordered from the N-terminus), that is characterised by one short and cwo long connecting loops (labelled by the helices they join). This h.as been shown tor several ligands that are members of this family, including granulocyte colony- stimulating factor (G-CSF) (Hill et al, 1993), granulocyte-macrophage CSF (GM- CSF) (Diederichs et al, 1991), growth hormone (GH) (de Nos et al, 1992; Abdel- Meguid et al, 1987), IL-2 (Brandhuber et al 1987), IL-4 (Powers et al, 1992) and IL-5 (Milburn et al, 1993), whose structures have been solved, either by x-ray crystallography or by ΝMR studies. The three-dimensional structure of LIF is predicted to be most similar to the structures of OSM, CΝTF, IL-6 and G-CSF, amongst others.

Whilst murine LIF (mLIF) is unable to bind to the human LLF (hLIF) receptor (hLIF-R), hLIF is able to bind to both high- and low-affinity mouse LLF receptors (mLIF-R), and is fully biologically active on mouse cells. Intriguingly, hLIF binds to both the naturally occurring soluble form of the mLIF-R α-chain, mLIF-binding protein (mLBP) (Layton et al, 1992), and the high-affinity mLIF-R on PC.13 cells with a higher affinity than mLIF, due to markedly different dissociation kinetics. Competitive displacement curves showed that unlabelled mLIF and hL F had a similar ability to complete with [¹²⁵I]mLIF for binding to mLBP, while unlabelled hLIF was consistently 1000- to 5000-fold more effective than mLIF in competing with [¹²⁵I]hLIF for binding to mLBP (Layton et al, 1992). Mouse LBP is also able to act as a competitive inhibitor of LIF binding to its cellular receptor, leading to inhibition of biological responses to LIF. Again, mLBP was an approximately 1000-fold more potent inhibitor of hLIF than mLIF in this system. Thus, at least two features of hLIF distinguish it from mLIF: first, its capacity to bind to the hLIF-R where mLIF cannot; and second, its capacity to bind to the mLIF-R with higher affinity than does mLIF. Understanding the way in which a cytokine interacts with its receptor at a molecul level is required for the rational design of agonists and antagonists to growth fact and cytokine action as well as the design of growth factor and cytokine .analogue In work leading up to the present invention, the inventors exploited the differenc in binding characteristics of murine and human LIF to identify a major determina responsible for the binding of human LIF (hLIF) to the α-chain of its receptor. Th present invention provides, for the first time, a rational approach to the generatio of a new range of therapeutics based on LIF by providing non-naturaUy occurrin molecules capable of binding to hLIF receptor. Such therapeutics may be bot proteinaceous and non-proteinaceous molecules.

Accordingly, one aspect of the present invention contemplates a non-naturall occurring molecule comprising a tertiary structure which presents a functiona binding face for the α-chain of hLIF binding receptor.

This aspect of the present invention is predicated in part on the identification of th amino acid residues on hLLF which constitute the binding face for the hLIF recepto and in particular the α-chain of the hLIF receptor. Since a major contribution to th binding face is the three dimensional arrangement of chemically interactive group on the side chains of the amino acid residues, a preferred aspect of the presen invention relates to the non-naturally occurring molecule wherein the functiona binding face comprises the chemically interactive groups of the amino acid residue which constitute the binding determinant on hLIF for the α-chain of hLLF bindin receptor. The most preferred chemically interactive groups including groups selecte from at least one of a methyl group, a hydroxyl group, an imino-nitrogen group an a positively charged nitrogen group or chemical equivalents homologues or analogue thereof. These chemical groups are in spatial arrangement such that they presen a binding face for the α-chain of the LIF binding receptor. In a most preferre embodiment, the chemical groups are presented on the amino acid residue themselves, or their chemical homologues, analogues or equivalents, which amin acid residues are selected from Gin 21 to Lys 160 of hLIF. The amino acid residue are more particularly selected from at least one of each of the following regions o hLIF:

(i) a region between A and B helices; (ii) a region between the B and C helices; (iii) the C helix; and

(iv) a region between the C and D helices.

Even more particularly, amino acids are Ser 107, His 112, Ser 113, Nal 155 and Ly 158 of hLIF or chemical homologues, .analogues or equivalents thereof. In thi embodiment of the present invention, the critical aspect of these amino acids is th chemically interactive group(s) on the side chains of each residue which, in define spatial arrangement, constitute the binding face of hLLF for the α-chain of the hLL binding receptor. It is clear, therefore, that the binding face does not need' t comprise amino acid residues but the chemical interactive groups thereon.

The non-naturally occurring molecule may be any convenient carrier as describe below including a solid support or matrix, or a non-proteinaceous molecule, or protein, polypeptide or peptide. In a most preferred embodiment, the non-naturall occurring molecule in the form of a carrier is a mammalian cytokine such as but no limited to mLIF carrying the hLIF receptor-binding face.

The non-naturally occurring molecule may not necessarily carry the exact hLI receptor-binding face of hLIF but will contain an amount of the hLLF receptor binding face of hLIF to enable the non-naturally occurring molecule to exhibit a least about 35% and more preferably at least about 50% hLLF-like activity a described below.

According to a more preferred embodiment, there is provided a non-naturall occurring chemical entity such as a proteinaceous or non-proteinaceous molecul and/or a solid or non-immobilised support, said chemical entity comprising a tertiar structure which presents one or more chemically interactive groups selected from a least one of each of a methyl group, a hydroxyl group, an imino-nitrogen group an a positively charged nitrogen group as a functional binding face for the α-chain o hLIF binding receptor. In a preferred embodiment, the non-naturally occurrin chemical entity exhibits at least about 50% hLIF-like activity determined -a described below.

Preferably, the chemicaΗy interactive groups .are presented on amino acid residue selected from one or more of Ser, His, Nal and Lys or their chemical equivalents homologues or analogues. More preferably, the chemicaΗy interactive group correspond to those on Ser 107, His 112, Ser 113, Nal 155 and Lys 158 of hLIF o chemical homologues, analogues or equivalents thereof.

Another .aspect of the present invention is directed to a molecule which is non naturally occurring and which comprises a carrier portion and an active portio wherein said active portion comprises amino acid residues or chemical equivalent thereof which constitute a binding determinant for the α-chain of the hLIF bindin receptor.

The term "molecule" is used in its broadest sense to refer to .any chemical entity o compound, .or a protein, polypeptide, or peptide or a non-peptide analogue an which is capable of carrying a sufficient number of amino acid residues or thei chemical equivalents to constitute an effective binding determinant for the α-chai of the hLIF binding receptor. A "sufficient number" of amino acid residues is th minimum number required to exhibit at least 35% hLIF-like activity in terms of th ability to compete for ¹²⁵I-hLIF binding to murine LIF-binding protein (mLBP) whic is determined as described below. A "chemical equivalent" includes active an interactive groups present on a particular amino acid residue and, hence, extends t non-peptide or non-amino acid mimics of the amino acid residues involved in th binding face.

The "molecule" is considered a "carrier" of the amino acid residues or their analogue which constitute all or a functional part of a binding determinant for the α-chain o the hLIF binding receptor. The selection of an appropriate carrier molecule i determined by its tertiary structure. The amino acid residues or their analogues ar required to form a receptor binding face in the tertiary structure of the carrie molecule. The present invention provides, therefore, a carrier molecule having tertiary structure such that a sufficient number of amino acid residues or thei analogues including chemical equivalents, constituting a binding determinant for th α-ch n of the hLIF binding receptor, when introduced into s.aid carrier molecule t form a receptor-binding face capable of at least 35% hLLF-like activity.

Preferably, the carrier molecule is a protein, polypeptide or peptide or a molecul substantially consisting of amino acid residues linked via peptide bonds. Th molecules may also be engineered to contain non-naturally occurring amino aci residues such as some D-amino acids, non-naturally occurring substituted amino acid or chemical equivalents of amino acids which direct or otherwise influence th terti.ary structure of the molecule. For example, a carrier molecule such as a protein polypeptide or protein may be engineered to contain an amino acid, chemica derivative thereof or other chemical entity e.xhibiting conf ormational restraints on th overall molecule or a part thereof. Molecules exhibiting pre-determined tertiar conformations are particularly useful in their selection as carrier molecules i accordance with the present invention.

In a preferred embodiment, the carrier molecule is a mammalian cytokine such as but not limited to a non-human LIF, a colony stimulating factor (CSF) (e.g. G-CSF, GM-CSF, growth hormone), an interleukin (e.g. IL-2, IL-4, IL-6), oncostatin M (OSM), ciliary neurotrophic factor (CNTF) or other factor (e.g. a haemopoieti growth factor, interferon). Preferred non-human LIF molecules are those fro livestock animals (e.g. from sheep, cattle, donkeys, horses, deer, goats and pigs) laboratory test animals (e.g. mice, rats and guinea pigs), companion animals (e.g. dogs and cats) or captive wild animals (e.g. kangaroos, foxes, dingos, wild boar an emus). The preferred non-human LIF is murine LIF. The preferred cytokine is G- CSF.

