HU222983B1 - Monomeric and dimeric antibody-fragment fusion protein - Google Patents

Abstract

FIELD OF THE INVENTION The present invention relates to novel antigen binding molecules comprising an Fv fragment of a single antibody but not containing constant antibody domains. In addition, they are capable of dimerizing with other antibody molecules or non-hematopoietic fragment molecules, and thus producing fusion proteins of either a multifunctional antibody fragment or so-called miniature antibodies. The fusion proteins of the invention can be applied in a number of areas of diagnostics and therapy. ŕ

Description

The fusion proteins of the present invention can be used in many areas of diagnostics and therapy.

The scope of the description is 18 pages (including 4 pages)

HU 222 983 B1

HU 222 983 Bl

The present invention relates to novel antigen binding molecules comprising an Fv fragment of an antibody but not containing the constant antibody domains. In addition, they are able to dimerize with other antibody molecules or non-antibody fragment fragments to form bi- or multifunctional antibody fragment fusion proteins as well as so-called minibodies.

The fusion proteins of the present invention are useful in a variety of diagnostic and therapeutic applications.

For some years now, there has been a keen interest in biotechnology to modify naturally occurring antibodies in order to obtain more specific and even more specific types of antibodies. Therefore, numerous attempts have been made to produce modified antibody fragments.

All antibodies of any class have at least two binding sites. This allows them to bind to a surface with higher functional affinity (also called avidity) than monovalent fragments, i.e., Fab fragments. Several methods have been described in the last few years (Skerra and Pluckthun, Science 240: 1038-1040 (1988); Better et al., Science 240: 1041-1043 (1988)] which can be used to produce functional antibody fragments in Escherichia coli. These include the Fv fragment (heterodimer containing VH and VL) and the Fab fragment (which contains the complete light chain together with the VL and CL domains and the first two domains of the VH and CH1 heavy chains).

However, the Fv fragment tends to dissociate to VH and VL, so it is preferable that the two domains are covalently linked. One way of linking them is to design a peptide link between the two, either in the VH-linkerVL orientation or in the VL-linker-VH orientation [Bírd et al., Science, 242, 423 (1988); Houston et al., Proceedings of the National Academy of Sciences, USA, 85, 5879 (1988)]. The resulting ffagments are called single-stranded Fv ffagments.

However, all of these ffagments are monovalent. The present invention provides a method for the formation of small dimerization domains based on peptides forming amphipathic helices. These peptides will be referred to as "intercalators," but this term is merely an expression of the ability of the targeted association, and is not limited to a particular structure of the dimerization interface.

While the method described is applicable in principle to both Fab and Fv or scFv fragments, the latter is the one that is most preferred. In this case, very small bivalent ffagments can be made, and even dissociation to VL and VH, as well as malfunctioning of ffag chains (e.g., VL-VL) can be prevented.

In many applications, small antibody fragments are particularly preferred. In diagnostic applications (such as ELISA, RIA, etc.), the smaller molecular surface area reduces problems caused by non-specific interactions, which are known to occur frequently in the constant domain domain. The same is true when antibody fragments are used as a ligand in affinity chromatography. In tumor diagnostics or therapy, it is important that a significant portion of the injected antibody penetrate the tissue and localize to the tumor, which depends on the molecular dimensions [Colcher et al., J. Natl. Cancer Inst., 1990, 82, 1191-1197. Expression expression of recombinant proteins and the efficiency of their secretion are also dependent on the size of the chain (Skerra & Pluckthun, Protein Eng., 4, 971 (1991)), and therefore smaller proteins are preferred. Therefore, small molecules are more advantageous for a number of reasons.

Previously, reducing the molecular size of the antibody involved producing proteolytic fragments. The smallest divalent fragments, the (Fab) '2 fragments, are still about twice as large as the fragments of the present invention. Therefore, these new fragments combine three properties:

a) small size,

(b) Bivalent or Bivalent Function; and

c) Ability to express functional expression in Escherichia coli.

There is a great deal of interest in bifunctional antibodies in many areas. Bifunctional antibodies can be defined as antibodies that have two different specificities either against two different antigens or against two different epitopes on the same antigen.

At present, many methods are known for producing bifunctional antibodies. However, none of the existing methods allows the production of only bifunctional antibodies in vivo, but rather a mixture of molecular species, which requires complicated and expensive separation procedures.

There are four basic methods. The former utilizes chemical crosslinks, in which heterobifunctional crosslinking agents are advantageously used. Using this method complete antibodies (Staerz et al., Naturre, 314, 628 (1985); Perez et al., Naturre, 316, 354-356 (1985)], Fab fragments (Carter et al., Bio / Technology, 10, 163 (1992)) and scFv fragments (Cumber et al., J, Immunol). 149,120 (1992)] after chemical purification.

The second, previously used method is based on the fusion of two hybridomas to give so-called heterohybridomas or "quadromas". In this process, any light chain can pair with any heavy chain, and the two heavy chains can form homodimers or heterodimers, resulting in very complex product mixtures (Milstein & Cuello, Naturre, 305, 537 (1983)).

The third method is related to the second and consists in transfecting two expression plasmids into a hybridoma cell encoding the heavy and light chain of the second antibody (Lenz & Weidle, Gene, 1990, 87, 213) or a retroviral vector [De Monté et al. ., Acad. Sci., 87, 2941-2945 (1990)]. After delivery, however, the product is the same as in the second process.

Finally, the antibodies were reduced, mixed and re-oxidized (Staerz & Bevan, Immunology Today 7 (1986)). thus again a very complicated product mix2

EN 222 983 ΒΙ which requires sophisticated separation and quality control procedures.

Thus, there is still a need for a technique that permits the direct isolation of heterodimeric antibodies only, without the need for complicated separations in the formation of chemical crosslinks. The present invention solves this problem by (i) covalently linking the corresponding VH and VL domains in a scFv fragment, and (ii) using dimerization domains that allow only heterodimers to be formed, such as certain leucine zippers and their derivatives.

Another important consideration of the present invention is to keep the molecular weight of the bispecific antibody as low as possible for the reasons detailed above. This was achieved using scFv fragments.

Many functions of bispecific antibodies have been described, and most of them can benefit from this new technology. Bispecific antibodies are of great interest in tumor therapy. One arm of the antibody may be linked to a tumor marker, the other to a T-cell epitope, a toxin, or a radioactive atom-binding peptide or protein, to bring about a destructive function close to the tumor cell. In diagnosis, one arm may be related to the component of interest to us, and the other to a component that is readily quantifiable, such as an enzyme. Finally, in cellular applications, it may be advantageous to achieve greater selectivity in binding if two different epitopes are recognized, or the same protein complex, or if two different proteins are recognized on the same cell surface.

Thus, it is an object of the present invention to provide novel unique and stable antibody fragment fusion proteins having bi- or multifunctional binding sites.

It has been found that antibody fusion proteins containing Fv fragments can be produced by genetic engineering techniques and have specific and improved properties.

