Engineered heterodimeric protein domains

[0001] The present invention relates to engineered heterodimeric immunoglobulin domains and methods for their production.

Background Art [0002] Nature provides many heterodimeric proteins and protein domains that belong to families of related proteins. Such proteins and domains often form homodimers on their own, but do not form heterodimers with members of other families. On the other hand, heterodimeric and heteromultimeric proteins are often useful. They provide new therapeutic agents and research tools. For example, bispecific antibodies ( bispecific antibodies BsAbs) capable of binding to at least two different antigens have significant potential in a wide spectrum of clinical applications as target agents for immunodiagnosis and in vitro and in vivo therapy and for diagnostic immunoassays. In the field of diagnostics, BsAbs have proved to be very useful in probing the functional properties of cell surface molecules and in defining the ability of various Fc receptors to mediate cytotoxicity (publication: Fanger et al. (1992) Crit. Rev. Immunol, 12: 101-124; which the scope of the disclosure is hereby incorporated by reference into this description). However, during the simple generation of BsAbs by the co-expression of a wide variety of components that can interact without specificity, a large number of species are often generated and it is often difficult to separate desired species from undesirable species. Therefore, it is desirable to provide techniques for efficiently producing heteromultimers. It is particularly desirable to provide a method for generating antibody subunits that form heterodimers preferentially to homodimer formation so that BsAbs can be effectively recovered from recombinant cell culture.

[0003] The prior art contains reports of methods for producing heterodimeric proteins. For example, Stahl and Yancopoulos described the use of fusion proteins containing two different receptor subunits to create

- soluble heterodimeric receptors that could bind to a given circulating cytokine and thereby block the activity of that cytokine (see US Patent Application Publication No. 6,472,179). In the publication of the US patent application: Carter et al. US 5,807,706; describes a "protuberance-into-cavity" approach for generating a heterodimeric Fc unit.

[0004] Such existing methods allow the construction of individual heterodimers but do not provide general techniques for constructing multimeric proteins involving the interaction of various domains. Therefore, there is a need to provide a generic system for designing heterodimeric pairs that can be specifically assembled in an environment containing many different potential mating partners.

SUMMARY OF THE INVENTION [0005] The invention provides a new approach to designing immunoglobulin domains that preferentially heterodimerize. In particular, the invention uses a strategy for engineering domains constructed with thread exchange (Strand Exchange constructed with engineering methods).

Domain - SEED) for the engineering construction of a protein-protein interaction interface within the heterodimerization of said immunoglobulin domains. The invention also provides immunoglobulins containing domains constructed by engineering methods using the method of the invention.

[0006] According to one aspect, the invention relates to a multi-domain heterodimeric immunoglobulin comprising at least a first and a second non-identical engineered domain, each of which contains a protein-protein interaction interface containing amino acid sequence segments derived from two or more naturally occurring homologous parent domains, thereby giving the first and second engineered domains the specificity of splicing these parent domains, said first and second engineered domains forming heterodimers in a preferential manner over homodimers (for example, heterodimers represent more than 55%, 65%, 75 %, 80%, 85%, 90% or 95% of the total dimers). First and second domain

Engineered methods are not antibody variable domains. In some embodiments, the multi-domain immunoglobulin of the invention comprises a first subunit comprising a first engineered domain and a second subunit comprising a second engineered domain. As used herein, the expression "amino acid sequence segment" means any sequence segment containing two or more amino acids (for example, three or more, four or more, five or more, six or more, seven or more, eight or more, nine or more or ten or more).

[0007] In preferred embodiments of the invention, said multi-domain immunoglobulin comprises non-identical engineered domains from naturally occurring homologous parent domains, for example, CH3 antibody domains. In particular, the engineered domains are derived from the IgG and IgA CH3 domains.

[0008] In certain preferred embodiments of the invention, the multi-domain immunoglobulin according to the invention comprises engineered domains that are part of polypeptide chains that are connected by a disulfide bond.

[0009] In one embodiment of the invention, one of the engineered domains in said multi-domain immunoglobulin according to the invention comprises at least two non-adjacent sequence segments derived from the same parent domain. In another embodiment, each of the first and second engineered domains comprises at least two, three or four or more non-adherent sequence segments derived from the same parent domain. According to another embodiment of the invention, at least one of the engineered domains comprises sequence segments from each parent domain that are at least two amino acids in length. In another embodiment of the invention, at least one of the engineered domains comprises sequence segments from each parent domain that are at least three, four, five or six amino acids in length.

[0010] In certain embodiments of the invention, the multi-domain immunoglobulin according to the invention comprises a first bioactive domain. Such a first bioactive domain may occupy the N-terminal or C-terminal position relative to said first engineered domain.

[0011] According to further embodiments of the invention, the multi-domain immunoglobulin may further comprise a second bioactive domain in addition to the first bioactive domain. In one embodiment of the invention, the second bioactive domain is associated with a second engineered domain and may occupy the N-terminus or C-terminus position relative to the second engineered domain. In an alternative embodiment of the invention, the second bioactive domain is also associated with the first engineered domain and may occupy the opposite position to the first bioactive domain. For example, the first and second bioactive domains may occupy positions N and C, respectively, relative to the first engineered domain.

[0012] The multi-domain immunoglobulin according to the invention can be used to generate bispecific antibodies. For example, a multi-domain protein may comprise a first bioactive domain containing an antibody variable domain and a second bioactive domain containing a second antibody variable domain with different specificity.

[0013] In another aspect, the invention provides a multi-domain immunoglobulin, wherein the first bioactive region comprises two or more antibody variable domains with a first specificity or a first combination of specificities. Such a multi-domain protein may also contain a second bioactive region containing two or more antibody variable domains with a second specificity or with a second combination of specificities. For example, such a multi-domain protein may contain one or more single chain Fv units, a diabody (one VH-VL chain), a single chain diabody [VH (1) - VL (2) ----- VH (2) - VL (1) ] or other single chain repeats of fused linked single chain Fvs (with the same or different specificities).

[0014] In another aspect, the invention provides a multi-domain immunoglobulin, wherein the first bioactive region comprises two or more antibody variable domains with a first specificity or a first combination of specificities. Such multi-domain immunoglobulin may further comprise a second bioactive region containing two or more antibody variable domains with a second specificity or with a second combination of specificities that are substantially different from the first combination of specificity.

[0015] The present invention further provides a method for co-localizing bioactive domains when administered to a biological system. The method comprises the step of administering to the biological system a heterodimeric immunoglobulin comprising a first and a second bioactive domain as described in the various above embodiments of the invention. In one embodiment of the invention, such biological system is a mammal. In a more preferred embodiment, the biological system is a human.

[0016] In another aspect, the invention provides a multi-domain heterodimeric immunoglobulin comprising at least a first and a second non-identical engineered domain that are in contact at an interface. The interface of the first engineered domain contains at least two segments of the amino acid sequence, each segment derived from a different naturally occurring homologous parent domain, thereby giving a splicing specificity different from the parent splicing specificity, with the first and second engineered domains forming heterodimers. In a preferred embodiment of the invention, the second engineered domain also comprises at least two amino acid sequence segments, each segment being derived from a different naturally occurring homologous parent domain, thereby giving a splicing specificity different from the splicing specificity of the parent domains, the first and a second engineered domain formed by heterodimers.

[0017] In yet another aspect, the present invention provides a multi-domain immunoglobulin comprising at least the first and second engineered domains that are in contact at an interface,

- (1) that the first and second engineered domains are derived from two or more naturally occurring homologous parent domains, (2) the interface from the first engineered domain contains at least one segment of the amino acid sequence entering into interaction with the segment of the amino acid sequence on the interface of the second engineered domain derived from the same parent domain, and (3) the first and second engineered domains form heterodimers.

[0018] In a further aspect, the present invention provides an engineered immunoglobulin domain comprising a protein-protein interaction interface containing amino acids from two or more parent immunoglobulin domains, such that this protein-white interaction interface confers such an engineered immunoglobulin domain of splicing specificity different from the specificity of splicing immunoglobulin parent domains, wherein such an engineered immunoglobulin domain is not an antibody variable domain. In preferred embodiments, the engineered immunoglobulin domain is assembled with a partner domain with improved specificity compared to parent domains. In certain embodiments, the partner domain is an engineered immunoglobulin domain of the invention.

In yet another aspect, the present invention provides an engineered immunoglobulin superfamily domain comprising a protein-protein interaction interface containing amino acids from two or more parental immunoglobulin superfamily parent domains, such that this protein-protein interaction interface gives this engineered domain immunoglobulin superfamily interaction properties, which are different from the interaction properties of the parent immunoglobulin superfamily domains.