It is not a requirement of the present invention nor is the invention so limited tha every amino acid residue forming the receptor binding face for the α-chain of hLIF binding receptor on hLIF be existent on the carrier molecule. It is a requirement however, that a sufficien number of the residues be present in sufficien conformational proximity such that the receptor-binding face is formed exhibiting a least about 35% hLIF-like activity. Furthermore, the amino acids present in hLI forming the receptor-binding determinant may be substituted by one or mor functionally equivalent amino acids or chemical derivatives thereof such that recepto binding activity is not completely lost. Accordingly, the present invention extends t a carrier molecule carrying amino acid residues or their chemical equivalents o derivatives or amino acid equivalents, homologues or analogues of the .amino acid in hLIF which constitute the binding determinant for the α-chain of the hLLF bindin receptor, wherein the number of amino acid residues and their proximity ar sufficient for the molecule to exhibit at least about 35% hLIF-like activity and mor preferably at least about 50% hLIF-like activity.

The .amino acid residues constituting the binding determinant are selected from Gi 48 to Lys 160 of hLIF using the numbering system depicted in Figure 4. Preferably, at least one amino acid residue is selected from each of the following regions o hLIF: (i) a region between the A and B helices; (ii) a region between the B and C helices; (iii) the C helix; and (iv) a region between the C and D helices.

Even more preferably, the amino acid residues are selected from one or more of a Ser, His, Nal and/or Lys or chemical equivalents thereof including chemically interactive groups thereon. The actual binding determin.ant on LLF comprises Ser 107, His 112, Ser 113, Nal 155 and Lys 158 which, in a most preferred embodiment, forms the same arrangement or a functional equivalent arranged in the non-naturally occurring molecule. These amino acids are arranged in the carrier molecule such that at least two and more preferably at least three and even more preferably at least four residues are non-contiguous relative to each other. Reference is made herein to the subject molecule having at least about 35% hLLF like activity. More preferably, the molecule has at least about 50% hLLF-like activit and even more preferably at least about 70% hLIF-like activity. hLLF-like activit is defined in terms of the ability for a molecule to compete for ¹²⁵I-hLIF binding t mLBP based on the equation:

log x_m - log x_c

100 = % hLIF-like activity log x_m - log x_h

wherein ^ is the dose of unlabelled mLIF; X_jj is the dose of unlabelled hLLF; _p i the dose of the molecule (e.g. protein, polypeptide or peptide) required to give 50% inhibition of ¹²⁵I-hLIF binding to mLBP.

Accordingly, a preferred embodiment of the present invention contemplates a hybrid molecule comprising:

(i) a polypeptide backbone having a tertiary structure;

(ii) .amino acid residues constituting a binding determinant for the α chain of hLIF binding receptor inserted into said polypeptide backbone such that the arrangement of said amino acid residues in the tertiary structure of the polypeptide provides a hLIF receptor binding face; and wherein said hybrid molecule exhibits at least about 50% hLLF-like activity as defined by the equation:

log x_m - log x_c

≥ 0.50 log x_m - log x.

wherein ^ is a dose of unlabelled mLIF; _j, is a dose of unlabelled hLLF; and x,. is a dose of hybrid molecule required to give 50% inhibition of ¹²⁵I-hLLF binding to murine LIF binding protein (mLBP).

With respect to this embodiment, the molecule is a hybrid between a carrier molecule in the form of a polypeptide and the amino acid residues forming the binding determinant for the α-chain of the hLIF-binding receptor. Preferably, the polypeptide is a mammalian cytokine as hereinbefore described. Preferably, the amino acid residues are selected from hLIF as hereinbefore described. This definition is also applicable when the carrier molecule anu/or the binding interface are non-proteinaceous in a form such as a chemical compound capable of mimicking a polypeptide or peptide.

This embodiment of the present invention is conveniently elucidated by the following diagram which shows the competitive inhibition of mLIF, hLLF .and a hybrid LLF peptide or polypeptide with iodinated hLIF for the mLIF receptor:

X, X. ^km

A "peptide" is taken herein to be a molecule comprising ten or less amino acid residues. A "polypeptide" is a molecule comprising eleven or more amino acid residues and includes a protein. The peptides and polypeptides are preferably recombinant or synthetic molecules but extend to fragments or parts of naturally occurring LIF molecules and to hybrid molecules comprising regions from two or more LIF molecules from different species. Preferably, the peptides and polypeptides are "isolated" meaning that they have undergone at least one step toward biological purity. A peptide or polypeptide is deemed herein to be "isolated" or "biologically pure" where a preparation comprises at least 35%, preferably at least 45%, more preferably at least 55-60%, still more preferably at least 65-75% and even more preferably at least 80-90% of the peptide or polypeptide relative to other components in the preparation as determined by weight, LIF receptor-binding activity, amino acid content or other convenient means. The present invention is particularly exemplified using a soluble murine LIF-binding protein (mLBP) (Layton et al, 1992) or a recombinant human LLF receptor which are convenient since human LIF readily binds or otherwise associates with these molecules. It is not intended, nowever, for the present invention to be limited solely to the murine receptor or to the elucidation of the binding determinant solely on the human LIF molecule.

As stated above, the molecules of the present invention are useful inter alia in rendering a LIF molecule from one species more like a LIF molecule from .another species. The hybrid LEF molecule is also useful in designing drug .analogues, agonists and antagonists. Preferably, the hybrid LLF molecule is between non-human LIF and human LLF and most preferably between murine LIF and human LIF. In accordance with these preferred embodiments, the hybrid LIF molecule is s d to be a "humanised" form of non-human LIF and, in such a case, the humanised non- human LIF molecule may be more efficacious in humans and/or may provoke no or a reduced immune response compared to the non-human LLF molecule per se.

The present invention is particularly useful in designing .analogues, antagonists and agonists based on the chemically interactive groups which constitute the binding face on hLIF for. hLIF receptor. This permits a rational design of such molecules which can be readily assayed in vitro and in vivo. The present invention extends to aU such analogues, antagonists and/or agonists capable of binding or otherwise interacting with the hLIF receptor binding face on a non-naturally occurring molecule, chemical entity or hybrid molecule as hereinbefore described.

Another embodiment of the present invention is directed to genetic sequences comprising a sequence of nucleotides which encode or are complementary to nucleotide sequences which encode the peptides and polypeptides encompassed by the present invention. The genetic sequences may be cDNA or mRNA .and may be single or double stranded, linear or covalently closed, circular molecules. Conveniently, the genetic molecules are part of an expression vector capable of expression in a prokaryotic cell (e.g. E. colt) or a eukaryotic cell (e.g. an animal or ma.mmali.an cell).

Tixe present invention further contemplates a method for mapping LIF receptor binding activity associated with a mammalian LLF to one or more amino aci residues or to a sequence of amino acid residues on said LLF, said metho comprising substituting for regions on said mammalian LLF, structurally simila regions from a LIF from a different species of mammal wherein both LIF molecule before substitution differ in respect of the activity to be mapped, assaying th substituted LIF molecules so obtained and identifying regions on one or both of sai LIF molecules responsible for LIF receptor-binding activity.

Preferably, the substitutions are based on secondary structural predictions (se gener.ally Bazan, 1991) such that the hybrids so constructed differ in the activity't be mapped. Preferably also, the substitution of regions are at the DNA level whic are then expressed to form recombinant hybrid LIF molecules.

This aspect of the present invention is exemplified .and described hereinafter b determining the LIF receptor-binding determinant on human LLF using a series o mouse-human LIF hybrid molecules some of which retain hLIF-like activity a previoxisly described. The present invention, however, extends to the use of LI molecules from other mammalian species.

The preferred hybrids comprise a murine LIF carrier molecule and a sufficient o effective number of amino acid residues from hLIF to constitute a receptor bindin face for the α-chain of hLIF-binding receptor. Even more preferred amino aci residues or peptides are those which carry the hLIF binding determinant in MH5, MH6, MH16, MH17, MH18, MH31, MH19, MH20, MH21, MH22, MH23, MH24, MH25, MH29, MH30, MH32, MH33, MH34, MH35, MH36, MH37, MH38, MH39, MH40 and MH41. Still more preferred amino acids or peptides are those in MH33, MH35, MH36, MH37, MH38, MH40 and MH41. The most preferred amino acids or peptides are those in MH33, i.e. Ser 107, His 112, Ser 113, Nal 155 and Lys 158. The present invention is further described by the following non-limiting Figures an Examples.