The present invention thus relates to a monomeric antibody fragment fusion protein comprising essentially an Fv fragment of an antibody and a peptide capable of dimerizing with another peptide by non-covalent interactions.

The term "non-covalent interaction" refers to any interaction that is normally stable under normal conditions and is non-covalent bond, such as Van Der Waal's forces, (steric) interdigitation or amphiphilic peptides, particularly peptide helices, or peptides having opposite charge residues. . Suitable effective peptides are referred to above and below as interactive or intercalating peptides.

The amphiphilic peptides can contain up to 50 amino acids. Preferably they contain 10-30 amino acids. In a preferred embodiment of the invention, the interactive peptide is a peptide-helix bundle (comprising a helix, a hairpin, and another helix, see above). In another embodiment, the interactive peptide is a leucine zipper comprising a peptide comprising a plurality of repeating amino acids in which every seventh amino acid residue is a leucine residue. In other cases of the invention, the peptide contains positively or negatively charged groups, such as lysine (positively charged) or glutamic acid (negatively charged) such that this peptide is capable of binding to another peptide (belonging to another monomeric moiety), which carries opposite charges.

The Fv fragment and the intercalation peptide are linked either directly or via a linker peptide, preferably via a linker peptide. In a preferred embodiment, the linker peptide is the hinge region sequence of an antibody.

By definition, the Fv fragment contains the VL and VH regions of an antibody. Preferably, the Fv fragment of the invention is a single chain fragment. Single chain fragments may be prepared by standard techniques using standard linker molecules.

The invention further relates to a dimeric fusion protein consisting essentially of two monomeric fusion proteins, wherein the relationship between the two monomeric units is based on non-covalent interaction between the same or different peptides, wherein at least one monomeric unit is an antibody Fv fragment fusion. protein as defined above.

When the dimer contains two Fv fragments, the Fv fragments may be the same (same antigen binding sites) or different (different antigen binding sites). In these cases, mono- and bispecific (Fv) -antibodies are available. Bispecific minibodies are preferred according to the invention.

The interactive peptides may be the same or different; it is advantageous if they are the same. The intercalating peptides may be linked in one way or in the opposite direction.

Accordingly, the present invention relates, in particular, to a dimeric fusion protein comprising two Fv fragments of different specificity (antigen binding sites) and the same intercalation helix peptides, and the antibody fragments and peptides joined by a hinge region sequence.

The present invention further provides a dimer comprising an Fv fragment and another monomeric unit in which the Fv fragment is replaced by a non-antibody peptide. The non-antibody peptide may be a toxin such as ricin, a chelator or metal binding peptide or an enzyme (e.g. a marker enzyme) or a peptide carrying a detectable label (e.g. a radioactive isotope).

The non-antibody peptide may also be a suitable binding site for said groups, including binding sites for T cells or T cell fragments. In addition, the invention relates to monomers and dimers as defined above, while the interactive peptide (s) is fused at the C-terminus to the target proteins / peptides as described above, including the appropriate binding sites. Thus, the resulting fusion proteins and minibodies are multifunctional.

The invention further relates to a process for the preparation of monomeric antibody fusion proteins according to the invention.

Characterized in that the genes encoding the Fv fragment, the interactive peptide and, if necessary, the linker peptide are cloned into an expression plasmid, transformed into a host cell with said expression plasmid, and the transformed host cell grown in a suitable medium, and the monomeric fusion protein is expressed in the cell or secreted into the medium.

Finally, the present invention provides a method for producing the dimeric fusion proteins as defined above, characterized by cloning the genes encoding the complete monomeric fusion proteins, or a portion thereof, into at least one expression plasmid, transforming a host cell with said expression plasmid (s), and and / or the complete dimeric fusion protein is expressed in the cell or in the medium or the monomeric fusion proteins are expressed separately and the non-covalent linkage between the two monomeric units is established in the medium or in vitro and only when the fusion protein is some were cloned, and protein engineering steps were still performed using standard techniques.

The dimeric Fv fragment containing the dimeric fusion proteins comprising the fusion proteins of the present invention has a high avidity to the corresponding antigens and satisfactory stability. These novel bivalent or bifunctional molecules can be produced as naturally-conforming molecules in Escherichia coli. These minibodies are compatible with secretory functional expression.

The oligomerization domains were chosen to be sufficiently low in molecular weight and compatible with membrane transport of fusion proteins. They are based on two different types of amphiphilic helix.

The amphiphilic helices are known to be predominantly but not exclusively associated with two different molecular structures: four helix bundles and looped loops. The design and creation of helix bundles has been previously studied (Eisenberg et al., Proteins, 1986, 1,16-22; Ho and de Grado, 1987, Journal of the American Chemical Society, 109, 6751-6758; Regan and de Grado, Science, 241: 976-978 (1988); Hill et al., Science, 294,543-546 (1990)]. This molecular association is also known for these natural proteins [Richardson, Adv. Prot. Chem., 34, 167 (1981)].

The four helix bundles are formed from either four separate molecules (each contributing to one helix), or two molecules containing two helices (helix-bend-helix linked), or a molecule that is a helix-bend-helix-bend -helix-bend-helix motif. Only the first two are suitable for dimerization or multimerization.

Three variants of the latter were examined. The first uses a helix of the sequence given by Eisenberg et al. (Proteins, 1, 16-22 (1986)). In the second, this sequence was extended with a small hydrophilic peptide terminating in cysteine. When the helixes associate with each other, the hydrophilic peptides remain in close contact to collide and form a disulfide bridge under oxidizing conditions such as in the periplasmic space of Escherichia coli. In the third variation, two helices are used sequentially, separated by two curves encoding short peptides.

The second design uses peptides that are capable of forming a so-called looped structure. Such peptides are found in transcription factors such as GCN4 in yeast and are called leucine locks (Landschultz et al., Science 240: 1759-1764 (1988)). Its crystal structure has recently been determined (O'Shea et al., Science, 254, 539-544 (1991)) and has been shown to have a parallel arrangement with helices.

Covalent attachment of helixes is possible with small peptide protrusions which again contain a cysteine. Because the helices are now parallel, the peptide projection may be much shorter since the distance is much smaller.

However, different dimerizing structures (intercalating helices) do not fuse directly to the antibody domain. Preferably, a flexible peptide is inserted between the end of the scFv fragment and the beginning of the helix. As an example, the upper hinge region of mouse IgG3 is used. However, many different wrists can be used. It is not necessary for dimerization alone, but it provides space for the two scFv domains, similar to the antigen binding sites of the whole antibody. In this way, the two binding sites can be spatially spaced and reach adjacent antigens on a solid surface.

Naturally occurring hinge regions of antibodies are preferred embodiments of hinge regions for divalent minibodies. In the case of bifunctional minibodies, the hinge regions may be shorter, since molecules on different surfaces often need to be cross-linked as closely as possible and dimer flexibility is not required. The choice of wrist is guided by the length of the desired residue sequence (Argos, 1990, Journal of Molecular Biology, 211,943-958), compatibility with the formation and stability of the amphiphilic helix structure (Richardson & Richardson, Science 240: 1648-1652 (1988)). ], secretion and resistance to proteases.