[0020] The invention also provides a multi-domain immunoglobulin comprising an engineered domain with the following properties. First, the engineering-constructed domain contains a protein-protein interaction interface. Secondly, such a domain constructed by engineering methods

Homologous to the family of naturally occurring domains, preferably in such a way that the amino acid sequence of such engineered domain can be aligned with the amino acid sequences of naturally occurring domains that can be further aligned with each other. Preferably, such alignment of the amino acid sequences of such naturally occurring domains corresponds to alignment of the three-dimensional structures of naturally occurring domains. Third, the interface of interaction of such engineered domain contains amino acids from corresponding sequential positions from two or more naturally occurring parent domains. Fourth, the amino acids in the interface of such a engineered domain considered as a group are not all found in the corresponding interface of any single member of homologous naturally occurring domains. Fifth, the interaction interface of such a domain-engineered interface confers properties different from any parent domain. Preferably, the assembly properties of such engineered domain are different because the interaction interface has amino acids from two or more different parents who create specific contacts with assembly partners, thereby acquiring assembly specificity that is hybrid between assembly domain specificity.

[0021] Furthermore, the present invention provides a nucleic acid encoding a multi-domain immunoglobulin as described in the various above embodiments of the invention. In particular, the invention provides a nucleic acid encoding a multi-domain protein containing at least one bioactive domain. The invention also provides cells comprising the nucleic acid of the invention.

[0022] In another aspect, the present invention provides a method of designing multi-domain immunoglobulin with domains that heterodimerize. The method comprises the following steps: (a) selecting a first polypeptide, a second polypeptide, a third polypeptide and a fourth polypeptide, wherein the first and third polypeptides dimerize with each other but do not dimerize with the second or fourth polypeptide, and wherein the second and fourth polypeptides dimerize with each other, (b) composing sequences

- the amino acid of the first domain of the first and second polypeptides comprising at least one component consisting of the first polypeptide, and (c) composing the amino acid sequence of the second domain of the third and fourth polypeptides comprising at least one component consisting of the third polypeptide, yes that the components consisting of the first and third polypeptides assemble together, promoting heterodimerization of the first and second domains.

[0023] Other features, objects and advantages of the present invention will become apparent upon reading the detailed description of the invention below. Nevertheless, it should be understood that this detailed description, indicating preferred embodiments of the invention, is illustrative only, without limiting the scope of the invention in any way. A variety of changes and modifications within the scope of the invention are apparent to those skilled in the art of the invention and are based on the following detailed description.

[0024] In summary, the essence of the present invention is:

• Multi-domain heterodimeric immunoglobulin containing at least the first and second non-identical engineered domains, each of the first and second engineered domains containing a protein-protein interaction interface containing amino acid sequence segments derived from two or more naturally occurring homologous parent domains, thanks to which the first and the second engineered domain of splicing specificity differ from the parent domain's splicing specificity, whereby the first and second engineered domains form heterodimers.

• A suitable multi-domain immunoglobulin, wherein the multi-domain immunoglobulin comprises a first subunit containing a first engineered domain and a second subunit comprising a second engineered domain;

• A suitable multi-domain immunoglobulin, wherein the immunoglobulin domains are the CH3 antibody domains;

• A suitable multi-domain immunoglobulin, wherein the CH3 domains comprise the Ig3 and IgA CH3 domains;

• A suitable multi-domain immunoglobulin, wherein the first and second engineered domains are part of polypeptide chains that are joined by a disulfide bond;

• A suitable multi-domain immunoglobulin, wherein one of the first and second engineered domains contains at least two nonadjacent sequence segments from the same parent domain;

• An appropriate multi-domain immunoglobulin, each of the first and second engineered domains having at least two nonadjacent sequence segments from the same parent domain;

• A suitable multi-domain immunoglobulin, wherein each of the amino acid sequence segments comprises two or more amino acids;

• An appropriate multi-domain immunoglobulin, wherein the protein-protein interaction interface of the first engineered domain contains at least two amino acids from each parent domain;

• A suitable multi-domain immunoglobulin that contains the first bioactive domain;

• A suitable multi-domain immunoglobulin, wherein the first bioactive domain occupies the N-terminal position of the first engineered domain;

• A suitable multi-domain immunoglobulin which further comprises a second bioactive domain;

• A suitable multi-domain immunoglobulin, wherein the second bioactive domain occupies the C-terminus position of the first engineered domain;

• A suitable multi-domain immunoglobulin, wherein the first bioactive domain comprises an antibody variable domain;

- A multi-domain immunoglobulin which further comprises a second bioactive domain comprising a second antibody variable domain with different specificity;

• Multi-domain immunoglobulin containing at least the non-identical first and second engineered domains contacting the interface, with the interface of each of the first and second engineered domains containing at least two amino acid sequence segments, each derived from a different naturally occurring homologous parent domain, thanks to which the splicing specificity is different from the splicing specificity of the parent domains, where the first and second engineered domains are formed by heterodimers;

• A suitable multi-domain immunoglobulin containing at least a non-identical first and second engineered domain that are in contact on an interface, where (1) the first and second engineered domain are from two or more naturally occurring homologous parent domains, (2) the interface from the first engineered domain contains at least one segment of the amino acid sequence interacting with the amino acid sequence segment on the interface of the second engineered domain from the same parent domain, and (3) the first and second engineered domains form heterodimers;

• An engineered immunoglobulin domain that contains a protein-protein interaction interface containing amino acids from two or more parent immunoglobulin domains in such a way that this protein-protein interaction interface gives this engineered immunoglobulin domain specific splicing specificity that is different from splicing specificity parent immunoglobulin domains, wherein such engineered immunoglobulin domain is not an antibody variable domain;

• A suitable engineered immunoglobulin domain that is assembled with a partner domain with improved specificity compared to parent domains;

• An engineered constant domain of an immunoglobulin superfamily that contains a protein-protein interaction interface containing amino acids from two or more parent immunoglobulin domains in such a way that this protein-protein interaction interface gives this immunoglobulin domain constructed by engineering methods interactive properties that are different from properties interactive parent immunoglobulin domains;

• A multi-domain immunoglobulin containing an engineered domain that contains a protein-protein interaction interface, said domain being homologous to a family of naturally occurring domains, said interface containing amino acids that have been found at the appropriate sequence positions in two or more naturally occurring domains . wherein not all said amino acids are found at appropriate sequence positions in any single member of this family of naturally occurring domains;

• A suitable multi-domain immunoglobulin, wherein the interaction interface of such engineered domain gives assembly properties that are different from those of any parent domain;

• Heterodimeric engineered immunoglobulin molecule containing the first naturally occurring immunoglobulin chain from the first member of the immunoglobulin superfamily and the second naturally occurring immunoglobulin chain from the second other member of the immunoglobulin family, each of these immunoglobulin chains having a bioactive domain, which is not an antibody variable region and contains a protein-protein interface domain that contains complementary amino acid segments from another immunoglobulin superfamily chain; said first-chain protein-protein interface interacts with the second-chain protein-protein interface domain by dimerization,

Preferably homodimerization of the corresponding amino acid segments derived from the same immunoglobulin superfamily within the interaction domains;

• A suitable engineered heterodimeric immunoglobulin molecule, wherein the bioactive domain is a CH3 antibody domain, a CH2-CH3 domain, or a CH1-CH2-CH3 domain;

• Nucleic acid encoding a multi-domain immunoglobulin containing at least a non-identical first and second engineered domain, each of the first and second engineered domains having a protein-protein interaction interface containing amino acid sequence segments derived from two or more naturally occurring homologous parent domains . thanks to which the first and second domain specificity of domain engineering constructed different from the domain specificity of domain assembly are given, (1) the first and second engineered domain form heterodimers with each other in a preferential manner in relation to the creation of homodimers, and (2) the first and second constructed domain engineering methods are the variable domains of the antibody;

• A cell containing the nucleic acid disclosed herein;

• Nucleic acid encoding the multomeric immunoglobulin disclosed herein.

Brief Description of the Drawings [0025] The attached drawing is for illustrative purposes only and is not intended to limit the present invention.

[0026] Figure 1A schematically illustrates an example method for designing SEED constructs. Two related parent X and Y domains are aligned. The sequences of the two SEED subunits (XY and YX) are then generated by selecting for one SEED subunit of alternating sequence segments from these two parent sequences and selecting complementary sequence segments to generate another sequence

- SEED subunits. Engineered SEEDs using this method are called "Full" SEEDs.

[0027] Figure 1B schematically shows a second example method for designing SEED constructs, which is similar to the method of Figure 1A, except that only the amino acids forming the dimerization interfaces are selected from one of the parent sequences. Engineered SEEDs using this method are also referred to as "Surface" SEEDs.