The following single and three letter abbreviations for amino acid residues are use in the specification:

Amino Acid Three-letter One-letter Abbreviation Symbol

Alanine Ala A

Arginine Arg R

Asparagine Asn Ν Aspartic acid Asp D

Cysteine Cys C

Glutamine Gin Q

Glutamic acid Glu E

Glycine Gly G Histidine His H

Isoleucine He I

Leucine Leu L

Lysine Lys K

Methionine* Met M Phenylalanine Phe F

Proline Pro P

Serine Ser S

Threonine Thr T

Tryptophan Trp w Tyrosine Tyr Y

N.aline Nal N The following other abbreviations are used in the specification:

LIF Leukaemia Inhibitory Factor

LIF-R LIF receptor m murine h human

P porcine

LBP LIF binding protein

In the Figures:

Figure 1 is a summary of data from biological assays and competitive binding assay of mLIF, hLIF and chimeric LLF proteins. Mouse LLF sequence is represented b blue boxes and hLIF sequence is represented by yellow boxes. The hybrid protein are numbered from 1 to 41 and are prefixed by "MH" to denote that they contai both mLIF and hLIF amino acid sequence. The relative positions of the predicte α-helices and connecting loops in the hybrid proteins can be determined by visua alignment of the schematic representations with the diagr.am immediately below The percentage human score of mLIF, hLIF and hybrid proteins is calculated fro the ID_ζQ for inhibition of [¹²⁵I]hLIF binding to mLIF-binding protein. Most assay were performed two or more times with essentially identical results. N.D., no determined. The exact amino acid construction of each hybrid according to SEQ I NO. 2 is indicated in Table 1.

Figure 2 (A) Competitive inhibition of [¹²⁵I]mLIF binding to mLBP (soluble mLLF- α-chain) by mLIF, hLIF and a selection of m-hLIF hybrids, (•) mLIF; (o) hLIF ( A) MH1; (Δ) MH2; (B) MH3; (D) MH4; ( ♦) MH5; (0) MH6; ( T) MH14. Result for all competition assays were expressed as the number of counts bound to th receptor at a particular concentration of unlabelled competitor (B) divided by th number of counts bound to the receptor when no unlabelled competitor was presen (B₀). (B) Competitive inhibition of [¹²⁵I]hLIF binding to mLBP by mLLF, hLIF an a selection of m-hLIF hybrids. (•) mLIF; (o) hLIF; ( A) MH1; (Δ) MH2. (C Competitive inhibition of [¹²⁵I]hLIF biding to mLBP by mLIF, hLIF .and a selection of m-HLEF hybrids, (β) mLIF; (o) hLLF; (B) MH3; ( G) MH4; ( ♦) MH5; ( 0) MH6; ( τ) MH14.

Figure 3 (A) Competitive inhibition of [¹²⁵I]hLLF binding to COS cells transfected with the hLIF-R α-chain by mLLF, hLLF .and a selection of m-hLIF hybrids, (•) mLLF; (o) hLEF; ( ι) MHl; (Δ) MH2; (D) MH4; (♦) MH5; (0) MH6; (D) MH14. (B) Competitive inhibition by normal mouse serum of Ml colony differentiation induced by 200 U/ml mLIF, hLIF or m-hLIF hybrids, (•) mLLF; (o) hLIF; (Δ) MH2; (D) MH4; ( ♦) MH5; (0) MH6.

Figure 4 Comparison of mouse, human and porcine LIF and G-CSF amino acid sequences. Amino acid residues that are identical between mLIF .and hLLF are indicated by asterisks and predicted α-helices A, B, C and D in hLIF are marked within boxes. Numbering of the amino acid residues of each protein is indicated as starting at the first serine of the mature protein. The first glycine is -1 and is derived by thrombin cleavage of GST-LIF fusion. The serine + 1 is equiv.alent to serine +24 of the immature native protein.

The figure also shows the alignment of LIF proteins to hG-CSF and the secondary structure assignments for G-CSF. (H, α-helrx; G, 3₁₀-helix turn; T, turn; S, bend).

Figure 5 is a graphical representation showing:

(A) competitive inhibition of ¹²⁵I-mLIF binding to mLBP (soluble mLLF-R α- chain) by mLIF (•), hLIF (o), MH33 ( A) and MH2 (Δ);

(B) competitive inhibition of ¹²⁵I-hLIF binding to mLBP by mLLF (β), hLIF (o), MH33 ( A) and MH2 ( Δ); and

(C) competitive inhibition of ¹²⁵I-hLIF binding to COS cells transfected with the hLEF-R α-chain by mLIF (•), hLIF (o) and MH33 ( A); results for all competition assays are expressed as the number of counts bound to the receptor at a particular concentration of unlabelled inhibitor (B) over the number of counts bound to the receptor when no unlabelled inhibitor is present (B₀). Figure 6 is a pictorial representation showing a ribbon diagram of the model of hLLF with the receptor binding site amino acid residues shown in CPK form.

EXAMPLE 1 CONSTRUCTION OF HYBRID LLF PROTEINS

Hybrid cDNAs in which portions of mouse LLF were replaced with homologous regions of human LIF were either synthesized from discrete restriction fragments of the two LIF cDNA molecules, or they were generated by the "splicing by overlap extension" method (Ho et al, 1989) which is a PCR-based technique using Pfu polymerase (Stratagene). The cDNA encoding porcine LIF was constructed from a genomic porcine LIF clone (Willson et al 1992) in which the second intron was removed using the above PCR-based technique. The mutated cDNAs were subcloned into the E. coli expression vector pGEX-2T. All constructs were sequenced in both directions using T7 DNA polymerase (Pharmacia) and a Promega dideojcy sequencing kit.

Clones designated herein as MH1 and MH2 (Table 1 and Figure 1) were constructed from plasmids pmLIF (in pGEX-2TH) and phLIF (in pGEX-2TH). Plasmid pmLIF was cut with the restriction enzymes Smal and Hindlll to produce fragments of approjcimately 257 bp and 5241 bp. Plasmid phLIF was also cut with the restriction enzymes Smal and Hindlll to produce fragments of approximately 258 bp and approximately 247 bp. The approximately 293 bp Smal-Hindlll mLLF fragment was ligated with the approximately 247 bp Smal-Hindlll hLLF fragment to give construct MH1. The approximately 258 bp Smal-Hindlll hLLF fragment was ligated with the approximately 240 bp Smal-Hindlll mLLF fragment to give construct MH2. LLF hybrid molecules .are shown in Table 1 and Figure 1. EXAMPLE 2 PROTEIN EXPRESSION

.All cDNAs were expressed in E. cυiiNM522 except for MH11 (E. coli strain TOPP4, Stratagene). Growth and induction of transformants, lysis of cells, adsorption on glutathione agarose and thrombin cleavage of fusion proteins was carried out essentially as described (Smith and Johnson, 1988; Gearing et al 1989) except that culture were induced by 50μM IPTG in exponentially growing bacteria at 30 °C rather than 37 °C. All mutants were further purified on a 10 cm x 2 cm i.d. S- Sepharose column (Pharmacia, Uppsala) with a linear gradient of 0.2 - 0.45 M NaCl in 50 mM sodium phosphate pH7.0 over 45 minutes at a flow rate of 2.5 ml/min. The purity of each hybrid was confirmed by their elution as a single peak from a 100 mm x 4.6 mm i.d. C8 RP-HPLC column (Brownlee) using a 0-100% v/v acetonitrile gradient in 0.1% v/v trifluoro acetic acid over 60 minutes at a flow rate of 1 ml/min and by electrophoresis on 15% w/v SDS-polycrylamide gels (Laemmli, 1970) in a Mini-Protean II system (Bio-Rad) followed by silver staining (Butcher and Tomkins, 1985). The concentration of protein in each preparation of purified hybrid was quantified by amino acid analysis on a Beckman amino acid analyser (model 6300) (Simpson et αl, 1986) with norleucine added as an internal standard in all samples.

EXAMPLE 3

RADIOIODINATIONOF LIGANDS

Recombinant mLIF or hLIF (1-2 μg) produced in E. coli was purified and iodinated as previously described (Hilton et αl, 1988b). The iodinated materials retained biological activity and had specific activities of 3.5-4.5 x 10⁵ c.p.m./pmol for [¹²⁵I]mLIF and 4-8 x 10⁵ c.p.m./pmol for [¹²⁵I]hLIF. EXAMPLE 4 HUMAN LIF RECEPTOR cDNA

A 4.1 kb cDNA encoding the human LIF receptor was isolated from a foetal liver cDNA library (Clonetech) by plaque hybridisation with a ³²P-radiolabelled DNA fragment corresponding to amino acids 13-177. The nucleotide sequence corresponding to the coding region of this cDNA was determined by the dideoxynucleotide chain termination method (Sanger et al, 1977) using synthetic oligonucleotide primers. With the exception of a T to C nucleotide substitution at the third position of codon 555, which does not alter the predicted amino acid sequence, the sequence of this cDNA was otherwise identical to that previously reported (Gearing et al, 1991). The LIF-R cDNA was subcloned into the m.ammalian expression vector pCDM8 (Seed, 1987) and designated pCDM8/16C. The pl.asmid pCDM8/16CP, encoding a truncated soluble form of the hLLF-R that includes both haemopoietin domains and two of the three fibronectin repeat structures, was constructed from pCDM8/16C by deletion of a Pstl fragment that codes for 97 membrane proximal residues, the transmembrane and cytoplasmic domains. The C-terminus of the soluble LIF-R encoded by pCDM8/16CP includes 10 residues encoded by vector sequences.