The present invention relates to peptides as dimerization devices which should be as small as possible. In one preferred embodiment, peptides are used which are capable of generating amphipathic peptides. Such helices cover the hydrophobic surface by dimerization or even multimerization. Helixes of this type are characterized by having hydrophobic moieties on one side of the helix and containing a sufficient number of helix-forming groups. The rules for such peptides are described by Eisenberg et al., Proteins, 1,16-22 (1986), O'Shea et al., O'Shea et al., Science 254: 539-544 (1991); Cell, 68, 699-708 (1992)].

Peptides of this type are found in the form of so-called leucine zippers characterized by

Leucine is periodically present in it (every seventh group), and other hydrophobic groups (such as valine) also occur in a weekly period. Now that these principles are known [O'Shea et al., Science, 254, 539-544 (1991); Cell 68: 699-708 (1992)], the sequence may be varied such that groups which do not promote association of homodimers but promote association of heterodimers may be incorporated. Such a sequence change may include, for example, the incorporation of charge bridges such that charges are repelled in homodimers and attracted in heterodimers (see below).

The present invention may also be extended to bifunctional minibodies. In this case, the dimerization tools (intercalating peptides) which allow only heterodimers to be formed, not homodimers, should be used. A preferred embodiment of this aspect of the invention is two different looped loop constructs, such as those derived from naturally occurring leucine zippers such as jun and fos transcription factor proteins (O'Shea et al., Science, 245: 646-648 (1989)).

In another embodiment of the invention, the constant scFv wrist helix can be extended at the C-terminus to obtain a fusion protein. For example, fusion with an enzyme can be used to use such bivalent constructs as diagnostics. Examples of such enzymes are alkaline phosphatase, luciferase or horseradish peroxidase. An advantage of such an antibody enzyme fusion protein is that the bivalent nature of the antibody may result in enhanced binding to the surface-bound antigen. The advantage of using conventional technology (i.e., chemical coupling of the enzyme to the enzyme of choice) may be greater similarity per fusion; homogeneity of the product and simpler preparation, namely in Escherichia coli, in one step.

Similarly, minibodies can be extended at the C-terminus to incorporate a toxin. Such immunotoxins may be bivalent or even bispecific and thus combine the above coupling of such antibody fragments with the benefits of known tumor therapy for immunotoxins. Similarly, the protein or peptide of a metal binding agent may be genetically linked to be used in radioimmunotherapy or in tumor imaging. The same advantages apply to the protein of any genetically encoded hybrid as given above for antibody-enzyme fusions.

In another embodiment of the invention, a scFv wrist-helix construct can be constructed to dimerize with another protein fused to a dimerization domain, in a manner analogous to that described above for bispecific minibodies. In this way, the scFv fragment would, for example, fit into the helix of the fos protein. Such a foreign protein, which can be produced by the formation of heterodimers with scFv fragments, including diagnostic enzymes, toxins, metal binding proteins or peptides, which can be used in radioimmunotherapy or in radioimmunoassay.

Using the principles of the present invention, the dimerization domains disclosed herein may also serve purification purposes. Any type of recombinant protein can be fused to a dimerization domain, such as a wrist zipper. After co-expression with an scFv wrist-jun, the heterodimer can be purified in a single step on a scFV-specific affinity column. Alternatively, the "opposite" zipper, when attached to a column carrier, "traps" the protein wrist zipper when passed through the column in the form of a crude cell extract.

Elution of the pure fusion protein from the column is possible using the temperature required to disrupt the zipper spatial structure. Subsequent separation from the dimerization domain can be achieved by introducing a proteolytic cleavage site into the hinge region, e.g., for coagulation factor Xa (Nagai & Thogerson, Methods in Enzymology, 152, 461-481 (1987)).

A particular advantage of the minibodies disclosed in this invention is their ability to assemble in Escherichia coli functionally. For homobivalent constructs, a dimerization principle is used which allows the formation of homodimers. Examples described above include the looped helix (leucine zipper) of the yeast GCN4 protein, or helices derived from an antiparallel 4 helix bundle. In this case, the scFv fragment is expressed in the presence of a bacterial signal sequence containing codons for a hinge region and dimerization or helix-helix helix at the end of the scFv fragment gene. The helices are compatible with secretion into the periplasmic space of Escherichia coli, where the formation of white space structure, the formation of disulfide bridges and the formation of the final structure take place. Under these conditions, homodimeric proteins are spontaneously formed and can be directly isolated in dimeric form.

If heterobivalent constructs are required, two different scFv fragments or another protein-associated scFv fragment must be associated with each other. In a preferred embodiment of the present invention, both proteins to be assembled are expressed in the same cell, preferably from the same plasmid, preferably one. The design of artificial bicistronic operons is explained, for example, by Skerra et al. (Protein Eng., 4, 971 (1991)). Since the formation of the final structure must take place in the periplasmic space and the scFv fragment can only take up the final space structure in an oxidizing environment, both proteins must be transported and provided with the same signal sequence. The dimerization peptides must be chosen to promote association of the two different proteins, but to prevent association of the corresponding homodimers. Examples of such proteins are the leucine zippers in the fos and jun proteins (see above).

If not expressed in the same cell, the different scFv wrist-zipper constructs should be mixed as a crude cell extract or purified protein and treated at high temperature. The

In the absence of an "opposite" zipper, for example, the scFv-wrist-zipper construction is capable of generating heterodimers. After being held at a melting temperature of about 40 ° C for a short period of time, the zippers of the unwanted homodimer disintegrate and form a much more stable heterodimer (O'Shea et al., Cell, 68, 699-708 (1992)). Without raising the temperature, it is not possible to form heterodimers in vitro, according to experimental results.

BRIEF DESCRIPTION OF THE DRAWINGS AND SEQUENCE LIST

Figure 1 The pLISC-SE scFv expression vector containing the scFv fragment.

Figure 2 The dicistronic scFv hinge zipper expression vector pACKxFyJ.

Figure 3 Functional ELISA.

The concentration of affinity purified proteins measured by OD280 (vertical axis) is related to the molar number of binding sites per well (horizontal axis). ELISA plates were coated with phosphocholine-BSA, and purified phosphocholine-specific minibody proteins were bound and detected with anti-McPC603 antiserum.

(a) Comparison of different minibodies.

(b) Comparison of the scHLXc minibody with scFv and total IgA.

Figure 4 Functional anti-isozyme ELISA.

PC-purified samples of co-expressed anti-PC antilysozyme bispecific minibodies. + and - on the horizontal axis: plus inhibitor (+) and without inhibitor (-).

The attached Sequence Listing contains the following SEQID NOs:

SEQ ID NO: 1: Full nucleotide amino acid sequence of vector pLISC-SE.

SEQ ID NO: 2: Gene cassette (nucleotide and amino acid sequence) of the intercalating GCN4 leucine zipper.