[0028] Figure 1C schematically shows the configurations of the SEED heterodimer composed of the first SEED derivative (white oval) and the second SEED derivative (black oval) and a fusion partner such as a bioactive domain (white diamond with a tail). The SEED unit and the fusion partner may be coupled with a linker segment (not shown). In configurations with more than one fusion partner, the fusion partners may be identical or may differ from each other, although in the diagrams they are generally depicted as a white diamond with a tail. The fusion partner may be N-terminal (A) or Cterminal (B) relative to the SEED unit. There may be multiple coupled fusion partners at one end of the SEED as shown in case (C), or fusion partners may be at opposite ends of the SEED (D). One fusion partner can be located at the N terminus of the first SEED derivative while the other fusion partner can be located at the N (F) or C (G) end of the second SEED derivative. The SEED heterodimer may contain three (H) or four (I) fusion partners.

[0029] Figure 2 shows the structural alignment of the human CH3 IgG1 (SEQ ID NO: 51) and human CH3 IgA domains (SEQ ID NO: 52). Residue numbers are shown above and below the sequences given. IgG1 is numbered according to Kabat EU numbering, while IgA is numbered sequentially as in PDB 1OW0 structure. Bold letters indicate skeletal positions that were included in the alignment described in Table 2 in Example 1. Diamonds indicate residues that contact or are close to the dimerization interfaces in IgG1 and IgA homodimers.

[0030] Figure 3A shows sequence alignment and secondary sequence structure of human IgA (SEQ ID NO: 52), IgG1 (SEQ ID NO: 51) and the derivative "Surface" SEED "AG SURF" ( SEQ ID NO: 10) and "GA SURF" (SEQ ID NO: 11), while Fig. 3B shows sequence alignment and secondary structure of human IgA, IgG1 and the derivative of "Full" SEED "SEED AG" (SEQ ID NO: 3 ) and "SEED GA" (SEQ ID NO: 6). IgG1 is numbered according to Kabat EU numbering, while IgA is numbered sequentially as in PDB 1OW0 structure (natural numbers at the center of alignment). For the purposes of this drawing, the sequential numbering of the SEED sequences is interrupted at additional loop residues, which are indicated by the letters "A", "B", etc. (e.g., 18A), to illustrate the structural alignment of molecules. Thread exchange points are indicated in bold letters of the sequence. Two exchange points that do not contain any common residues are written in italics. The modeled secondary structures (arrows above and below the sequence) of the two SEEDs illustrate strand exchanges and are colored to indicate the manner in which these domains were split as indicated in Figures 6B and 6C.

n

The white segments □ are from IgA; gray segments are from IgG and black segments are common residues at exchange points. Twelve (12) residues in IgA segments are underlined. They are residues that are maintained as IgG because of their proximity to the CH3 / CH2 interface region. These residues do not participate in CH3 dimerization, but they are potentially important for interaction with CH2 and / or with the FcRn complex. Because CH2 is human IgG for both SEEDs, these residues were maintained to maintain both the native CH2 / CH3 interaction and the various well-known advantages conferred by FcRn binding.

[0031] Figure 4 is a representation of an IgG antibody molecule illustrating the symmetry of a CH3 homodimer. The vertical bar indicates the axis of double rotational symmetry.

[0032] Figure 5 is a representation of a bispecific antibody-like molecule having two different Fab domains, paired by the heterodimeric SEED analogue of the CH3 domain. The gray, hatched part represents the IgG-derived part, while the white part □ represents the IgA-derived part.

- EP 1 999 154 B1

The symmetry of the CH3 complex is broken in the Ag / GA hetermodimer, which is represented by an "X" on the vertical bar indicating the axis of double rotational symmetry.

[0033] Figure 6A-C is a schematic representation of the IgG CH3 secondary structure and two CH3-based SEEDs. Figure 6A shows the wild type secondary structure CH3.

[0034] Figure 6B illustrates the secondary structure of "SEED GA" and shows the thread exchange pattern. Gray represents the IgG sequence; white □ represents the IgA sequence; and black represents exchange points, with the widened black bar indicating residues that are conserved in both IgA and IgG.

[0035] Figure 6C shows the secondary structure of "SEED AG", which has a pattern opposite to the pattern "SEED GA".

[0036] Figure 7A-C shows ribbon diagrams representing the three-dimensional structure of the CH3 domains 'SEED GA' and 'SEED AG' and their putative heterodimeric structure showing cross-exchange point and CH3 domain interactions. In all diagrams, white or slightly gray ribbons represent the IgA sequence and structure, the dark gray area corresponds to the IgG sequence and structure, and the black sections indicate where the sequence changes from G to A or vice versa. Except for the two exchange sites at 55-56 and 101-102 (numbering according to Figure 3B), all black residues are shared by IgA and IgG, in sequence and in the basic structure.

[0037] Figure 7A shows "SEED GA", where the C-terminus begins as an IgG sequence and ends as IgA after seven exchanges. In this structure, the upper layer of β thread is in the outer layer while the back layer forms an interface with another CH3 domain.

[0038] Figure 7B shows "SEED AG" beginning with the IgA sequence. In this case, the front β strands form the interface, while the back β chains are outside the dimer.

[0039] Figure 7C shows the putative heterodimeric structure "SEED GA" and "SEED AG". The translation of the structure shown in Fig. 7A into the structure shown in Fig. 7B brings the interface surfaces together. Black residues form an approximately plane that is oriented vertically and perpendicular to the page. All residues on the left are dark gray (IgG) while all residues on the right are white (IgA). Thus, when placing white against white and gray against gray, all interfaces are well formed as a fusion of IgA and IgG interfaces. Alternative homodimers (AG / AG and GA / GA) would each have their IgA sites juxtaposed next to their IgG site (on both sides of the dividing planes) and would therefore not be favored.

[0040] Figure 8-10 schematically shows a series of protein molecules that can be produced using the SEED units described herein. For all figures of the drawing, different units are indicated as follows. In Figure 8 and Figure 9, the polypeptide chains that contain SEED GA are blackened while the polypeptide chains containing SEED AG are whitened. Within such polypeptide chains, the V antibody regions that are part of the GA SEED containing polypeptide chain are black with white stripes, while the V antibody regions that are part of the SEED AG containing polypeptide chain are white with thin black stripes. The light chain constant regions are marked with a checkerboard pattern. Antibody hinge regions are depicted as narrow ovals joined by an "SS" and a thick line to represent the disulfide bond between the hinge regions. Polypeptide linkers are represented by dashed lines.

[0041] The parts of figures Fig. 8, Fig. 9 and Fig. 10 include the following reference numbers. In some cases, to simplify the drawing figures, numerical designations are not indicated, and the identity of the different domains and regions may be inferred from the figures based on the corresponding domains and regions. "1" means GA-associated set of heavy and light chain V regions.

"2" means the AG-associated set of heavy and light chain V regions.

"3" means GA-related light chain V region.

"4" means the Fab region.

"5" means GA related heavy chain V region.

"6" means AG-associated heavy chain V region.

"7" means AG-associated light chain V region.

"8" means the light chain constant region.

"9" means the Fc region containing the SEED pair.

"10" stands for SEED pair.

"11" stands for artificial connector.

"12" means GA related single domain V or camelid V region.

"13" means AG-associated single-domain V or camelid V region.

"14" means a diabody or single chain fusion antibody that is incorporated into a polypeptide chain comprising GA SEED.

"15" means a diabody or single chain fusion antibody that is incorporated into a polypeptide chain comprising SEED AG.

"16", "17", "18" or "19" refer to any protein or peptide, such as a non-Ig domain. Such domains may contain, for example, cytokines, hormones, toxins, enzymes, antigens and extracellular domains of cell surface receptors.

"20" means the canonical homodimeric Fc region.

"21" stands for canonical homodimeric CH3 domain pair.

[0042] Figure 8 illustrates antibody-type SEED configuration types comprising units with substantially naturally occurring V regions such as Fab (Fig. 8A and Fig. 8B), single-chain Fab (Fig. 8C and Fig. 8D) and single-domain or single-domain regions V camel (Fig. 8E and Fig. 8F). Fig. 8A, Fig. 8C and Fig. 8E show molecules containing a substantially intact Fc region, including CH2 domains, as well as a hinge. Fig. 8B, Fig. 8D and Fig. 8F depict molecules without a CH2 domain in which the hinge is optionally replaced by a linker that optionally has or does not have cysteine ​​residues capable of disulfide binding.