Transient expression of the hLIF-R in COS cells was achieved by electroporation with plasmid DNA. Briefly, 2 x 10⁷ cells were harvested in 0.8 ml of phosphate- buffered saline (PBS), mixed with 20 μg of pCDM8/16C DNA at 4 °C and subjected to electroporation at 300 N and 500 μFD. Viable cells were harvested by centrifugation through a cushion of FCS and incubated in 50 ml of medium at 37 °C in an atmosphere of 10% v/v C0₂. Seventy-two hours post-transfection, cells were detached in HRF containing 0.04 M EDTA and 0.1 mg/ml chondroitin sulfate, harvested by centrifugation and resuspended in HRF. EXAMPLE 5 BIOLOGICAL ACTIVITY OF MOUSE-HUMAN HYBRLD LLF PROTEINS

The functional activity of the m-hLLF hybrids was assayed by their ability to induce differentiation in murine Ml leukaemic colonies. Ml differentiation assays were performed as described (Metcalf et al, 1988). The specific activity of each hybrid was expressed as the units of Ml differentiation-inducing activity per milligr.am of protein.

For serum dose inhibition assays, the specific activity of each LLF preparation was determined by titration in cultures of Ml cells. Aliquots of normal mouse serum (containing approximately 5 μg/ml mLBP) were then added in serial 2-fold dilutions to cultures of Ml cells that also contained a just maximal concentration (200 U) Of mLIF, hLLF or m-hLLF hybrid. Assays including mLLF and hLLF gave identical results when either crude mouse serum or a highly purified preparation of mLBP was used. Hybrids were assessed for mLIF or hLLF character from the dilution of serum required to block 50% of their Ml cell differentiation-inducing activity.

EXAMPLE 6 BINDING OF MOUSE-HUMAN HYBRLD LIF PROTEINS TO MOUSE

LIF-BLNDING PROTEIN (mLBP)

Normal mouse serum was vised as a source of mLBP. For competitive binding experiments, 50 μl aliquots of unlabelled LIF or a mouse-human LLF hybrid were added to 96-well filtration assay plates containing a 0.65 micron Durapore membrane (Millipore, MA, USA) with 20 μl aliquots of a 1/10 to 1/20 dilution of normal mouse serum, 10 μl radiolabelled ligand and 25 μl Concanavalin-A Seph-arose 4B (ConA-Sepharose, Pharmacia, Uppsala) (diluted 1:4 in 100 mM sodium acetate pH6.0 containing 1 mM MnCl₂, 1 mM MgCl₂ and 1 mM CaCl₂; this and all subsequent buffers contained 0.02% w/v sodium azide and 0.02% v/v Tween 20) and incubated at room temperature overnight with agitation. Bound and free radioactivity were separated by vacuum filtration of the supernatant and the ConA- Sepharose pellet was washed once with 200 μl cold PBS. Assay plates containing the cell pellet were then exposed to a phosphor screen (Molecular Dynamics, CA, USA) for 16-24 hours and the results quantified using Imagequant version 3.0 software (Molecular Dynamics).

EXAMPLE 7

BINDING OF THE MOUSE-HUMAN LIF HYBRLDS TO COS CELLS TRANSFECTED WITH A hLIF-R cDNA CLONE

A cDNA encoding the hLIF-R was isolated, subcloned and expressed in COS cells as described in the above examples.

For competitive binding experiments, 20 μl aliquots of cells, resuspended in HRF at 5 x 10⁵ - 8 x 10⁵ cells/ml, were added to 96 well filtration assay plates containing a 0.65 micron Durapore membrane (Millipore, MA, USA) with 50μl unlabelled LIF or m-hLLF hybrid and lOμl ¹²⁵I-hLIF and incubated overnight on ice. Bound and free radioactivity were separated by vacuum filtration of the supernatant and the cell pellet was washed once with 200 μl cold PBS. .Assay plates containing the cell pellet were then exposed to a phosphor screen (Molecular Dynamics, CA, USA) for 2-3 days and the results quantified using Imagequant version 3.0 software (Molecular Dynamics).

EXAMPLE 8

MAPPING OF RECEPTOR-BINDING DETERMINANT ON LIF

1. Hybrids MH1 to MH17

Secondary and tertiary structural predictions have suggested that the LLF molecule is an anti-parallel, four α-helical bundle, a topology common to a number of growth factors and cytokines (Diedrichs et al, 1991; Parry et al, 1991; de Nos et al, 1992). The model of Bazan (1991) was used to divide the LIF amino acid sequence into a series of modules of predicted α-helices and connecting loops. A series of plasmids was then designed to encode mouse-human chimeric LLF (m-hLLF) molecules (Figure 1) in which regions of hLIF sequence were incorporated into a mLLF molecular framework. As the sequences of mLLF and hLLF are - 80% identical (Gough et al, 1988) (Figure 4), swapping even large domains of mLLF and hLIF sequence resulted in changes of relatively few amino acid residues.

The protein concentration of each purified sample was determined by amino acid analysis, and the amount of biologically active protein in each sample was estimated by its ability to induce differentiation in mouse Ml myeloid leukaemic colonies. .An internal standard of mLIF (10⁴ U/ml) was used to normalise .all Ml cell bioassays (50 U/ml of LIF is defined as the concentration of LIF required for half ma^dm^ stimulation). The specific biological activity of each hybrid on mouse cells w-as defined by the number of units of LIF activity per milligram of protein and was found to be within the normal range for both mLIF and hLLF (1-3 x 10⁸ U/mg) (Figure 1) indicating that joining structural elements of these homologous proteins did not significantly disrupt their tertiary structure, nor affect their activity on mouse cells. ^' '

The structural integrity of all m-hLIF chimeras was also verified by evaluating their ability to bind correctly to the mLIF-R α-chain. Mouse LLF, hLLF and all hybrids had a similar ability to compete for [¹²⁵I]mLIF binding to mLBP (Figure 2A).

The first feature of hLIF that distinguishes it from mLLF, i.e. its ability to bind to the hLLF-R α-chain, was evaluated by the ability of each hybrid to compete with [¹²⁵I]hLIF for binding to COS cells expressing the hLIF-R α-chain (COS hLIF-R cells). The second feature of hLIF that distinguishes it from mLIF, i.e. its ability to bind to the mLIF-R with a higher affinity than mLIF, was evaluated by the ability of each hybrid to compete with [¹²⁵I]hLIF for binding to mLBP and by its sensitivity to biological inhibition by mLBP.

The 1000- to 5000- fold difference in the ability of mLIF and hLLF to compete with [¹²⁵I]hLIF for binding to mLBP provided a large window in which to measure the degree of hLIF-like specific binding of each chimeric protein. In each assay, the doses of hLIF, mLIF and m-hLLF hybrid required to inhibit 50% of [¹²⁵I]hLIF binding to mLBP (ID₅₀) were measured, and hLLF and mLLF were defined as having 100% and 0% hLIF-like binding activity, respectively. .Assays could then be norm ised for inter-assay variations by using a logarithmic scale to convert the ID for hLIF and mLLF to a score of 100% and 0%, respectively, then converting th ED_JO for each hybrid to a percentage score between these two extremes.

Initial experiments found that a hybrid LIF molecule (MH2), in which the amino terminal half contained mLLF amino acid residues and the carboxy-terminal hal consisted of hLLF residues, had 87 ± 4% (mean ± SEM) hLLF-like activity fo binding to mLBP, and had gained the capacity to bind to the hLIF-R α-chain. I contrast, the reverse hybrid (MH1) had only - 23 ± 3 % hLIF-like binding to mLBP (Figure 2B) and could not bind to COS hLIF-R cells. These results defined th carboxy-terrhinal half of hLIF as containing a major receptor binding determinant.

Hybrids MH3, MH4, MH5 and MH6 were constructed to resolve which structural modules were involved in the receptor binding site. When the predicted D-helix was composed of hLIF residues (MH3), the hybrid had no hLIF-like binding to mLBP. When the predicted C-helix was composed of hLIF residues (MH4), the hybrid displayed 26 ± 1% hLIF-like activity in its ability to compete for [¹²⁵I]hLIF binding to mLBP (Figure 2C). Hybrid MH5, which contained the predicted C-D loop region substituted for the equivalent hLIF residues, displayed 52 ± 1% hLLF-like activity in the same assay. The combination of hLIF residues in both the predicted C-helix and the C-D loop (MH6) resulted in a hybrid with 78 ± 2% hLIF-like activity as assessed by its ability to compete for [¹²⁵I]hLIF binding to mLBP (Figure 2C). Hybrids MH7, MH8 and MH9 were constructed in order to test whether the D-helix co-operated with either the C-helix or the C-D loop to enhance hLLF-like binding specificity. Hybrid MH7, which contained hLIF sequences in the C-D loop and in the D-helix, behaved like hybrid MH5, which contained hLIF residues in the C-D loop only. Hybrid MH9, which contained hLIF residues in the C-helix and the D-helix, behaved like hybrid MH4, in which only the C-helix comprised hLIF sequence. Hybrid MH8, which contained hLIF residues in the C-helix, C-D loop and D-helix, behaved like hybrid MH6, in which only the C-helix and C-D loop comprised hLIF sequence (Figure 1). All ch.imer.as were also tested for their ability to compete with [¹²⁵I]hLIF for binding to COS hLIF-R cells. A typical example of a competitive binding assay between the chimeras and [¹²⁵I]hLIF on COS hLIF R cells is shown in Figure 3A. When the hybrids were r.anked according to the dose required to produce 50% inhibition (TD^) of [¹²⁵I]hLIF binding to either mLBP or COS hLIF-R cells, the hierarchy was the same in each assay (Figure 1) indicating that the two features of hLIF that distinguish it from mLIF map to the same regions on the hLLF molecule.