SEQ ID NO: 3: Gene cassette (nucleotide and amino acid sequence) encoding an intercalating antiparallel helix turn helix.

SEQ ID NO: 4: Gene cassette encoding an intercalating jun zipper and IgG3 hinge region.

SEQ ID NO: 5: Gene cassette encoding an intercalating phospho-zipper and IgG3 hinge region.

SEQ ID NO: 6: Gene cassette encoding an intercalating jun zipper and designed link.

SEQ ID NO: 7: Gene cassette encoding an intercalating phospho-zipper and a designed link.

Example 1

Construction of vectors for the construction of secreted single-stranded phage fragments containing a restriction site cleavage site to introduce genes encoding intercalating peptides

Recombinant DNA techniques were used according to Sambrook et al., Maniatis et al., Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY (1989). Functional expression of single-stranded Fv fragments and minibodies in Escherichia coli JM83 was accomplished with vectors similar to pASK-lisc (Skerra et al., 1991, Protein Eng. 4, 971). Site-specific mutagenesis in these vectors was performed according to the method of Kunkel et al., Methods in Enzymology, 154,367-382 (1987) and Geisselsoder et al., (Biotechniques, 5, 786-791 (1987)), M13KO7 helper phage [Vieira & Messing]. , Methods in Enzymology, 153, 3-11 (1987)]. The SDS polyacrylamide assay was performed according to Fiing and Gregerson (Analytical Biochemistry, 155, 83-88 (1986)). The concentration of affinity purified proteins is determined by OD280 using calculated extinction coefficients (Gill & von Hippel, 1989, Analytical Biochemistry, 182, 319-326). A vector such as pASK40 (Skerra et al., Protein Eng., 4, 971 (1991)) is used which has an origin of replication, a regulatory promoter, a bacterial signal sequence (followed by a multiple cloning site), a transcriptional site, terminator and an origin of replication used by single-stranded phages. The gene encoding the single-stranded Fv-ftagment is designed as follows: the nucleotide sequence of a VH domain is directly followed by a linker sequence that preferably encodes about 15 residues, preferably the (Gly4Ser) 3 sequence directly followed by the VL domain. Alternatively, the sequence of the VL domain may be directly followed by the linker sequence, followed by the sequence of the VH domain.

Once the sequence of the antibody is known, the complete gene of the scFv fragment can be constructed from synthetic oligonucleotides. A detailed experimental description of such gene synthesis of an antibody gene is provided, for example, by Plückthun et al., Cold Spring Harbor Symp. Quant. Biol., 52, 105-112 (1987)].

If the genes of the VH and VL domains are on different vectors, the gene for the scFv fragment can be assembled from restriction fragments. For example, a restriction ffag fragment encoding most of the VH domain may be excised from another plasmid, and an ffag fragment encoding the largest portion of the VL domain may be excised from a plasmid. The remainder of VL and VH and the linker of the scFv fragment can be provided by synthetic oligonucleotide cassettes which must be ligated by standard methods (Maniatis et al., Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY, 1982). The mixture of phage fragments is ligated into the vector pASK40 or a similar plasmid containing one pair of suitable restriction enzymes.

If the antibody genes have not yet been cloned, they can be obtained directly from the antibody-producing hybridoma cell using the polymerase chain reaction [PCR; the PCR method is described in McPherson et al., PCR-A Practical Approach Oxford University Press, New York (1991). Numerous starter molecules for amplifying the VH and VL domains have been described in Orlandi et al., 1989, Proceedings of the National Academy of Sciences, 86, 3833-3937 (1989); Huse et al., Science, 246: 1275-1281 (1989); Larrick et al., 1989, Bio / Technology 7: 934-938]. The mRNA

The technique of making the hybridoma from HU 222 983 B1 is cloned into separate vectors and the scFv gene is constructed according to the principles described above.

If the ligated fragments do not result in the scFv fragment in the correct reading phase, precise fusion with the codons of the signal sequence on the plasmid can be accomplished by site-specific mutagenesis. The design of the oligonucleotides and the implementation of the method are simple to those skilled in the art.

The resulting scFv expression plasmid contains codons for a bacterial signal sequence directly followed by the first variable domain (VH or VL) under the control of a regulatory promoter.

At the 3 'end of these genes, corresponding to the C-terminus of the scFv proteins, a unique restriction site is introduced into the expression plasmid to allow insertion of cassettes encoding the intercalating peptides. The restriction site was introduced by site-specific mutagenesis using the method of Kunkel (Methods in Enzymology, 154: 367-382 (1987)).

The complete sequence of a suitable single-stranded Fv expression plasmid capable of accepting an intercalating peptide, pLISC-SE, is shown in SEQ ID NO: 1 (see Sequence Listing), and the map of the plasmid is shown in Figure 1.

Example 2

Design and construction of a gene cassette encoding the intercalating peptides of a leucine zipper

The gene cassette provided with restriction sites that are compatible with the 3 'end of the scFv fragment gene must encode the sequence of a hinge region (attachment of the scFv fragment to the intercalating peptide) as well as the intercalating peptide itself. However, the wrist region can be omitted.

An example is the sequence of the upper hinge region of mouse IgG3 (Dangl et al., EMBO J., 1988, 7, 1989-1994) followed by the leucine zipper sequence of the yeast GCN4 protein (Oas et al., Biochemistry T29, 2891- 2894 (1990)] is converted to the frequently used codon for Escherichia coli (SEQ ID NO: 2). Oligonucleotides were synthesized and ligated into the pLISC-SE vector previously digested with EcoRI and HindIII.

Example 3

Design and construction of a gene cassette encoding the intercalating peptides of a four-helix bundle

As in Example 2, the upper hinge region sequence of mouse IgG3, followed by the helix-bend helix of a quadruple bundle (Eisenberg et al., Proteins, 1, 16-22 (1986)), is transformed into the commonly used Escherichia coli. codon (SEQ ID NO: 3). Oligonucleotides were synthesized and ligated into the pLISC-SE vector previously digested with EcoRI and HindIII.

Example 4

Design, production and co-expression of two gene cassettes encoding intercalating peptides of a leucine zipper

As in Example 2, the upper hinge region sequence of mouse IgG3, followed by the zipper sequence of the jun protein (O'Shea et al., 1992), was converted to the frequently used codon of Escherichia coli (SEQ ID NO: 4). Oligonucleotides are synthesized and ligated into the pLISC-SE vector previously digested with EcoRI and HindIII.

In a parallel reaction, the upper hinge region sequence of mouse IgG3, followed by the zipper sequence of fos protein (O'Shea et al., Cell, 68, 699-708 (1992)), is converted to the frequently used codon of Escherichia coli (SEQ ID NO: 2). NO: 5). Oligonucleotides were synthesized and ligated into the pLISC-SE vector previously digested with EcoRI and HindIII. The two vectors thus encode antibody scFv son judgments, followed by one wrist peptide and one different zipper peptide. To coexpress the two gene cassettes, the entire scFv hinge zipper gene of the fos-containing product is excised from the vector in the form of an XbaI-HindIII fragment and ligated into the pLISC-SE-scFv-jun vector, which already contains the other ) scFv gene.