[0043] Figure 9 illustrates antibody type SEED configuration types comprising units with artificially configured V regions, such as single chain Fv

- (Fig. 9A and Fig. 9B), antibodies (Fig. 9C and Fig. 9D) and single-chain Fv with additional units attached to the N and / or C terminus of the dipeptide chains (Fig. 9E and Fig. 9F). Figures 9A, 9C and 9E show molecules containing a substantially intact Fc region, including CH2 domains, as well as a hinge. Fig. 9B, Fig. 9D and Fig. 9F depicts molecules without a CH2 domain in which the hinge is optionally replaced by a linker that optionally has or does not have cysteine ​​residues capable of disulfide binding.

[0044] Figure 10 schematically illustrates a molecule in which the GA / AG SEED pair substantially replaces CH1-CL evaporation in an antibody. Additional units, indicated by X and Y, can be replaced at the N-terminus with SEED GA and AG. The X unit and the Y unit may be, for example, Fab, single chain Fab, single domain V camel, single chain Fv, single chain antibody such as illustrated by "14" and "15" in Figure 9C and Figure 9D. Additional units may be fused to the C-terminus of the CH3 domains indicated by "21".

[0045] Figure 11: Figure 11A shows an Fc heterodimer prepared as described in Example 5 in which the SEED AG unit has an IL-2 unit fused to the C-terminus. In this case, the CH2 and hinge units are identical. Fig. 11B shows the antibodies produced as described in Example 7, wherein the SEED AG unit has an IL-2 unit fused to the C-terminus. Each antibody domain is represented by an oval and the IL-2 unit is represented by a white square. The units CH2, CH1, hinged, VH, VL and CL are identical in this case. The hinge regions are attached by disulfide bonds represented in this figure by "SS". The light chain constant region is represented by a checkerboard pattern. The V region of the light chain is represented by a pattern of vertical stripes. VH, CH1 and CH2 are black.

[0046] In Fig. 12: Figs. 12A-C show a preferential assembly

SEED AG / GA into heterodimers as presented by the results of Fc and Fc-IL2 expression in the same cell. Figure 12A shows the possible configurations of molecules resulting from the co-expression of Fc and Fc-IL2, such that each dimer species has a different molecular weight. Figure 12B shows a non-reducing SDS gel, w

Which were loaded with the following samples: lane 1 - standard molecular weights; lane 2-4 - about 1, 2 and 4 micrograms of total protein from "Full" Fc (GA SEED) / Fc (AG SEED) -IL2 expressed from NS / 0 cells; lane 5-7 - about 1, 2 and 4 micrograms of total protein from "Surface" Fc (GA SEED) / Fc (AG SEED) -IL2 expressed from NS / 0 cells; lane 8-10 - about 1, 2 and 4 micrograms of total protein from parent IgG Fc / Fc-IL2 expressed from NS / 0 cells. Figure 12C shows a reducing gel showing the level of IgG-derived Fc and Fc-IL2 expression.

[0047] Figure 12D-E shows Western blot analysis of non-reduced samples (panel D) and reduced samples (panel E) of the Fc / Fc-IL2 proteins of Figures 12BC. Duplicate samples of 'Full' Fc (GA SEED) / Fc (AG SEED) -IL2 (lanes 1 and 4), 'Surface' Fc (GA SEED) / Fc (AG SEED) -IL2 (lanes 2 and 5) and native Fc / Fc-IL2 (lanes 3 and 6) were loaded and the nitrocellulose blot sheet was sampled using anti-human Fc IgG (lanes 1-3) and anti-human IL-2 (lanes 4-6).

Detailed description of the invention [0048] The present invention provides methods for designing heterodimeric immunoglobulin domains that preferentially heterodimerize or heteromultimerize. In particular, the invention uses the "Strand Exchange Engineered Domain" (SEED) strategy to engineer a protein-protein interaction interface that promotes heterodimerization or heteromultimerization. The invention provides multi-domain proteins containing domains engineered using this approach. In this way, the invention represents a significant advance in the field of engineering protein construction.

[0049] Hereinafter, various aspects of the present invention are set forth in detail, which should in no way be construed as limiting the scope of the invention in any way.

Each of the following sections of the description may refer to any other aspect of the invention.

[0050] As used herein, the term "multi-domain immunoglobulin" includes any immunoglobulin comprising two or more domains. These domains can be on a single polypeptide; they can also be on different polypeptides. The term "heterodimerization" refers to non-identical domains forming multimeric complexes mediated by domain interactions. The term "heteromultimeric protein" means a protein molecule comprising at least a first subunit and a second subunit, each of which contains a non-identical domain. The heteromultimer may comprise a "heterodimer" formed by the first and second subunits, or may form higher order structures (e.g., triple) in which subunit polypeptides are present in addition to the first and second subunits. Typically, each subunit contains a domain. Exemplary heteromultimer structures include heterodimers, heterotrimers, heterototramers (e.g., bispecific antibodies) and further oligomeric structures.

[0051] As used herein, the term "domain" includes any region of a polypeptide that is responsible for selective assembly with the assembly partner of interest (for example, another domain, ligand, receptor, substrate or inhibitor). Exemplary domains include an immunoglobulin superfamily constant domain, such as a CH2 or CH3 domain, a receptor binding domain, a ligand binding domain, an enzyme domain, or any polypeptide that is preferably engineered and / or selected for binding to the target. When two domains are put together, they meet at the protein-protein interaction interface. As used herein, the term "protein-protein interaction interface", "interaction interface" or "interface" includes those "contact" residues (amino acid or other non-amino acid residues such as bicarbonate, NADH, biotin, FAD or heme group) in the first domains that interact with one or more "contact" residues (amino acid or other non-amino acid groups) on the interface of the second domain. As used herein, the term "contact" residue refers to any amino acid or non-amino acid residue from one domain that interacts with another amino acid or non-amino acid residue from another domain through van der Waals forces, hydrogen bonds, hydrogen bonds mediated by water, electrolytic bridges or other electrostatic forces,

Attraction interactions between aromatic side chains, disulfide bond formation, or other forces known to those skilled in the art. Typically, the distance between the alpha carbons of two interacting contact amino acid residues in the interaction interface is not greater than 12 A. More typically, the distance between the alpha carbons of two interacting contact amino acid residues in the interaction interface is not greater than 11 A.

[0052] As used herein, the term "parent domain" refers to any existing splicing domain as described above that can be used as a parent sequence for designing an engineered domain in a thread exchange strategy. Suitable parent domains are typically associated or homologous and have specific assembly specificity. The term "homologous" typically means two domains characterized by 35%, 40%, 45%, 50%, 55%, 60%, 62%, 65%, 68%, 70%, 75%, 80%, 85%, 90 %, 95% or 99% sequence identity. If the parent domains are in ordinary solution, they may tend to homodimerize rather than heterodimerize with each other. As used herein, the term "existing splicing domains" includes wild-type sequences or naturally occurring sequences from organisms such as human, mouse, yeast, bacteria, just a few examples, as well as derived sequences that have been modified from wild-type sequences, such such as sequences that have been stabilized; sequences whose immunogenicity has been reduced; sequences with given changed, improved or reduced assembly specificity, changed enzymatic properties, changed solubility or improved expression; abbreviated; or fused to another polypeptide. The term "existing splicing domains" can also mean partly or completely synthetic sequences that are synthesized based on molecular design, by in vitro or in vivo selection methods (for example, a two yeast hybrid system, phage display technique) or combinations thereof.

[0053] As used herein, the term "engineered domain" refers to an engineered domain from at least two non-identical parent domains. A domain constructed by engineering also includes a derived daughter domain. Typically, a constructed domain

The engineering methods of the invention comprise segments of amino acid sequences derived from two or more existing homologous parent domains. Preferably, the engineered domain interface contains amino acids derived from more than one parent domain. The presence of amino acids from different parent domains gives a splicing specificity different from the splicing specificity of the parent domains. For example, the presence of amino acids from different parent domains promotes or improves heterodimerization or heteromultimerization.

[0054] A Structured Replacement Domain (SEED) is an engineered structured domain that is engineered from at least two non-identical parent domains by engineering strand exchange as described in detail hereinafter.