Single amino acid substitutions were then used to identify the individual amino acids within the C-helix that were critical for hLIF-like binding to both mLBP and COS hLIF-R cells. Hybrid MH8, which had 77 ± 3% hLIF-like binding to mLBP, was used as a framework for these constructs. The residues most likely to contribute to hLIF-specific binding in the C-helix are those that are different in mLIF and hLIF, and are predicted on the basis of a helical wheel projection to be on the external (hydrophilic) surface of the molecule. There were therefore three .amino acids in the C-helix of hLLF (HI 12, SI 13, and 1121) that were candidates for investigation by site- directed mutagenesis, although residue 1121 represented a conservative substitution between mLIF (A) and hLIF (I) and so was discounted. Residues H112 and S113 were mutated individually to their equivalent mLIF residue in hybrids MH11 and MH12, respectively. If either residue was important for hLIF-like binding, its substitution, to the equivalent mLIF residue should reduce the hLIF-like binding of the hybrid, perhaps to a level similar to that of hybrid MH5, in which only the C-D loop contains hLIF residues. The hLIF-like binding of hybrids MH11 and MH12 to mLBP was scored as 67 ± 1% and 73 ± 3%, respectively, indicating that the substitution H112Q was the more significant of these mutations, but alone was insufficient to abrogate completely the contribution of the C-helix to hLLF-like binding to mLBP. Hybrid MH10, which comprised an MH8 framework with a residue in the D-helix, K168, swapped to its mLIF equivalent (T168), showed that changing the least conserved residue on the external face of the D-helix did not affect hLIF-like binding. In order to test whether hLLF-like binding activity of hybrid MH5, in which the C-D loop contaiiiS hLIF sequence, could be restored to the level of hybrid MH6, in which both the C-D loop .and the C-helix consist of hLIF residues, mLIF amino acids Q112, VI 13 in the C-helix and T168 in the D-helix were substituted with their equivalent hLIF residues. Hybrid MH14, which was based on hybrid MH5 but with additional substitutions Q112H and V113S, showed 71 ± 2% hLIF-like binding to mLBP, indicating that these two residues define the contribution of the C-helix to hLIF- specific binding (Figure 2C). Hybrid MH13, which was identical to hybrid MH14 except for an additional T168K mutation, also had 73 ± 3% hLLF-specific binding to mLBP, confirming that the mutated residues in the C-helix were sufficient to restore MH6-like mLBP binding. These chimeras were also tested for their ability to compete with [¹²⁵I]hLIF binding to both mLBP and COS hLDF-R cells and gave qualitatively the same results in both assays (Figure 1). This strategy has, therefore, identified two residues within the C-helix, H112, and S113, as critical for hLLF-like binding to both mLBP and the hLLF receptor.

A 1:4 to 1:8 dilution of normal mouse serum (equivalent to - 0.5-1 μg/ml mLBP) is required to block 50% of the Ml differentiation-inducing activity of up to 200 U/ml mLIF, whereas, due to the higher affinity of hLIF for mLBP, a 1:8192 dilution of serum ( - 0.6 ng/ml mLBP) is sufficient to block 50% of the Ml activity of up to 200 U/ml hLIF. This ability of mLBP to inhibit the differentiation-inducing activity of LIF in a species-specific manner was utilized as a second assay that distinguished hLIF from mLIF. Hybrids were also assessed for hLIF-like activity in this assay according to the dose of serum required to inhibit 50% of their Ml cell differentiation-inducing activity when they were present in the culture dish at a just maximal stimulatory concentration ( - 200 U/ml). An example of a typical serum dilution Ml assay is shown in Figure 5B. All data from this assay were consistent with the data obtained from the competitive binding assays with mLBP and [¹²⁵I]hLLF (Figure 3B), thus eliminating the possibility that the higher affinity of hLIF for mLBP was an artefact of the binding assays themselves or of the use of iodinated hLIF. A strategy similar to that used to define the important residues for hLIF binding to the hLIF-R in the predicted C-helix has been used for identifying single amino acid residues in the C-D loop region that .are critical for hLIF-like activity. Hybrids in which the C-D loop had been subdivided were constructed. Hybrids MH16 and MH17 are based on a MH4 framework, with MH16 having the first half of the C-D loop (residues 131-153) substituted for hLIF residues .and MH17 having the second h-alf of the C-D loop (residues 154-160) swapped (see Table 1).

When the predicted C-helix plus residues 131-153 of the C-D loop consisted of hLLF residues (MH16), the hybrid had a greater hLLF-like character than MH4 (50% vs. 23%), indicating that there was a small contribution to hLIF-like binding to the hLIF-R α-chain by residues 131-153. When the C-helix plus residues 154-160 of the C-D loop consisted of hLIF residues (MH17), the hybrid displayed 94% hLLF-like activity in its ability to compete for ¹²⁵I-hLIF binding to mLBP. This result defined D154-K166 as the major binding determinant within the C-D loop.

2. Hybrids MH18-MH41

The strategy for mutagenesis was to divide the LIF molecule into predicted secondary structural units of α-helix and connecting loop, test the contribution of each structural unit to hLIF-specific binding, then to ascertain the individual amino acids in that region that were responsible for that contribution.

Residues in the C-D loop were initially investigated. The assays used were found to be best able to discriminate between hybrids in the middle of their range, so hybrid MH6 which had 78 ± 2 % hLIF-like activity, was chosen as a basis for hybrids MH18-25 and MH31. .Amino acid residues in the C-D loop that were different between mLIF and hLLF were individually swapped from the hLLF residue to the equivalent mLIF residue, and the resulting chimeras tested for a decrease in hLLF-like activity relative to MH6. Hybrids MH24 .and MH25 showed a significant decrease in their % human score and their affinity for the hLIF-R α-chain relative to hLIF, and thus defined residues Val 155 and Lys 158 as being the two residues in the C-D loop that were responsible for its contribution to the overall hLIF-like activity.

The residues in the B-C loop region were then investigated, as previous comparisons of hybrids MH6 and MH8 versus MH2 and MH15 (Example 8(i)) indicated that the region between He 103 and Leu 109, which is predicted to form the B-C loop as well as the last two residues of the B-helix, contributed to hLLF-specific binding. Hybrid MH5 (52 ± 1 % hLIF-like activity) was chosen as a basis for hybrids MH29, MH30 .and MH32. Amino acid residues in the region between residues 103-109 that were different between mLIF and hLIF were individually swapped from the mLIF residue to the equivalent hLIF residue, and the resulting chimeras tested for an increase in hLIF-like activity relative to MH5. Hybrid MH29 showed an increase in hLIF-like activity in terms of binding to both the mLIF-R α-chain and d e hLIF-R α-chain, as did hybrid MH32 to a lesser e.xtent, thus defining residues Ser 107 as being the residue in the region between residues 103-109 that was responsible for its contribution to the overall hLIF-like activity.

Assuming that the contribution of each amino acid to hLIF-specific activity was additive, hybrid MH33 was constructed in order to test that the residues identified in the C-helix, the C-D loop and the B-C loop could be substituted onto a mLLF framework, and reconstitute hLIF-like activity. Hybrid MH33 was a mouse framework molecule that contained the human LIF residues His 112 and Ser 113 from the C-helix, Val 155 and Lys 158 from the C-D loop, and had 72 ± 3 % hLIF- like activity (Figure 5B) and had only an 8-fold lower affinity for the hLIF-R α-chain compared to hLIF (Figure 5C), indicating that only 5 residues out of 180 conferred approximately three-quarters of the binding energy of hLLF.

The additional 25% of the binding energy of hLLF must reside in the N-terminal half of the molecule, so hybrids MH34-36 were based on MH33, but either the N-terminal tail and the A-helix, the A-B loop or the B-helix were swapped from mLIF sequence to hLIF sequence. There was no contribution to hLLF-specific activity from the N-terminal tail, the A-helix or the B-helix, but there was a large contribution from the A-B loop (Figure 1).