The newly obtained vector then expresses the scFvlinker-phosph zipper and scFv-linker _2- jun zipper protein from a single promoter, as a two-cistron operon.

For the zip and zip zippers, a better sequence for the wrist region is given in SEQ ID NOs: 6 and 7. This wrist is shorter and therefore insensitive to proteolysis. In cases where the distance between the two binding sites is less important, such shortened joints may be preferred. In this case, the "tail" of the scFv fragment was truncated and the Eco RI site, which hosts the genes encoding the intercalating peptides, was moved up four amino acids.

Example 5

Purification of divalent antibody from Escherichia coli

Escherichia coli JM38 cells, containing plasmids prepared according to Examples 2 and 3, were grown to OD550 = 0.5 and then induced with a final concentration of 1 mM IPTG. The cells were centrifuged, suspended in BBS buffer (200 mM Naborate, 160 mM NaCl, pH 8.0) and passed through a French press. In these examples, a phosphorylcholine binding antibody is used. The minibody is purified by passing it through an affinity chromatography column as described (Chesebro and Metzger, Biochemistry, 11, 766-771 (1972)).

Example 6

Purification of bispecific antibody from Escherichia coli

Escherichia coli JM83 cells containing plasmids prepared according to Examples 2 and 3, which carry two different scFv-encoding dicistronic structural genes (Figure 2), were grown to OD550 = 0.5, followed by IPTG with a final concentration of 1 mM. . Cells were centrifuged and resuspended in BBS buffer (200 mM Na borate, 160 mM Na

Brine chloride, pH 8.0) and the suspension was passed through a French press.

In this example, a bispecific mininant antibody having both phosphorylcholine and benzoylampicillin specificity is used. The minibody as described in Chesebro and Metzger, Biochemistry,

11, 766-771 (1972)] by purification through an affinity chromatography column.

Example 7

Binding of divalent minibodies to surface

ELISA plates (Nunc, Macrosorp) were coated with phosphocholine-BSA at a concentration of 400 µg / ml in PBS buffer (20 mM phosphate, pH 7.2, 115 mM sodium chloride). The hapten reagent was prepared from nitrophenylphosphocholine (SIGMA), which was reduced and diazotized according to Chesebro and Metzger, Biochemistry, 11, 766-771 (1972) and coupled to BSA (SIGMA) in borate buffer (52.5 mmol). (1 L sodium borate, pH 9, 20-120 mmol / L sodium chloride) at 4 ° C for 48 hours and then dialyzed against PBS. The surface of the uncoated plate was blocked with 5% skimmed milk (Nestle) in PBS buffer for at least 2 hours and then the periplasmic extract of the purified protein was incubated for 90 minutes at room temperature in BBS buffer. After thorough washing (three times), residual functional antibody fragments were detected using standard procedures (Harlow & Lane, "Antibodies, A Laboratory Manual", Cold Spring Harbor Laboratory, 555-592 (1988)), rabbit anti-McPC603 serum and peroxidase-linked anti-rabbit immunoglobulin. as described by Gallati, Clin. Chem. Clin. Biochem.

17.1-4. (1979)].

Compared to the monomeric scFv fragments, much better binding and hence sensitivity to all the minibody constructs can be observed. This is equivalent to having two or more binding sites simultaneously on the same surface. This avidity of the scHLXc fusion protein is comparable to that of the native antibody McPC603, which was detected by antigen-coated ELISA, whereas the monomeric scFv fragment was only detected at a hundred-fold higher concentration (Figures 3a and 3b). Each of the bonds is almost completely inhibited by soluble hapten except for the monomeric scFv fragment. The natural antibody has a thermodynamic affinity for soluble phosphocholine of about Ι, όχΙΟ ⁵ (mol / l) -1 and is thus relatively weak [Metzger et al., Proceedings of the Lst 15 Congress of Immunology, New York, 253-267 (1971)] and apparently is not sufficient for a monomeric fragment-hapten complex to survive repeated washing steps in a functional ELISA [Chimney & Challacombe, ELISA and other solid phase immunoassays, Wiley & Sons, New York (1988). )].

Example 8

Surface binding of bifunctional minibodies

Co-expressed bifunctional minibodies recognizing phosphorylcholine with one arm and lysozyme recognizing the other with their arms are purified by phosphocholine (PC) affinity chromatography and assayed for lysozyme specificity. An ELISA plate is coated with lysozyme and the ELISA is performed as described in Example 8. Three different formulations show binding to the antigen surface which is completely inhibited by soluble lysozyme (Figure 4).

The sequences to which we refer in the text are given below (the Hungarian translation of the English subtitles is given at the end of the list).

SEQUENCE LIST

SEQ ID NO: 1

ACCCGACACC ATCGAATGGC GCAAAACCTT TCGCGGTATG GCATGATAGC 50

GCCCGGAAGA GAGTCAATTC AGGGTGGTGA ATGTGAAACC AGTAACGTTA 100

TACGATGTCG CAGAGTATGC CGGTGTCTCT TATCAGACCG TTTCCCGCGT 150

GGTGAACCAG GCCAGCCACG TTTCTGCGAA AACGCGGGAA AAAGTGGAAG 200

CGGCGATGGC GGAGCTGAAT TACATTCCCA ACCGCGTGGC ACAACAACTG 250

GCGGGCAAAC AGTCGTTGCT GATTGGCGTT GCCACCTCCA GTCTGGCCCT 300

GCACGCGCCG TCGCAAATTG TCGCGGCGAT TAAATCTCGC GCCGATCAAC 350

TGGGTOCCAG CTGTGTGGTG TCGATGGTAG AACGAAGCGG CGTCGAAGCC 400

TGTAAAGCGG CGGTGCACAA TCTTCTCGCG CAACGCGTCA GTGGGCTGAT 450

CATTAACTAT CCGCTGGATG ACCAGGATGC CATTGCTGTG GAAGCTGCCT 500

GCACTAATGT TCCGGCGTTA TTTCTTGATG TCTCTGACCA GACACCCATC 550

AACAGTATTA TTTTCTCCCA TGAAGACGGT ACGCGACTGG GCGTGGAGCA 600

TCTGGTCGCA TTGGGTCACC AGCAAATCGC GCTGTTAGCG GGCCCATTAA 650

GTTCTGTCTC GGCGCGTCTG CGTCTGGCTG GCTGGCATAA ATATCTCACT 700

CGCAATCAAA TTCAGCCGAT AGCGGAACGG GAAGGCGACT GGAGTGCCAT 750

GTCCGGTTTT CAACAAACCA TGCAAATGCT GAATGAGGGC ATCGTTCCCA 800

CZGCGATGCT GGTTGCCAAC GATCAGATGG CGCTGGGCGC AATGCGCGCC 850

ATTACCGAGT CCGGGCTGCG CGTTGGTGCG CATGTCTCGG TAGTGGGATA 900

CGCAGATACC GAAGACAGCT CATGTTATAT CCCGCCGTTA ACCACCATCA 950

AACAGGATTT TCGCCTGCTG GGGCAAACCA GCGTGGACCG CTTGCTGCAA 1000

CTCTCTCAGG GCCAGGCGGT GAAGGGCAAT CAGCTGTTGC CCGTCTCACT 1050

GGTGAAAAGA AAAACCACCC TGGCGCCCAA TACGCAAACC GCCTCTCCCC 1100

GCGCGTTGGC CGATTCATTA ATGCAGCTGG CACGACAGGT TTCCCGACTG 1150

HU 222 983 Bl

GAAAGCGGGC AGTGAGCGCA ACGCAATTAA TGTGAGTTAG CTCACTCATT 1200 AGGCACCCCA GGCTTTACAC TTTATGCTTC CGGCTCGTAT GTTGTGTGGA 1250 ATTGTGAGCGTGATGATGATGATGATTA