[0055] As used herein, the term "polypeptide" generally refers to any polypeptide or protein having more than about ten amino acids. For engineering construction of SEED, mammalian polypeptides (polypeptides that originally originated from the mammalian body) are preferably used, and even more preferably those polypeptides that are secreted directly into the medium. Examples of bacterial polypeptides include, for example, alkaline phosphatase and β-lactamase. Examples of mammalian polypeptides include molecules such as renin, growth hormone including human growth hormone; calf growth hormone; growth hormone releasing factor; parathyroid hormone, thyroid stimulating hormone, lipoproteins, α-1-antitrypsin; insulin chain A; insulin B chain; proinsulin; follitropin; calcitonin; luteinizing hormone; glucagon; coagulation factors such as factor VIIIC, factor IX, tissue factor, and von-Willebrands factor; anti-coagulant agents such as Protein C; atrial natriuretic factor, pulmonary sufractant; a plasminogen activator such as urokinase or human urine or a tissue type plasminogen activator tissue-type plasminogen activator - t-PA); bombesin; thrombin; hematopoietic growth factor; tumor necrosis factor-α and -β; enkephalinase; RANTES protein (regulated by normal activation of expressed and secreted T cells); human macrophage inflammatory protein (MIP-1-a); serum albumin such as human serum albumin; substance

Inhibiting the development of the Miler duct; relaxation chain A; relaxin B chain; prorelaxin; mouse gonadotropin-related peptide; DNase; inhibin; activin; vascular endothelial growth factor (VEGF); receptors for hormones or growth factors; integrin; protein A or D; rheumatoid factors; a neurotrophic factor such as a bone-derived neurotrophic factor ( bone-derived neurotrophic factor (BDNF), neurotrophin-3, -4, -5 or -6 (NT-3, NT-4, NT-5 or NT-6) or a nerve growth factor such as NGFbeta; platelet-derived growth factor PDGF; fibroblast growth factor such as AFGF and bFGF; epidermal growth factor (EGF); transforming growth factor ( transforming growth factor (TGF) such as TGF-α and TGF-β, including TGF-βΙ, TGFβ2, TGF-e3, TGF-e4 or TGF-e5; insulin-like growth factor I and II (IGF-I and IGFII); des (1-3) -IGF-I (brain IGF-I), insulin-like growth factor binding proteins; CD proteins such as CD-3, CD-4, CD-8 and CD-19; erythropoietin; osteoinductive factors; immunotoxins; bone morphogenetic protein ( bone morphogenetic protein - BMP); interferon such as interferon-alpha, -beta and -gamma; colony stimulating factors (CSFs), for example, MCSF, GM-CSF and G-CSF; interleukins (ILs), for example, IL-1 to IL-10; superoxide dismutase; T cell receptors; surface membrane proteins; factor accelerating degradation; transport proteins; colonization receptors; adrena addressins); regulatory proteins; immunoglobulins (antibodies); and fragments of any of the above-listed polypeptides.

[0056] As used herein, the term "first polypeptide" or "first subunit" means a polypeptide to be linked to a second polypeptide by interaction between domains constructed by engineering methods. The term "second polypeptide" or "second subunit" means any polypeptide to be linked to the first polypeptide by interaction between domains constructed by engineering methods. In addition to engineered domains, the first and / or second polypeptide may contain one or more additional bioactive domains such as, for example, an antibody variable domain, receptor binding domain, ligand binding domain or enzyme domain, or other "binding domains" such as antibody constant domains (or parts thereof) including CH3 and CH2 domains. For example, the first polypeptide may

Contain at least one engineered domain of the invention, such as an engineered immunoglobulin CH3 domain, and can form the interface of a first polypeptide. The first polypeptide may further contain another antibody heavy chain binding domain (e.g. CH1, CH2 or CH4), and additional bioactive domains such as receptor polypeptides (in particular those that form dimers with other receptor polypeptides, e.g. interleukin-8 receptor and heterodimers integrin, e.g. LFA-1 or GPIIIb / IIIa), ligand polypeptides (e.g. cytokines, nerve growth factor, nuerotrophin-3 and a brain-derived neurotrophin factor - see publication: Arakawa et al. (1994) J. Biol. Chem. 269 ​​(45): 27833-27839; and Radziejewski et al. (1993) Biochem. 32 (48): 1350) and antibody variable domain polypeptides (e.g. diabodies and BsAbs).

[0057] As used herein, the term "splicing" means the protein-protein interaction that occurs during the production of a multi-subunit protein. For example, during antibody production, heavy and light chains are synthesized from ribosomes associated with the endoplasmic reticulum. The individual chains are then folded and then assembled into mature antibodies by properly linking heavy and light chains. For example, in the case of IgG antibodies, assembly of Fab parts is initially carried out by interactions between the CH1 and CL domains, but also by interactions between the VH and VL regions. For these two heavy chains, the initial assembly reaction is the association of two CH3 domains. Such initial assembly reactions are usually followed, but not always, by the formation of a disulfide bond between assembled subunit polypeptides. As used herein, "folding" is separate from "binding"; splicing refers to protein interaction events that occur during the production of a mature protein, such as an antibody before it is secreted from the cell, while binding relates to protein interaction events that occur after secretion, such as the interaction of the antibody with the antigen or Fc receptor. In an operational sense, splicing of a therapeutic or diagnostic protein occurs during preparation of the therapeutic protein for and involves placing the product in a vial, and binding of the therapeutic or diagnostic protein relates to events that occur after

- administering a therapeutic protein to a patient or using a diagnostic protein in a diagnostic test.

[0058] As used herein, the term "binding" means the interaction of a protein with a target molecule after synthesis and assembly of this protein.

Thread Replacement by Engineering Methods [0059] The present invention uses the fact that natural protein domains mediating protein-protein interactions are often homologous or, in the case of homodimers, identical, and that such proteins and domains often only homodimerize each other but typically they do not heterodimerize with members of another family or do not heterodimerize with members of another family with an affinity equal to or greater than their affinity for homodimerization. According to the invention, such proteins can be used to design heterodimeric or heteromultimeric proteins using engineering strand exchange methods as described below. Such engineering-constructed domains are also referred to as "Strand Exchange Engineered Domains" SEEDs. Multi-domain proteins containing such engineered domains are also referred to as engineered stranded proteins.

[0060] Thread exchange engineering construction typically begins with a structural model of the dimeric domain of the parent protein. Two parent domains, each of which can homodimerize or dimerize with its own assembly partner, but which cannot heterodimerize with each other, are structurally aligned. Parent domains can dimerize face-to-face, i.e. dimer partners can be related to each other 180^about rotational symmetry. Parent domains can also dimerize in a front-to-back fashion.

[0061] Because of the rotational symmetry geometry of homodimeric proteins, there is usually a line of amino acids in the interaction surface that interacts in a homotypic manner. In other words, there are amino acids that interact with the corresponding parts in another subunit. For example, in

CH3 domain with IgG1, such amino acids include L351, P352, T366, T394, P395 and Y407. This amino acid line will be substantially parallel to the rotational symmetry axis of the dimer. When choosing parent domains, it is often useful to choose proteins that homodimerize so that the long axis of the dimerization interface is not strongly parallel to the axis of rotational symmetry. For example, SEED based members of the leucine zipper family are difficult to construct because the dimerization interface is parallel to the symmetry axis, and many amino acid interactions are homotypic. Accordingly, in some preferred embodiments, the engineered domains of the invention are not leucine zipper domains. In contrast, the CH3 family domains, on the contrary, are extremely advantageous because a significant portion of the interaction surface lies outside the symmetry line. However, one of ordinary skill in the art will recognize that the symmetry line (i.e., the line of amino acids interacting homotypically) can be an oversimplification. For example, the side chains of amino acids on the symmetry line may be directed towards the core of the hydrophobic domain.

[0062] The new dimerization interface is conceptually designed and divided into at least two regions that typically lie on either side of the homotype interaction (i.e., the symmetry line). The new domains are then designed by strand exchange, with two linear amino acid sequences of daughter daughter domains being constructed from two aligned amino acid sequences of the parent domains by selecting complementary segments from each parent sequence. As a result, in the dimerization interface regions, two daughter domain derivatives (i.e., two SEEDs) have complementary amino acid segments from parent domains. This concept is illustrated in Figs. 1A and 1B. As shown in Figure 1A, the two daughter SEED daughter sequences 1 and 2 are engineered from the two parent sequences, A and B, in a completely complementary manner. If Daughter 1 has an amino acid segment from Parent A at a given interface region, then Daughter 2 will have a corresponding amino acid segment from Parent B. This interaction interface is designed so that at least one segment of the amino acid sequence on Daughter 1 interacts with an amino acid sequence segment on Daughter 2 that is from the same parent domain. In Figure 1B, the SEED daughter domains mainly come from one domain

- parent. Nevertheless, the amino acids at one dimerization interface on one domain of a daughter SEED come from either parent or from another parent in a complementary manner.