EXAMPLE 9

LIF w.as aligned to G-CSF using the known hLLF and hG-CSF .amino acid sequences and the multiple sequence alignment software of Smith (1986). The alignment is shown in Figure 4. Modelling was performed using the Homology module of the Insight software system (Biosoft Inc, San Diego) and refined using the X-PLOR package (Brunger, 1992). LIF is most closely related to OSM, CNTF, G-CSF and IL-6 (Bazan, 1991), however, G-CSF was used as a basis for the model as it is the only known structure of this group. The helices of the 4 α-helical bundle were designated A, B, C and D (ordered from the N-terminus) .and the loops were labelled according to the helices they join. The helical segments in G-CSF were treated as structurally conserved regions and their co-ordinates copied to the LIF model. The region Cys 131- Ser 135 was deemed not to be part of the C-helix so as not to constrain the disulfide bridges. Both G-CSF and GH (de Vos et αl, 1992) have a 3₁₀-helix in the A-B loop, close to the A-helix, so this structural motif was also conserved and copied to the model, while the remainder of the A-B loop was copied from a loop search of the protein data bank (Bernstein, 1977). The B-C loop and C-D loop from hG-CSF were used as models for the equivalent loops in hLIF. The N-terminal loop was built by the homology module of Insight.

Refinement of the model was performed in several stages. The disulfide bond pairs (Nicola et αl, 1993) were defined, then the model was energy minimised with the Cα co-ordinates of all regions except for the N-terminal loop held fixed, to allow the segment joins and disulfide bridges to form. The model was then surrounded by 4 layers of water. The initial refinement used simulated-annealing starting at 600 °K, then reducing by 300 °K in steps of 50 °K over 0.5 or 1.0 pico-seconds, followed by 5 pico-seconds of molecular dynamics at 300 °K. To assist in maintaining a helical structure, the side chains and loops were allowed to relax with an i-i + 4 O-H distance constraint of 2.2A on the helices of the 4 α-helical bundle. This was then followed by 10 pico-seconds of unrestrained molecular dynamics. The resulting model is shown in Figure 10 as a ribbon diagram (Kraulis, 1991) with the receptor binding site amino acids shown in CPK form.

Since mLIF, hLIF and all m-hLIF chimeras had an approximately equd ability to compete with [¹²⁵I]mLIF binding to the soluble mLIF-R α-chain, and had nearly equal biological activities in a mouse cell bioassay they must contain a common binding site for the α-ch n of the mLIF-R. This binding site on the lig-and (site a) presumably comprises conserved amino acid residues in the mLIF and hLIF proteins and so is invisible in our assay system. This site, however, is not sufficient to mediate binding to the hLIF-R α-chain, or to generate enhanced binding to the mLLF-R α-chain, which are exclusive properties of hLLF. A second site on the hLLF molecule, which we have mapped in these studies (site b) is comprised of a small number of residues in the predicted C-D loop, C-helix and B-C loop of hLLF and is proposed to mediate the exclusive properties of hLLF. These residues form a cluster on one face of the predicted three-dimensional structure of hLLF (Figure 6).

The primary interaction site on the LIF ligand for its isologous receptor α-chain is not the same in the human and mouse systems. It is proposed that hLLF but not mLLF is able to recognise a site which is conserved in both the hLLF and mLLF receptor α-chains (site B), while the primary binding site for mLIF on the mLLF-R α-chain (site A), is not present on the hLIF-R α-chain. This proposed model potentially explains the low-affinity binding of mLIF and hLLF to their isologous receptor α-chains, the inability of mLIF to bind to the hLLF-R α-chain, and the apparent identity of the sites on the hLIF molecule that mediate both the high- .affinity binding to the mLIF-R α-chain and binding to the hLLF-R α-chain. This appears to be the simplest model consistent with these data. Those skilled in the art will appreciate that the invention described herein is susceptible to variations and modifications other than those specifically described. It is to be understood that the invention includes all such variations and modifications. The invention .also includes all of the steps, features, compositions and compounds referred to or indicated in this specification, individuaHy or collectively, and any and all combinations of any two or more of said steps or features.

Table Legends

Table 1 Amino acid sequence specifications for m-hLLF chimeric proteins. Sequences .ar designated according to the following example. Ml-130: H131-160: M161-180 (MH5 denotes that residues between 1 and 130 as well as between 161 and 180 are derive from mLIF sequence and residues 131-160 are derived from hLIF sequence. Singl amino substitutions are denoted as H138R, indicating that the histidine at positio 138 is changed to arginine.

TABLE 1

HYBRID AMINO ACID SEQUENCE SPECIFICATION

M M1-180

H H1-180

MH1 H1-102:M103-180

MH26 H1-109:M110-180

MH2 M1-102:H103-180

MH3 M1-160:H161-180

MH4 M1-109:H110-130:M131-180

MH5 M1-130:H131-160:M161-180

MH6 M1-109:H110-160:M161-180

MH15 M1-102:H103-160:M161-180

MH7 M1-130:H131-180

MH8 M1-109:H110-180

MH9 M1-109:H110-130:M131-160:H161-180

MH10 MH8+K168T

MH11 MH8+H112Q

MH12 MH8+S113V

MH13 MH5+Q112H:V113S:T168K

MH27 MH5+Q112H

MH28 MH5+V113S

MH14 MH5+Q112H:V113S

MH16 M1-109:H110-153:M154-180

MH17 M1-109:H110-130:M131-153:H154-160:M161-180

MH18 MH6+H138R

MH31 MH6+T145P

MH19 MH6+Y146P

MH20 MH6+G147V

MH21 MH6+T150H

MH22 MH6+G152D

MH23 MH6+D154E

MH24 MH6+V155A

MH25 MH6+K158R

MH29 MH5+T107S

MH30 MH5+V109L

MH32 MH5+V103I

MH33 M+T107S:Q112H:V113S:A155V:R158K

MH34 MH33+H1-47

MH35 MH33+H48-81

MH36 MH33+H82-104

MH37 MH33+V103I

MH38 MH33+V56L

MH39 MH33+V109L

MH40 MH33+V56L:V103I:V109L

MH41 MH33+V56L:V103I REFERENCES

Abdel-Meguid, SS et al., Proc Natl Acad Sci USA 84. 6434-7 (1987).

Bazan, JF Proc Natl Acad Sci USA 87: 693-8 (1990).

Bazan, JF Immunol Today 11: 350-4 (1990).

Bazan JF Neuron 7(2): 197-208 (1991).

Bernstein, /. Mol. Biol. 722: 535-542 (1977).

Blundell T et al Ew. J. Biochem. 172, 513-520 (1988).

Brandhuber, BJ, T. Boone, W. C. Kenney, D. B. McKay, J Biol Chem 262: 12306-

(1987).

Brunger, AT "X-PLOR, a system for crystallography and NMR." Yale Univ., New

Haven, CT (1992).

Butcher LA and JK Tomkins Anal. Bioch. 148: 384-388 (1985).

Cosman, D et al., Trends Biochem Sci 15: 265-70 (1990).

Diederichs, K, T. Boone, P. A. Karplus, Science 254: 1779-82 (1991).

Gearing DP, Nicola NA, Metcalf D, Foote S, Willson TA, Gough NM and Williams

L BioTechnology 7: 1157-61 (1989).

Gearing DP, et al., EMBO J 10: 2839-48 (1991).

Gearing DP, et al., Science 255: 1434-7 (1992).

Gough NM, et al., Proc. Natl. Acad. Sci. USA 85: 2623-2727, (1988).

Hill CP, T. D. Osslund, D. Eisenberg, Proc Natl Acad Sci USA 90. 5167-71 (1993).

Hilton DJ, Nicola NA and Metcalf D Anal. Biochem. 173(2): 359-67 (1988a).

HUton DJ, Nicola NA and Metcalf D Proc. Natl. Acad. Sci. USA 85(16): 5971-5

(1988b).

Ho SN, Hunt HD, Horton RM, Pullen JK and Pease LR, Gene 77(1): 51-9 (1989).

Kraulis PJ, J. Appl. Cryst. 24: 946-950 (1991).

Layton MJ, Cross BA, Metcalf D, Ward LD, Simpson RD and Nicola NA Proc.

Natl. Acad. Sci. USA 89: 8616-20 (1992) .

Metcalf D Int. J. Cell Cloning 9(2): 92-108 (1991).

Metcalf D et al., Leukaemia 2: 216-221, (1988).

Milburn MV, et al., Nature 3 63: 172-6 (1993). Nicola NA, B. Cross, R. J. Simpson, Biochem. Biophys. Res. Comm. 190. 20-6 (1993).

Owczarek CM, et al., EMBO J 123487-95 (1993).

Parry AD, et al., J. Mol. Recog. 4: 63-75, (1991).

Powers R, et al, Science 256. 1673-7 (1992).

Sanger FA, Nicklen S and Coulson AR Proc. Natl. Acad. Sci. USA 74: 5463-7

(1977).

Seed B Nature 329: 840-842 (1987).