Met Lys Lys Thr Alá lle 1 5

GCG ATT GCA GTG With 10 GCA Al a CTG Leu GCT Al a GGT Gly TTC Phe 15 GCT Al a ACC Thr GTA With GCG Al a CAG Gin 20 1387 Al a Ile Al a GCC GAA GTT AAA CTG GTA GAG TCT GGT GGT GGT CTG GTA CAG 1429 Al a Gl u With Lys Leu 25 With Glu Ser Gly Gly 30 Gly Leu With Gin CCG GGT GGA TCC CTG CGT CTG TCT TGC GCT ACC TCA GGT TTC 1471 Pro 35 Gly Gly Ser Leu Arg 40 Leu Ser Cys Al a Thr 45 Ser Gly Phe ACC TTC TCT GAC TTC TAC ATG GAG TGG GTA CGT CAG CCC CCG 1513 Thr Phe 50 Ser asp Phe Tyr Me t 55 Glu Trp With Arg Gin 60 Pro Pro GGT AAA CGT CTC GAG TGG ATC GCA GCT AGC CGT AAC AAA GGT 1555 Gly Lys Arg 65 Leu Glu Trp Ile Al a 70 Al a Ser Arg Asn Lys 75 Gly AAC AAG TAT ACC ACC GAA TAC AGC GCT TCT GTT AAA GGT CGT 1597 Asn Ly s Tyr Thr 80 Thr Glu Tyr Ser Al Ser With Lys Gly Arg 90 TTC ATC GTT TCT CGT GAC ACT AGT CAA TCG ATC CTG TAC CTG 1639 Phe Ile With Ser Arg 95 asp Thr Ser Gin Ser 100 Ile Leu Tyr Leu CAG ATG AAT GCA TTG CGT GCT GAA GAC ACC GCT ATC TAC TAC 1681 Gin 105 Me t Asn Al a Leu Arg 110 Al a Glu asp Thr Al Ile Tyr Tyr TGC GCG CGT AAC TAC TAT GGC AGC ACT TGG TAC TTC GAC GTT 1723 Cys Al at 120 Arg Asn Tyr Tyr Gly 125 Ser Thr Trp Tyr Phe 130 asp With TGG GGT GCA GGT ACC ACC GTT ACC GTT TCT TCT GGT GGT GGT 1765 Trp Gl y Al a 135 Gly Thr Thr With Thr 140 With Ser Ser Gly Gly 145 Gly GGT TCT GGT GGT GGT GGT TCT GGT GGT GGT GGT TCT DAM ATC 1807 Gly Ser Gly Gly 150 Gly Gly Ser Gly Gly 155 Gly Gly Ser asp Ile 160 GTT ATG ACC CAG TCT CCG AGC TCT CTG TCT GTA TCT GCA GGT 1849 With Me t Thr Gin Ser 165 Pro Ser Ser Leu Ser 170 With Ser Al a Gly GAA CGT GTT ACC ATG TCT TGC AAA TCT TCT CAG TCT CTG CTG 1891 Glu Arg With Thr Me t Ser Cys Ly s Ser Ser Gin Ser Leu Leu