[0063] It should be noted that Figs. 1A and 1B represent two extreme examples of the invention, and that SEEDs can be engineered by methods of the invention that include intermediate designs between Figs. 1A and Fig. 1B. For example, as described in more detail in the Examples, it is possible to construct a SEED based on parent domains from the CH3 immunoglobulin family. SEED daughters can be derived essentially in a complementary fashion from IgG and IgA, but the amino acids that interact with FcRn are derived from IgG to maintain interaction with FcRn.

[0064] Thus, SEEDs are typically constructed by engineering methods by joining two or more homologous parent domains. Parent domains are polypeptides that differ from each other by at least four amino acids. In the production of SEEDs, the sequences of the original polypeptides are aligned based on their homology, theoretical structural models, crystal structures or solution structures, or any combination thereof. There is at least one other amino acid at one or more aligned sequence positions or a different number of amino acids in at least one pair of aligned original sequences. Then the parent sequences are divided into at least two segments including at least one amino acid in each case. The SEED sequence can be composed by selecting, from the original sequences, one desired for each isolated segment. SEED will often differ from each individual parent sequence by at least two consecutive amino acids, and sometimes three, four and more consecutive amino acids. In addition to selecting sequences from original parent polypeptides, the SEED may contain any desired amino acids at any positions, such as positions outside of the desired interface, to meet other design requirements.

[0065] There are positions on the SEED sequence where the parent sequence changes from one parent to another parent. These items are called exchange points or exchange items. Exchange points or items may contain one or

- a greater number of amino acids whose identity can be shared by both parents. Typically, exchange points are selected from amino acids on or near the symmetry line, although other exchange points may also be selected. Exchange points may also contain amino acids that are not shared by parents. In this case, the sequence changes step by step from one parent to the other. In addition, exchange points may contain one or more new amino acids not belonging to any parent. In this case, typically, different parent sequences occur on one side of the new amino acids. If there are many different exchange points in the SEED sequence, the total number of parent segments can be greater than two, up to a number greater than the number of exchange points. Such parent segments can be selected from different parent domains. Thus, the present invention contemplates SEEDs that are engineered from more than two parent domains.

[0066] For convenience, each SEED is typically named according to the order of its parent sequences starting from the C-terminus of the SEED. In the examples given below, SEED AG has an IgA1 sequence segment at the N-terminus, which then changes to the IgG1 sequence segment at the first exchange point. The GA SEED has an IgG1 sequence segment at the N-terminus which then changes to the IgA1 sequence at the first exchange point.

[0067] Thus, the SEED interaction interface of the invention comprises amino acid sequence segments derived from two or more parent domains. As a result, the SEED interface has interactive properties that are different from the interactive properties of parent domains. In particular, the presence of amino acids from other parent domains gives a splicing specificity different from the splicing specificity of any of the parent domains. For example, the specificity of heterodimerization or heteromultidimerization is improved by the presence of amino acids from other parent domains on the SEED interface. As a result, the SEED pair forms heterodimers with each other preferentially over homodimer formation. And so, when the SEED pair is expressed in the expression system, the SEED heterodimers can be specifically assembled so that the heterodimeric SEED can be directly recovered from the cells

- the rearing system without the need for performing separation steps to remove homodimers.

CH3-based SEEDs [0068] Skeletal homology and differences between parenter dimerization interfaces are important for SEED production. Thus, according to one embodiment of the invention, classes of immunoglobulin proteins are a useful source for parent domains. SEEDs can be produced using parental sequences from two different immunoglobulin classes. For example, SEEDs may be constructed by engineering methods of CH3 family domains using the method of the present invention. Suitable for the design of SEED, the CH3 family domains include, but are not limited to, the CH3 domains of IgG1, IgG2, IgG3, IgG4, IgA and IgD, and the CH4 domains of IgE and IgM.

[0069] The human IgG1 and IgA CH3 domains form homodimers but do not form heterodimers with each other. Therefore, SEED pairs (e.g., SEED AG and SEED GA) can be engineered from the CH3 domains of IgG 1 and IgA so that they can heterodimerize each other, but their homodimerization ability is minimal. According to one embodiment of the invention, the CH3 domain splicing interface is divided into two regions that lie on one side of the homotype interaction line. Homotype interactions for CH3 domains with IgA and IgG1 can be determined by observing and testing the crystal structure with 1.4A sphere to determine if the two-sided chains are close enough to exclude water or not. If the surfaces are connected to each other via an interface, it means that the side chains are in close interaction. For example, in the CH3 wild-type domain of IgG1, homo-interacting amino acids include, but are not limited to L351, P352, T366, T394, P395 and Y407. For the wild-type CH3 domain with IgA1, the interacting amino acids include, but are not limited to, L352, P353, T368, W398, A399 and T414. In one exemplary SEED subunit, those amino acids with outwardly directed side chains that lie to the left of the homotype interaction line are taken from CH3 with IgG, and amino acids with outwardly directed side chains that lie to the right of the homotype interaction line are taken from CH3 with IgA. In another SEED subunit, directed amino acids

Outward side chains that lie to the left of the homotype interaction line are taken from CH3 with IgA and amino acids with outwardly directed side chains that lie to the right of the homotype interaction line are taken from CH3 with IgG1. The choice of amino acids along the homotype interaction line is based on structural considerations and performed on a case-by-case basis, although it is likely that amino acids from any parent domain may be selected from a particular SEED region.

[0070] For example, CH3-based SEED AG may have a polypeptide sequence as shown in SEQ ID NO: 1, wherein X1, X2 or X3 may be any amino acids. According to some embodiments of the invention, X1 is K or S, X2 is V or T and X3 is T or S. Preferably, X1 is S, X2 is V or T and X3 is S. The CH3-based GA SEED may have a polypeptide sequence as shown in SEQ ID NO: 2, wherein X1, X2, X3, X4, X5 or X6 may be any amino acids. According to some embodiments of the invention, X1 is L or Q, X2 is A or T, X3 is L, V, D or T, X4 is F, A, D, E, G, H, K, N, P, Q, R, S 'or T, X5 is A or T and X6 is E or D. Preferably, X1 is Q, X2 is A or T, X3 is L, V, D or T, X4 is F, A, D, E, G, H, K, N, P, Q, R, S or T, X5 is T and X6 is D. Exemplary SEED heterodimers may include one SEED subunit selected from the group consisting of SEED AG (F0) (SEQ ID NO: 3), SEED AG (f) (SEQ ID NO: 4) or SEED AG (f2) (SEQ ID NO: 5) and another SEED subunit selected from the group consisting of SEED GA (F0) (SEQ ID NO: 6), GA (f1) SEED (SEQ ID NO: 7), SEED GA (f2) (SEQ ID NO: 8) or SEED GA ( f3) (SEQ ID NO: 9). For example, the SEED heterodimer may contain the SEED AG (F0) (SEQ ID NO: 3) and SEED GA (f0) (SEQ ID NO: 6) subunits. In another example, the SEED heterodimer may comprise SEED AG (f2) (SEQ ID NO: 5) and SEED GA (f2) (SEQ ID NO: 8) subunits. In yet another embodiment of the invention, the SEED heterodimer may comprise SEED AG (s0) (SEQ ID NO: 10) and SEED GA (s0) (SEQ ID NO: 11) subunits.

Bioactive Domains [0071] The SEEDs of the present invention are extremely useful when conjugated to a fusion partner. The fusion partner (X) can be connected

Fused to the N SEED (X-SEED) end, and may also be fused to the C SEED (SEED-X) end. In addition, the fusion partner may be fused simultaneously to the N-terminus and C-terminus SEED (X-SEED-X). Two different fusion partners (X-SEED-Y) can also be fused to SEED.

[0072] Assuming that the two SEED sequences typically form heterodimers, it is possible to consider at least one, two, three or four fusion partners in the SEED heterodimer. For example, according to one embodiment of the invention, the first SEED daughter has one fusion partner and the second SEED daughter has no fusion partner, resulting in the following exemplary configurations: SEED-X heterodimerized to SEED; or X-SEED heterodimerized to SEED. In another example, the first daughter SEED has two different fusion partners (X, Y) and the second daughter SEED has two different fusion partners (W, Z) different from the fusion partners of the first daughter SEED. Possible exemplary configurations include, but are not limited to: XSEED-Y heterodimerized to W-SEED-Z; X-SEED-Y heterodimerized to Z-SEED-W; Y-SEED-X heterodimerized to W-SEED-Z; or Y-SEED-X heterodimerized to Z-SEED-W. According to the present invention, the SEED may also have two or more fusion partners (X) fused to, for example, the N-terminus (XX-SEED). Alternatively, in another embodiment of the invention, the first SEED daughter may have one fusion partner (X) and the second SEED daughter have one fusion partner (Y), resulting in the following exemplary configurations: X-SEED heterodimerized to Y-SEED; XSEED heterodimerized to SEED-Y; or SEED-X heterodimerized to SEED-Y. In yet another embodiment of the invention, the first SEED daughter has one fusion partner (X) and the second SEED daughter has two fusion partners (Z, Y). Possible exemplary configurations include without any limitation: X-SEED heterodimerized to Y-SEED-Z; X-SEED heterodimerized to Z-SEED-Y; SEED-X heterodimerized to Z-SEED-Y; or SEED-X heterodimerized to Y-SEED-Z. Example configurations are illustrated in Figure 1C.