Smith DK, "Concepts and use of software for pair-wise and multiple genetic sequence alignment" Technical Report (School of Information Sciences and

Engineering, 1986).

Smith DB and Johnson KS Gene 67(1): 31-40 (1988). de Vos AM, M. Ultsch, A. A. Kossiakoff, Science 255: 306-12 (1992).

Willson TA, Metcalf D and Gough NM Eur. J. Biochem. 204(1): 21-30 (1992):

SEQUENCE LISTING

(1) GENERAL INFORMATION:

(i) APPLICANT:

(COUNTRIES OTHER THAN US): AMRAD CORPORATION LIMITED (IN THE US): LAYTON, Meredith J, OWCZAREK, Catherine M,

NICOLA, Nicos A, GOUGH, Nicholas M and

METCALF, Donald

(ii) TITLE OF INVENTION: RECEPTOR-BINDING DETERMINANT

FROM LEUKAEMIA INHIBITORY FACTOR

(iii) NUMBER OF SEQUENCES: 4

(iv) CORRESPONDENCE ADDRESS:

(A) ADDRESSEE: DANTES COLLISOΝ CANE

(B) STREET: 1 LITTLE COLLINS STREET

(C) CITY: MELBOURNE

(D) STATE: VICTORIA

(E) COUNTRY: AUSTRALIA

(F) ZIP: 3000

(v) COMPUTER READABLE FORM:

(A) MEDIUM TYPE: Floppy disk

(B) COMPUTER: IBM PC compatible

(C) OPERATING SYSTEM: PC-DOS /MS-DOS

(D) SOFTWARE: Patentln Release #1.0, Version #1.25

(vi) CURRENT APPLICATION DATA:

(A) APPLICATION NUMBER: PCT - INTERNATIONAL

(B) FILING DATE: 03-FEB-1994

(vii) PRIOR APPLICATION DATA:

(A) APPLICATION NUMBER: PL7102

(B) FILING DATE: 03-FEB-1993

(viii) ATTORNEY/AGENT INFORMATION: (A) NAME: HUGHES, E JOHN L

(C) REFERENCE/DOCKET NUMBER: EJH/EK

(ix) TELECOMMUNICATION INFORMATION:

(A) TELEPHONE: (613) 254 2777

(B) TELEFAX: (613) 254 2770 (2) INFORMATION FOR SEQ ID NO:l:

(i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 181 amino acids

(B) TYPE: amino acid

(C) STRANDEDNESS: single

(D) TOPOLOGY: linear

(ii) MOLECULE TYPE: protein - murine LIF; the mature protein begins at Ser 1

(xi) SEQUENCE DESCRIPTION: SEQ ID NO:l:

Gly Ser Pro Leu Pro He Thr Pro Val Asn Ala Thr Cys Ala He Arg -1 1 5 10 15

His Pro Cys His Gly Asn Leu Met Asn Gin He Lys Asn Gin Leu Ala 20 25 30

Gin Leu Asn Gly Ser Ala Asn Ala Leu Phe He Ser Tyr Tyr Thr Ala 35 40 45

Gin Gly Glu Pro Phe Pro Asn Asn Val Glu Lys Leu Cys Ala Pro Asn ' 50 55 60

Met Thr Asp Phe Pro Ser Phe His Gly Asn Gly Thr Glu Lys Thr Lys 65 70 75

Leu Val Glu Leu Tyr Arg Met Val Ala Tyr Leu Ser Ala Ser Leu Thr 80 85 90 95

Asn He Thr Arg Asp Gin Lys Val Leu Asn Pro Thr Ala Val Ser Leu 100 105 110

Gin Val Lys Leu Asn Ala Thr He Asp Val Met Arg Gly Leu Leu Ser 115 120 125

Asn Val Leu Cys Arg Leu Cys Asn Lys Tyr Arg Val Gly His Val Asp 130 135 140

Val Pro Pro Val Pro Asp His Ser Asp Lys Glu Ala Phe Gin Arg Lys 145 150 155

Lys Leu Gly Cys Gin Leu Leu Gly Thr Tyr Lys Gin Val He Ser Val 160 165 170 175

Val Val Gin Ala Phe 180 (2) INFORMATION FOR SEQ ID NO:2:

(i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 181 amino acids

(B) TYPE: amino acid

(C) STRANDEDNESS: single

(D) TOPOLOGY: linear

(ii) MOLECULE TYPE: protein - human LIF; the mature protein begins at Ser 1

(xi) SEQUENCE DESCRIPTION: SEQ ID NO:2:

Gly Ser Pro Leu Pro He Thr Pro Val Asn Ala Thr Cys Ala He Arg -1 1 5 10 15

His Pro Cys His Asn Asn Leu Met Asn Gin He Arg Ser Gin Leu Ala 20 25 30

Gin Leu Asn Gly Ser Ala Asn Ala Leu Phe He Leu Tyr Tyr Thr Ala 35 40 45

Gin Gly Glu Pro Phe Pro Asn Asn Leu Asp Lys Leu Cys Gly Pro Asn ' > 50 55 60

Val Thr Asp Phe Pro Pro Phe His Ala Asn Gly Thr Glu Lys Ala Lys 65 70 75

Leu Val Glu Leu Tyr Arg He Val Val Tyr Leu Gly Thr Ser Leu Gly 80 85 90 95

Asn He Thr Arg Asp Gin Lys He Leu Asn Pro Ser Ala Leu Ser Leu 100 105 110

His Ser Lys Leu Asn Ala Thr Ala Asp He Leu Arg Gly Leu Leu Ser 115 120 125

Asn Val Leu Cys Arg Leu Cys Ser Lys Tyr His Val Gly His Val Asp 130 135 140

Val Thr Tyr Gly Pro Asp Thr Ser Gly Lys Asp Val Phe Gin Lys Lys 145 150 155

Lys L-eu Gly Cys Gin Leu Leu Gly Lys Tyr Lys Gin He He Ala Val 160 165 170 175

Leu Ala Gin Ala Phe 180 (2) INFORMATION FOR SEQ ID NO:3:

(i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 181 amino acids

(B) TYPE: amino acid

(C) STRANDEDNESS: single

(D) TOPOLOGY: linear

(ii) MOLECULE TYPE: protein - porcine LIF; the mature protein begins at Ser 1

(xi) SEQUENCE DESCRIPTION: SEQ ID NO:3:

Gly Ser Pro Leu Ser He Thr Pro Val Asn Ala Thr Cys Ala Thr Arg -1 1 5 10 15

His Pro Cys His Ser Asn Leu Met Asn Gin He Lys Asn Gin Leu Ala 20 25 30

His Leu Asn Ser Ser Ala Asn Ala Leu Phe He Leu Tyr Tyr Thr Ala 35 40 45

Gin Gly Glu Pro Phe Pro Asn Asn Leu Asp Lys Leu Cys Gly Pro Asn 50 55 60

Asx Thr Asn Phe Pro Pro Phe His Ala Asn Gly Thr Glu Lys Ala Arg 65 70 75

Leu Val Glu Leu Tyr Arg He He Ala Tyr Leu Gly Ala Ser Leu Gly 80 85 90 95

Asn He Thr Arg Asp Gin Arg Ser Leu Asn Pro Gly Ala Val Asn Leu 100 105 110

His Ser Lys Leu Asn Ala Thr Ala Asp Ser Met Arg Gly Leu Leu Ser 115 120 125

Asn Val Leu Cys Arg Leu Cys Asn Lys Tyr His Val Ala His Val Asp 130 135 140

Val Ala Tyr Gly Pro Asp Thr Ser Gly Lys Asp Val Phe Gin Lys Lys 145 150 155

Lys Leu Gly Cys Gin Leu Leu Gly Lys Tyr Lys Gin Val He Ser Val 160 165 170 175

Leu Ala Arg Ala Phe 180 (2) INFORMATION FOR SEQ ID NO:4:

(i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 175 amino acids

(B) TYPE: amino acid

(C) STRANDEDNESS: single

(D) TOPOLOGY: linear

(ii) MOLECULE TYPE: protein

(xi) SEQUENCE DESCRIPTION: SEQ ID NO:4:

Ala Thr Pro Leu Gly Pro Ala Ser Ser Leu Pro Gin Ser Phe Leu Leu 1 5 10 15

Lys Cys Leu Glu Gin Val Arg Lys He Gin Gly Asp Gly Ala Ala Leu 20 25 30

Gin Glu Lys Leu Cys Ala Thr Tyr Lys Leu Cys His Pro Glu Glu Leu 35 40 45

Val Leu Leu Gly His Ser Leu Gly He Pro Trp Ala Pro Leu Ser Ser 50 55 60

Cys Pro Ser Gin Ala Leu Gin Leu Ala Gly Cys Leu Ser Gin Leu His 65 70 75 80

Ser Gly Leu Phe Leu Tyr Gin Gly Leu Leu Gin Ala Leu Glu Gly He 85 90 95

Ser Pro Glu Leu Gly Pro Thr Leu Asp Thr Leu Gin Leu Asp Val Ala 100 105 110

Asp Phe Ala Thr Thr He Trp Gin Gin Met Glu Glu Leu Gly Met Ala 115 120 125

Pro Ala Leu Gin Pro Thr Gin Gly Ala Met Pro Ala Phe Ala Ser Ala 130 135 140

Phe Gin Arg Arg Ala Gly Gly Val Leu Val Ala Ser His Leu Gin Ser 145 150 155 160

Phe Leu Glu Val Ser Tyr Arg Val Leu Arg His Leu Ala Gin Pro 165 170 175

Claims

CLAIMS:

1. A non-naturally occurring molecule comprising a tertiary structure which presents a functional binding face for the α-chain of human leukaemia inhibitory factor (hLIF) binding-receptor.