175 180 185

HU 222 983 Bl

AAC TCT GGT AAC CAG AAA AAC TTC CTG GCG TGG TAT CAG CAA 1933

Asn Ser Gly Asn Gin Lys Asn Phe Leu Alá Trp Tyr Gin Gin

190 195 200

AAG CCT GGC CAA CCG CCG AAA CTG CTG ATC TAC GGT GCG TCG 1975

Lys Pro Gly Gin Pro Pro Lys Leu Leu Ile Tyr Gly Alá Ser

205 210 215

ACC CGT GAA TCT GGT GTT CCG GAC CGT TTT ACC GGT AGC GGT 2017

Thr Arg Glu Ser Gly Val Pro Asp Arg Phe Thr Gly Ser Gly

220 225 230

AGC GGT ACC GAC TTC ACT CTG ACC ATC TCT TCT GTA CAG GCT 2059

Ser Gly Thr Asp Phe Thr Leu Thr Ile Ser Ser Val Gin Alá

235 240

GAA GAT CTG GCT GTT TAC TAC TGT CAA AAC GAC CAC TCT TAC 2101

Glu Asp Leu Alá Val Tyr Tyr Cys Gin Asn Asp His Ser Tyr

245 250 255

CCG CTG ACC TTT GGC GCC GGC ACC AAA CTG GAA CTG AAG CGC 2143

Pro Leu Thr Phe Gly Alá Gly Thr Lys Leu Glu Leu Lys Arg

260 265 270

GCT AAC GGT GAA TTC-TGATAAGCTTGA CCTGTGAAGT GAAAAATGGC 2190

Alá Asn Gly Glu Phe *

275

GCACATTGTG CGACATTTTT TTTGTCTGCC GTTTACCGCT ACTGCGTCAC 2240

GGATCCCCAC GCGCCCTGTA GCGGCGCATT AAGCGCGGCG GGTGTGGTGG 2290

TTACGCGCAG CGTGACCGCT ACACTTGCCA GCGCCCTAGC GCCCGCTCCT 2340

TTCGCTTTCT TCCCTTCCTT TCTCGCCACG TTCGCCGGCT TTCCCCGTCA 2390

AGCTCTAAAT CGGGGCATCC CTTTAGGGTT CCGATTTAGT GCTTTACGGC 2440

ACCTCGACCC CAAAAAACTT GATTAGGGTG ATGGTTCACG TAGTGGGCCA 2490

TCGCCCTGAT AGACGGTTTT TCGCCCTTTG ACGTTGGAGT CCACGTTCTT 2540

TAATAGTGGA CTCTTGTTCC AAACTGGAAC AACACTCAAC CCTATCTCGG 2590

TCTATTCTTT TGATTTATAA GGGATTTTGC CGATTTCGGC CTATTGGTTA 2640

AAAAAATGAGC TGATTTAACA AAAATTTAAC GCGAATTTTA ACAAAATATT 2690

AACGTTTACA ATTTCAGGTG GCACTTTTCG GGGAAATGTG CGCGGAACCC 2740

CTATTTGTTT ATTTTTCTAA ATACATTCAA ATATGTATCC GCTCATGAGA 2790

CAATAACCCT GATAAATGCT TCAATAATAT TGAAAAAGGA AGAGTATGAG 2840

TATTCAACAT TTCCGTGTCG CCCTTATTCC CTTTTTTGCG GCATTTTGCC 2890

TTCCTGTTTT TGCTCACCCA GAAACGCTGG TGAAAGTAAA AGATGCTGAA 2940

GATCAGTTGG GTGCACGAGT GGGTTACATC GAACTGGATC TCAACAGCGG 2990

TAAGATCCTT GAGAGTTTTC GCCCCGAAGA ACGTTTTCCA ATGATGAGCA 3040

CTTTTAAAGT TCTGCTATGT GGCGCGGTAT TATCCCGTAT TGACGCCGGG 3090

CAAGAGCAAC TCGGTCGCCG CATACACTAT TCTCAGAATG ACTTGGTTGA 3140

GTACTCACCA GTCACAGAAA AGCATCTTAC GGATGGCATG ACAGTAAGAG 3190

AATTATGCAG TGCTGCCATA ACCATGAGTG ATAACACTGC GGCCAACTTA 3240

CTTCTGACAA CGATCGGAGG ACCGAAGGAG CTAACCGCTT TTTTGCACAA 3290

CATGGGGGAT CATGTAACTC GCCTTGATCG TTGGGAACCG GAGCTGAATG 3340

AAGCCATACC AAACGACGAG CGTGACACCA CGATGCCTGT AGCAATGGCA 3390

ACAACGTTGC GCAAACTATT AACTGGCGAA CTACTTACTC TAGCTTCCCG 3440

GCAACAATTA ATAGACTGGA TGGAGGCGGA TAAAGTTGCA GGACCACTTC 3490

TGCGCTCGGC CCTTCCGGCT GGCTGGTTTA TTGCTGATAA ATCTGGAGCC 3540

GGTGAGCGTG GGTCTCGCGG TATCATTGCA GCACTGGGGC CAGATGGTAA 3590

GCCCTCCCGT ATCGTAGTTA TCTACACGAC GGGGAGTCAG GCAACTATGG 3640

ATGAACGAAA TAGACAGATC GCTGAGATAG GTGCCTCACT GATTAAGCAT 3690

TGGTAACTGT CAGACCAAGT TTACTCATAT ATACTTTAGA TTGATTTAAA 3740

ACTTCATTTT REFUNDS GGATCTAGGT GAAGATCCTT TTTGATAATC 3790

HU 222 983 ΒΙ

TCATGACCAA AATCCCTTAA CGTGAGTTTT CGTTCCACTG AGCGTCAGAC CCCGTAGAAA AGATCAAAGG ATCTTCTTGA GATCCTTTTT TTCTGCGCGT AATCTGCTGC TTGCAAACAA AAAAACCACC GCTACCAGCG GTGGTTTGTT TGCCGGATCA AGAGCTACCA ACTCTTTTTC CGAAGGTAAC TGGCTTCAGC AGAGCGCAGA TACCAAATAC TGTCCTTCTA GTGTAGCCGT AGTTAGGCCA CCACTTCAAG AACTCTGTAG CACCGCCTAC ATACCTCGCT CTGCTAATCC TGTTACCAGT GGCTGCTGCC AGTGGCGATA AGTCGTGTCT TACCGGGTTG GACTCAAGAC GATAGTTACC GGATAAGGCG CAGCGGTCGG GCTGAACGGG GGGTTCGTGC ACACAGCCCA GCTTGGAGCG AACGACCTAC ACCGAACTGA GATACCTACA GCGTGAGCTA TGAGAAAGCG CCACGCTTCC CGAAGGGAGA AAGGCGGACA GGTATCCGGT AAGCGGCAGG GTCGGAACAG GAGAGCGCAC GAGGGAGCTT CCAGGGGGAA ACGCCTGGTA TCTTTATAGT CCTGTCGGGT TTCGCCACCT CTGACTTGAG CGTCGATTTT TGTGATGCTC GTCAGGGGGG CGGAGCCTAT GGAAAAACGC CAGCAACGCG GCCTTTTTAC GGTTCCTGGC CTTTTGCTGG CCTTTTGCTC ACATG

SEQ ID NO: 2

GGT GAA TTC CCC AAA CCT AGT ACT CCC CCT GGC AGC AGC CGC ATG

Gly Glu Phe Pro Lys Pro Ser Thr Pro Pro Gly Ser Ser Arg Met

T 5 10 15 l ----------- IgG3-hinge -------------------- ^{1 1} ------ AAA CAG CTG GAA GAT AAA GTT GAA GAG CTT CTT TCG AAA AAC TAC

Lys Gin Leu Glu Asp Lys Val Glu Glu Leu Leu Ser Lys Asn Tyr

25 30

----------------------- ------------------------ GCN4-zipper -CAC CTC GAA AAT GAA GTT GCG CGC CTC AAA AAA CTT GTT GGT GAA His Leu Glu Asn Glu Val Ala Arg Leu Lys Lys Leu Val Gly Glu

40 45

3840

3890

3940

3990

4040

4090

4140

4190

4240

4290

4340

4390

4440

4490

4515

135

CGC TGATAAGCTT GAC Arg t

--- ¹ stop

151

SEQ ID NO: 3

GGT GAA TTC CCC AAA CCT AGC ACC CCC CCT GGC AGC AGT GGT GAA Gly Glu Phe Pro Lys Pro Ser Thr Pro Pro Gly Ser Ser Gly Glu ι T 5 10 15 l --------- igG3-Hinge -------------------- ^{1 1} ------- CTG GAA GAG CTG CTT AAG CAT CTT AAA GAA CTT CTG AAG GGC CCC Leu Glu Glu Leu Leu Lys His Leu Lys Glu Leu Lys Gly Pro 20 25 30 ------ bundle-helix A ----- _] L CGC AAA GGC GAA CTC GAG GAA CTG CTG AAA CAT CTG AAG GAG CTG Arg Lys Gly Glu Leu Glu Glu Leu Leu Lys His Leu Lys Glu Leu 35 40 45 my needle --- ¹ ι ------------------ bundle-helix B ---------------------- CTT AAA GGT GAA TTC TGATAAGCTT GACCTGTGAA GTGAAAAAAT G Leu Lys Gly Glu Phe t 50 t ------------ 1 stop SEQ ID NO: 4 GGT GAA TTC CCC AAA CCT AGT ACT CCC CCT GGC AGC AGC CGT ATC Gly Glu Phe Pro Lys Pro Ser Thr Pro Pro Gly Ser Ser Arg Ile 1 hr 5 10 15 ι-IgG3 hinge ---- ---- J L GCT CGT CTC GAG GAA AAA GTT AAA ACC CTG AAA GCT CAG AAC TCC Alá Arg Leu Glu Glu Lys Val Lys Thr Leu Lys Alá Gin Asn Ser