[0073] In particular, the fusion partner may be one or more bioactive domains including any biologically active protein or biologically

Active part thereof. For example, the bioactive domain may contain a constant or variable region of the antibody, including without limitation the VL domain, VH domain, Fv, single chain Fv, diabody, Fab fragment, single chain Fab or F (ab ') 2.

[0074] According to the present invention, the fusion partners may be coupled to SEEDs directly or indirectly. For example, the fusion partner may be attached to the SEED unit via a peptide linker such as described in US Patent Application Publication Nos. 5,258,498 and US 5,482,858 by Huston et al. or US 5,856,456 and US 5,990,275 by Whitlow et al., the information contained in these documents is incorporated herein by reference. Typically, a suitable peptide linker may contain glycine and serine residues. Typically, a suitable peptide linker may also have other properties. For example, in some embodiments of the invention, the linker may further comprise a protease cleavage site, such as a matrix metalloproteinase recognition site.

[0075] Thus, the present invention provides a new method of producing multispecific antibodies based on SEED technology. A multispecific antibody is a molecule having binding specificities for at least two different antigens. While such molecules will typically bind only two antigens (i.e., BsAbs), antibodies with additional specificities such as trispecific and tetispecific antibodies are encompassed by this term if it is used herein. Examples of BsAbs include antibodies that bind to different antigens on the same cell surface or antibodies that bind to cell surface antigen and non-cell surface antigen. The non-cell surface antigen includes, but is not limited to, extracellular or intracellular antigen, soluble antigen or insoluble antigen. Multispecific antibodies can bind to different antigens simultaneously, although simultaneous binding is not required for the action of multispecific antibodies. In some applications, antigens are preferentially functionally related, such as EGFR and HER2. Extremely useful types of multispecific antibodies include, but are not limited to, anti-antibodies

EGFR / anti-HER2; anti-EGFR / anti-HER2 / anti-HER3; anti-EGFR / HER3-and; anti-EGFR / anti-HER2 / anti-IGF1R; anti-EGFR / anti-HER2 / anti-HER3 / anti-IGF1R; anti-EGFR / anti-HER3 / anti-IGF1R; anti-EGFR / anti-IGF1R; and anti-HER2 / anti-IGF1R. Other combinations of specificities include EGFR, the HER family and IGF1R, and are within the scope of the present invention.

[0076] A further example of BsAbs include cases with one arm directed against a tumor cell antigen and another arm directed against a cytotoxic trigger molecule such as anti-FcYRI / anti-CD 15, anti-p185^HER2/ FcYRIII (CD 16), anti-CD3 / anti-malignant B cells (1D10), anti-CD3 / anti-p 185^HER2, anti-CD3 / anti-p97, anti-CD3 / anti-renal tumor cell, anti-CD3 / anti-OVCAR-3, anti-CD3 / L-D1 (against colon cancer), anti-CD3 / anti-hormone analogue stimulating melanocytes, anti-EGF / anti-CD3 receptor, anti-CD3 / anti-CAMA1, anti-CD3 / antiCD19, anti-CD3 / MoV18, anti-neural cell adhesion (NCAM) / anti-CD3, anti-binding protein folacin (FBP) / anti-CD3, anti-keyland cancer antigen (AMOC-31) / anti-CD3; BsAbs with one arm that binds specifically to the tumor antigen and one arm that binds to a toxin such as anti-saporin / anti-Id-1, anti-CD22 / anti-saporin, anti-CD7 / anti-saporin, anti CD38 / anti -saporin, anti-CEA / anti-castor A chain, anti-interferon-α (IFNa) / anti-hybridoma idiotype, anti-CEA / anti-alkaloid plant from the genus VInca; BsAbs for converting enzyme activated prodrugs such as anti-CD30 phosphatase / anti-alkaline phosphatase (which catalyzes the conversion of the myytomycin phosphate prodrug to myytomycin alcohol); BsAba which can be used as fibrinolytic agents such as anti-fibrin / anti-tissue plasminogen activator (tPA), anti-fibrin / antiurokinase plasminogen activator (uPA); BsAbs for targeting immune complexes to cell surface receptors such as the anti-low density lipoprotein (LDL) / anti-Fc receptor (e.g., Fc? RI, Fc? RII or Fc? RIII); BsAbs for use in the therapy of infectious diseases such as anti-CD3 / anti-herpes simplex virus (HSV), anti-T cell receptor: CD3 / anti-influenza complex, anti-FcYR / anti-HIV; BsAbs for the detection of tumors in vitro or in vivo such as anti-CEA / anti-EOTUBE, anti-CEA / anti-DPTA, anti-CEA

HFR?

p185^HER2/ Anti-hapten; BsAbs as vaccine adjuvants; and BsAbs as diagnostic tools such as, for example, anti-rabbit IgG / anti-ferritin, horseradish antiperoxidase (HRP) / anti-hormone, anti-somatostatin / anti-substance P, anti-HRP / anti-FITC, anti-CEA / anti-e -galactosidase. Examples of trispecific antibodies include anti-CD3 / anti-CD4 / anti-CD37, anti-CD3 / anti-CD5 / anti-CD37 and anti-CD3 / anti-CD8 / anti-CD37.

[0077] According to the present invention, other bioactive domains include hormones, cytokines, chemokines, secreted enzymes, ligands, extracellular portions of transmembrane receptors or receptors. Hormones include, but are not limited to, growth hormones or glucagon-like peptides (GLP-1). Cytokines include, but are not limited to, interleukin-2 (IL-2), IL-4, IL-5, IL-6, IL7, IL-10, IL-12, IL-13, IL-14, IL-15 , IL-16, IL-18, IL-21, IL-23, IL-31; hematopoietic factors such as granulocyte macrophage colony stimulating factor (GM-CSF), G-SCF and erythropoietin; tumor necrosis factors such as TNF-α; lymphokines such as lymphotoxin; regulators of metabolic processes such as leptin; and interferons (IFN) such as IFN-α, IFN-β and IFN-γ.

[0078] Thus, engineered heteromeric immunoglobulins of the invention allow for co-localization of various bioactive domains in a biological system. This can be accomplished, for example, in the context of a multimeric protein incorporating two different antibody variable domains, wherein one antibody variable domain is fused to one engineered domain and the other antibody variable domain is fused to another engineered domain that preferentially assembles with this first domain constructed by engineering methods. The administration of such an engineered protein produces two different activities - in this case binding activities - that are to occur in the same molecule in the biological system, co-localizing these activities within the biological system. If these activities include binding to other molecules (such as variable domain / antigen interaction, ligand / receptor interaction, and the like), enzymatic activities, or a combination thereof, the present invention of the system for ordering these activities to occur at the same site allowing

For example, targeting therapeutic activity to a specific cell type or location; cross-linking of different receptors or cells; co-localization of the ari and adjuvant antigen; and the like. This can be accomplished by directly administering a heteromeric engineered protein to the biological system or by expressing the nucleic acid encoding the subunits in the biological system. Nucleic acid expression allows the engineering of additional levels of control in the system. For example, the expression of each subunit may be regulated differently, such that the complete heterodimeric protein and the resulting co-localization of activity occur only after all conditions required for expression of each subunit have occurred.

Engineered domain with reduced immunogenicity [0079] In another embodiment of the invention, the SEED sequences may be modified to reduce their potential immunogenicity. Because SEED polypeptides are hybrids between two different naturally occurring human sequences, they contain sequence segments at their junctions that are not found in natural human proteins. In the body, such sequence segments can be processed into non-specific T cell epitopes.

[0080] Methods for analyzing peptide sequences for their potential to form T cell epitopes are well known in the art. For example, ProPred (<a href="http://www.imtech.res.in/raghava/propred">http://www.imtech.res.in/raghava/propred</a>; Singh and Raghava (2001) Bioinformatics 17: 1236-1237) is a publicly available computer network-based tool that can be used to predict peptides that bind the HLADR alleles. ProPred is based on the matrix prediction algorithm described for the set of 50 HLA-DR alleles described in: Sturniolo et al., (1999) Nature Biotechnol. 17: 555-561). Using such an algorithm, different peptide sequences were discovered within the SEED AG and SEED GA polypeptide sequences that are predicted to bind to various MHC class II alleles with significant binding strength and are therefore potentially immunogenic.