2. A non-naturally occurring molecule according to claim 1 wherein the functional binding face comprises the chemically interactive groups of the amino acid residues which constitute the binding determinant for the α-chain of hLIF binding receptor.

3. A non-naturally occurring molecule according to claim 2 wherein the chemically interactive groups are selected from a methyl group, a hydroxyl group, an imino-nitrogen group and a positively charged nitrogen group or chemical equivalents, homologues or analogues thereof.

4. A non-naturally occurring molecule according to claim 1 wherein the functional binding face comprises the amino acid residues or chemical equivalents, homologues or analogues thereof which constitute a binding determinant for the α- chain of hLIF binding receptor on hLIF.

5. A non-naturally occurring molecule according to claim 4 wherein the amino acid residues constituting the hLIF receptor binding determinant are selected from amino acid residues Gin 48 to Lys 160 of hLIF.

6. A non-naturally occurring molecule according to claim 5 wherein the amino acid residues constituting the hLIF receptor binding determinant are selected from at least one of each of the following predicted structural regions of hLIF:

(i) a region between A and B helices;

(ii) a region between the B and C helices;

(iii) the C helix; and

(iv) a region between the predicted C and D helices. 7. A non-naturally occurring molecule according to claim 6 wherein the amino acid residues constituting the hLIF receptor binding determin.ant are selected from a lie, Ser, His, Nal and Lys.

8. A non-naturally occurring molecule according to claim 7 wherein the amino acid residues constituting the hLIF receptor binding determinant are Ser 107, His 112, Ser 113, Nal 155 and Lys 158 of hLIF or chemical equivalents, homologues or ■analogues thereof.

9. A non-naturally occurring molecule according to claim 1 wherein the molecule forms part of a solid support and /or is a non-proteinaceous molecule.

10. A non-naturally occurring molecule according to claim 1 wherein the molecule is a protein, polypeptide or peptide.

11. A non-naturally occurring chemical entity comprising a tertiary structure which presents chemically interactive groups selected from a methyl group, a hydroxyl group, an imino-nitrogen group and a positively charged nitrogen group as a functional binding face for the α-chain of hLIF binding receptor.

12. A non-naturally occurring chemical entity according to claim 1 or 11 wherein the non-naturally occurring molecule exhibits at least about 50% hLIF-like activity as determined by the equation: log x_m - log x_c

≥ 0.50 log x_m - log x.

wherein ^ is a dose of unlabelled mLIF; x_h is a dose of unlabelled hLIF; and x,. is a dose of unlabelled non-naturally occurring chemical entity to give 50% inhibition^" of ¹²⁵I-hLIF binding to mLIF-binding protein. 13. A non-naturally occurring chemical entity according to claim 12 wherein the chemical entity is a protein, polypeptide or peptide.

14. A non-naturally occurring chemical entity according to claim 13 wherein the chemically interactive groups are presented on amino acid residues or chemical equivalents, homologues or analogues thereof.

15. A non-naturally occurring chemical entity according to claim 14 wherein the amino acids are selected from Ser, His, Nal and Lys or their chemical equivalents, homologues or analogues.

16. A molecule which is non-naturally occurring and which comprises a carrier portion and an active portion wherein said active portion comprises amino acid residues or chemical equivalents, homologues or analogues thereof which constitute a binding determinant for the α chain of the human leukaemia inhibitory factor (hLIF) binding receptor.

17. A molecule according to claim 1 wherein the amino acid residues constituting the hLIF receptor binding determinant are selected from amino acid residues from Asn 21 to Gin 178 or chemical equivalents, homologues or analogues thereof of hLIF.

18. A molecule according to claim 17 wherein the amino acid residues constituting- the hLIF receptor binding determinant are selected from at least one of each of the following predicted structural regions of hLIF:

(i) a region between the A and B helices;

(ii) a region between the B and C helices;

(iii) the C helix; and

(iv) a region between the C and D helices.

19. A molecule according to claim 18 wherein the amino acid residues constituting the hLIF receptor binding determinant are Ser 107, His 112, Ser 113, Nal 155 and Lys 158 of hLIF.

20. A molecule according to claim 16 or 19 wherein the carrier portion of the molecule is a protein, polypeptide or peptide.

21. A molecule according to claim 20 wherein the carrier portion of the molecule is a mammalian cytokine.

22. A molecule according to claim 21 wherein the cytokine is a non-human LIF.

23. A molecule according to claim 22 wherein the non-human LIF is from a livestock animal, laboratory test animal, companion animal or captive wild animal'.

24. A molecule according to claim 23 wherein the non-human LIF is murine LIF.

25. A molecule according to claim 21 wherein the mammalian cytokine is a colony stimulating factor, an interleukin, oncostatin M, ciliary neurotrophic factor or a haemopoietic growth factor.

26. A molecule according to claim 25 wherein the mammalian 4cytokine is G-CSF.

27. A molecule according to claim 24 wherein said molecule e.xhibits at least about 50% hLIF-like activity as determined by the equation:

log x_m - log x_c

≥ 0.50 log x_m - log x_h

wherein ^ is a dose of unlabelled mLIF; ^ is a dose of unlabelled human LIF," and x,. is a dose of unlabelled molecule to give 50% inhibition of ¹²⁵I-hLIF binding to murine LIF-binding protein.

28. A hybrid molecule comprising: (i) a polypeptide backbone having a tertiary structure;

(ii) amino acid residues constituting a binding determinant for the α chain of hLIF binding receptor inserted into said polypeptide backbone such that the .arrangement of said amino acid residues in the tertiary structure of the polypeptide provides a hLIF receptor-binding face; and wherein said hybrid molecule exhibits at least about 50% hLIF-like activity as defined by the equation: log x_m - log x_c

≥ 0.50 log x_m - log x.

wherein ^ is a dose of unlabelled mLIF; x_h is a dose of unlabelled hLIF; and X_j. is a dose of hybrid molecule required to give 50% inhibition of ¹²⁵I-hLIF binding ϋo murine LIF binding protein.

29. A hybrid molecule according to claim 28 wherein the polypeptide backbone is a mammalian cytokine.

30. A hybrid molecule according to claim 29 wherein the cytokine is a non-human LIF.

31. A hybrid molecule according to claim 30 wherein the non-human LIF is from a livestock animal, laboratory test animal, companion animal or captive wild animal.

32. A hybrid molecule according to claim 31 wherein the non-human LIF is murine LIF.

33. A hybrid molecule according to claim 29 wherein the mammalian cytokine is a colony stimulating factor, an interleukin, oncostatin M, ciliary neurotrophic factor or a haemopoietic growth factor. 34. A hybrid molecule according to claim 33 wherein the mammalian cytokine is G-CSF.

35. A hybrid molecule according to claim 28 wherein the amino acid residues constituting the hLIF receptor binding determined are selected from at least one of the following predicted structural regions:

(i) a region between the A and B helices;

(ii) a region between the B and C helices;

(iii) the C helix; and

(iv) a region between the C and D helices.

36. A hybrid molecule according to claim 35 wherein the amino acid residues constituting the hLIF receptor binding determinant are Ser 107, His 112, Ser 113, Nal 155 and Lys 158 of hLIF.

37. A nucleic acid comprising a sequence of nucleotides which encodes or is complementary to a sequence which encodes a molecule according to claim 20 or claim 28.

38. A nucleic acid according to claim 37 wherein said nucleic acid molecule is cDΝA.

39. A nucleic acid according to claim 38 present in an expression vector capable of expression in a prokaryotic and /or eukaryotic cell.

40. A prokaryotic or euk.aryotic cell containing a nucleic acid according to claim 37 or 38 or 39. 41. A hybrid molecule according to claim 38 selected from the list consisting of MH5, MH6, MH16, MH17, MH18, MH31, MH19, MH20, MH21, MH22, MH23, MH24, MH25, MH29, MH3G, MH32, MH33, MH34, MH35, MH36, MH37, MH38, MH39, MH40 and MH41.

42. A hybrid molecule according to claim 38 selected from the list consisting of MH33, MH35, MH36, MH37, MH38, MH40 and MH41.

43. A hybrid molecule according to claim 38 wherein said molecule is MH33.