25 30 jun-zipper

135

191

HU 222 983 Bl

GAA CTG GCT TCC ACC GCT below AAC ATG CTG CGT GAA CAG GTT GCT CAG Gin 45 135 Glu Leu below Ser Thr 35 Asn Not t Leu Arg 40 Glu Gin With Al a CTG AAA CAG AAA GTT ATG AAC TAC TGATAAGCTT GACCTGTGAA G 180 Leu Ly s Gin Lys With Me t Asn Tyr 50 T _l Stop SEQ ID NO: ' GGT GAA TTC CCC AAA CCT AGT ACT CCC CCT GGC AGC AGC CTG ACC 45 Gly Glu Phe Pro Lys Pro Ser Thr Pro Pro Gly Ser Ser Leu Thr 1 t 5 10 15 L - ---- IgG3 hinge J L GAC ACC CTG CAG GCT GAA ACC GAC CAG CTG GAA GAC AAA AAA TCC 90 asp Thr Leu Gin Al a Glu Thr asp Gin Leu Glu As p Lys Ly s Ser

25 30

----------------------- fos-zipper

GCT CTG CAG ACC GAA ATC GCT AAC CTG CTG AAA GAA AAA GAA AAA 135 Al a Leu Gin Thr Glu Ile Al a Asn Leu Leu Ly s Glu Ly s Glu Lys

40 45 fos-zipper

CTG GAA TTT ATC CTG GCT GCT TAC TGATAAGCTT GACCTGTGAA G 180 Leu Glu Phe Ile Leu 50 Al a Al a Tyr _l Stop t SEQ ID NO: < 6 GGT GAA TTC CCG TCT GGT AAC GAA GCT CGT ATC GCT CGT CTC GAG 45 Gly Glu Phe Pro Ser Gly Asn Glu Al a Arg Ile Alá Arg Leu Glu 1 t 5 10 15 1____ - linker - _l L GAA AAA GTT AAA ACC CTG AAA GCT CAG AAC TCC GAA CTG GCT TCC 90 Glu Ly s With Lys Thr Leu Ly s Al a Gin Asn Ser Glu Leu Alá Ser 20 25 30

----------------------- ------------------------ jun-zipper --- ACC GCT AAC ATG CTG CGT GAA CAG GTT GCT CAG CTG AAA CAG AAA 135

Thr Alá Asn Me t Leu Arg Glu Gin Val Alá Gin Leu Lys Gin Lys

40 45

----------------------- ------------------------ jun-zipper --- GTT ATG AAC TAC TGATAAGCTT GACCTGTGAA GTGAAAAATG GCG 180

Val Met Asn Tyr t

--------------- 1 _s top

SEQ ID NO: 7

GGT GAA TTC GGT CCG TCT GGT AAC GAA CTG ACC GAC ACC CTG CAG 45

Gly Glu Phe Gly Pro Ser Gly Asn Glu Leu Thr Asp Thr Leu Gin t 5 10 15 ¹ ---- linker -------------- ^{1 1} ------ fos-zipper -------- GCT GAA ACC GAC CAG CTG GAA GAC AAA AAA TCC GCT CTG CAG ACC 90

Alá Glu Thr Asp Gin Leu Glu Asp Lys Lys Ser Alá Leu Gin Thr

25 30

------------------- f _os - _z ipper -------------------------- ----- GAA ATC GCT AAC CTG CTG AAA GAA AAA GAA AAA CTG GAA TTT ATC 135

Glu Ile Alá Asn Leu Leu Lys Glu Lys Glu Lys Leu Glu Phe Ile

40 45

---------------------------- ------------------- fos-zipper --- 12

HU 222 983 Bl

CTG GCT GCT TAC TGATAAGCTT GACCTGTGAA GTGAAAAATGGCG 180

Leu Alá Alá Tyr t

--------------- 1 _s top

Translation of English subtitles in the Sequence Listing:

"IgG3 Hinge": IgG3 Hinge "GCN4 Zipper": GCN4 Zipper "Bundle Helix": Bundle Helix

Jun-zipper ": jun zipper" fos-zipper ": fos zipper" linker ": linking peptide

Claims (18)

PATENT CLAIMS CLAIMS 1. A monomeric antibody fragment fusion protein comprising a Fv fragment of an antibody and a peptide capable of dimerizing by non-covalent interactions with another peptide. 20 The monomer of claim 1, wherein the Fv moiety is a single chain ffagment. Monomer according to claim 1 or 2, characterized in that the interactive peptide comprises 10-50, preferably 10-30 amino acids. 25 4. A monomer according to any one of claims 1 to 4, characterized in that the peptide comprises at least one helix. 5. The monomer of claim 4, wherein the helix peptide is of helix-other helix structure. 6. The monomer of claim 4, wherein said peptide comprises a leucine zipper molecule comprising a plurality of repeating amino acids, of which each seventh is leucine. 35 The monomer of claim 4, wherein the peptide contains charged groups. 8. A monomer according to any one of claims 1 to 3, characterized in that a linker peptide is present between the Fv fragment and the peptide. 40 9. The monomer of claim 8, wherein the linker peptide is a hinge peptide of an antibody or fragment thereof. 10. A dimeric fusion protein comprising two monomeric fusion proteins in which the non-covalent interaction between the monomeric units 45 is provided by the same or different peptides, wherein at least one of the monomeric units is one of the amino acids 1-9. The antibody fragment fusion protein of any one of claims 1 to 3. 50 11. The dimeric fusion protein of claim 10, wherein the interacting peptides are identical. 12. The dimer according to claim 10 or 11, wherein the second monomer unit is a dimer according to any one of claims 1-9. The antibody fragment fusion protein of any one of claims 1 to 6, which has a different specificity. 13. The dimer of claim 10 or 11, wherein the second monomer unit is a dimer according to any one of claims 1-9. The fusion protein of any one of claims 1 to 6, wherein the antibody fragment (Fv) is replaced by a non-antibody protein or peptide. 14. The dimer of claim 13, wherein the protein or peptide is a toxin, a chelator peptide, a metal binding protein, or an enzyme, or the corresponding specific binding site. 15. The dimer of claim 13, wherein the protein or peptide has an α-cell or T-cell fragment specific binding site. 16. A dimer according to any one of claims 1 to 5, characterized in that another peptide is fused to the C-terminus of one or another intercalating peptide. 17. The dimer of claim 16, wherein the fused protein is a toxin, a chelator peptide, a metal binding protein, or an enzyme, or has a corresponding specific binding site or a T cell (fragment) specific binding site. 18. The construction kit is illustrated in FIGS. The antibody fragment for producing selective dimers of fusion proteins according to any one of claims 1 to 8, which (a) is an antibody fragment of claims 1-8. The monomeric antibody fragment of any one of claims 1 to 6, and (b) a second monomeric fusion protein as defined in (a), wherein the antibody fragment has the same or different antigen specificity, or wherein the antibody fragment unit is replaced by a non-antibody protein or peptide.