[0081] For example, in one embodiment of the invention, the SEED AG and SEED GA sequences are modified to remove one or more T cell epitopes in the SEED sequence. Such modification may include the substitution, deletion or modification of one or more amino acid residues to remove a T cell epitope. Table 1 provides a list of peptide sequences that are potential T cell epitopes in SEED AG and SEED GA, and possible amino acid substitutions that are predicted to reduce or remove the T cell epitope.

<td colspan="3">SEED AG (f0)</td>

<td>Position</td><td>peptide</td><td>Amino acid substitution</td>

<td> 32</td><td>FYPKDIAVE (SEQ ID NO: 12)</td><td>K35S</td>

<td> 67</td><td>FAVTSKLTV (SEQ ID NO: 13)</td><td rowspan="2">V75T</td>

<td> 69</td><td>VTSKLTVDK (SEQ ID NO: 14)</td>

<td> 99</td><td>YTQKTISLS (SEQ ID NO: 15)</td><td>T103S</td>

<td></td><td colspan="2">SEED GA (f0)</td>

<td>Position</td><td>peptide</td><td>Amino acid substitution</td>

<td> 18</td><td>LALNELVTL (SEQ ID NO: 16)</td><td></td>

<td> 20</td><td>LNELVTLTC (SEQ ID NO: 17)</td><td>L23Q</td>

<td> 23</td><td>LVTLTCLVK (SEQ ID NO: 18)</td><td></td>

<td> 54</td><td>YLTWAPVLD (SEQ ID NO: 19)</td><td>A58T</td>

<td> 55</td><td>LTWAPVLDS (SEQ ID NO: 20)</td><td>L61V, D, T</td>

<td> 61</td><td>LDSDGSFFL (SEQ ID NO: 21)</td><td>L61V, D, T.</td>

<td> 67</td><td>FFLYSILRV (SEQ ID NO: 22)</td><td>F67A, D, E, G, H, K, N, P, Q, R, S, T.</td>

<td> 68</td><td>FLYSILRVA (SEQ ID NO: 23)</td><td>A76T</td>

- EP 1 999 154 B1

<td> 69</td><td>LYSILRVAA (SEQ ID NO: 24)</td><td rowspan="3">E78D</td>

<td> 70</td><td>YSILRVAAE (SEQ ID NO: 25)</td>

<td> 72</td><td>ILRVAAEDW (SEQ ID NO: 26)</td>

Table 1 shows the peptides in SEED AG (f0) or SEED GA (f0) for which they are predicted to bind to HLA-DR alleles and that they are potential T cell epitopes, and amino acid substitutions at specific residues (indicated in bold) inside peptides for which they are predicted to reduce binding to HLA-DR alleles. The "Pos" designation indicates the position of the peptide within the sequence. Amino acid numbering is sequential and implemented relative to the first amino acid of the SEED molecule.

[0082] Primary "full" SEED AG polypeptides (SEED AG (f0) (SEQ ID NO: 3)) and

SEED GA (SEED GA (f0) (SEQ ID NO: 6)), and some exemplary variant polypeptides, including SEED AG (f1) (SEQ ID NO: 4), SEED AG (f2) (SEQ ID NO: 5), SEED GA (f1) (SEQ ID NO: 7), GA (f2) SEED (SEQ ID NO: 8), and SEED GA (f3) (SEQ ID NO: 9) are shown in the following equations.

SEED AG alignment (does not indicate residual identity)

GQPFRPEVHLLPPSREEMTKNQVSLTCLARGFYPKDIAVEWESNGQPENNYKTTPSRQEP AG &lt;£ 0&gt; SEED

............... .................................. S .......... AG (fl) SEED

............... .................................. S ........ · .. AG (£ 2) SEED

SQGTTT FAVTSKLTVDKS RWQQGNVFSCSVMHEALHNHYTQKTIS L

.......................................... S ...

........................... .............. T S ...

AG (fO) SEED AG (fi) SEED AG (£ 2) SEED

GA SEED alignment (does not indicate residual identity)

GQPREPQVYTLPPPSEELALNELVTLTCLVKGFYPSDIAVEWLQGSQELPREKYLTWAPV GA &lt;£ 0) SEED

Q ........................... ...................... ....... T. . GA (fl) SEED

GA (£ 2) SEED

Q ........................... ...................... ....... T. . GA (f3) SEED

LDSDGSFFLYSILRVAAEDWKKGDTFSCSVMHEALHNHYTQKSLDR

V TD ........................... ..............

D ..... ........ TD H ...........................

T ..... ........ TD D ...........................

GA &lt;fO&gt; SEED GA (fl) SEED GA &lt;f2&gt; SEED GA &lt;f3) SEED [0083] Further embodiments of the invention are detailed in the Examples below.

- EP 1 999 154 B1

EXAMPLES

EXAMPLE 1: Identifying homologous structures for becoming parents for

SEED.

[0084] In this set of examples, the goal is to produce two distinct homologous CH3 SEEDs that will form dimers that favor heterodimer formation over the formation of two possible homodimers, thereby resulting in an advantage of homologous CH3 heterodimers. The first task is to identify one or more CH3 domains that can give this result when used as parents of a SEED pair. The CH3 homodimer creates a dimerization interface between harmonica □. It is important to find two CH3 domains that have significant differences in this interface in order to generate an effective SEED pair that will preferentially heterodimerize.

[0085] Ig3 CH3 domains are structurally highly preserved in the animal kingdom, containing the classic fold of the □ -sandwicz type immunoglobulin domain. While there are significant differences between species in the amino acid identities found on the outer surface, the surface of the dimerization interface that is hidden after dimerization is largely preserved.

[0086] Each other immunoglobulin class has its own Fc, and in particular its own CH3 IgG sequence equivalent and structure. Examination of the CH3 domain for the crystal structure of the human IgA1 Fc part (PDB number 1OW0, resolution 3.1A) revealed that the total folding was homologous to human CH3 IgG. Square root deviation ( root mean square deviation (RMSD) skeletal alignment of single CH3 domains with Fc IgA 1OW0 and with Fc IgG 1L6X, excluding bends where the alignment had different lengths, was about 0.99 A (see Table 2). Nevertheless, the interface of the CH3 homodimer with IgA1 is clearly different from the interface with IgG1. Thus, each of the SEEDs made from CH3 human IgA1 and CH3 human IgG1 contains some interface with IgA1, and some with IgG1, and they are not designed not to dimerize with each other or with any other parent CH3, but to preferentially dimerize with another complementary SEED.

- 39 EP 1 999 154 B1

<td colspan="2">Table 2. Structural alignment of CH3 domains with IgG and IgA *</td>

<td>Human IgG</td><td>Human IgA</td>

<td>Q342 - M358</td><td>N343 - L359</td>

<td>N361 - P387</td><td>E363 - E389</td>

<td>N390 - S400</td><td>K394 - P404</td>

<td>G402 - L443</td><td>T409 - R450</td>

* Parts from the IgG and IgA sequences were used to cover structures and determined skeletal RMSD. The InsightII program (Accelrys, San Diego,

CA) coats the skeletal atoms of the residues contained in the structurally homologous sequences listed above. The RMSD between these two skeletons (within the ranges of the table) was 0.99 A.

[0087] For example, the CH3 domain from human IgA1 and the CH3 domain from human IgG1 were used as parent polypeptides. For structural alignment and modeling, the PDB IgG1 1DN2 (2.7A resolution) and 1L6X (CH3 sequence highly homologous to 1DN2, with two smaller differences, resolution 1.65 A) and PDB IgA1 1OW0 (3.1A resolution) were used. Figure 2 shows the structural alignment of these two sequences. The CH3 IgG1 domain is numbered according to the Kabat EU Index (Kabat et al., (1991) Sequences of Proteins of Immunological Interest, 5th Edition, NIH Publication 91-3242), while the CH3 IgA domain is sequentially numbered as in PDB 1OW0 structure. Bold letters indicate skeletal positions that were included in the alignment described in Table 2, which were then used to design cross-connect points when designing SEEDs.

EXAMPLE 2: Selection of exchange points [0088] After determining the structural alignment and identifying interface residues, it is possible to determine exchange points to produce the SEED. The CH3 homodimer is 180^about rotational symmetry around the axis that runs between these domains approximately perpendicular to the beta thread (Figure 4). Each domain has an N end and a C end on opposite sides of the symmetry axis. Therefore, the CH3 domains dimerize in an agreed manner, but only in a line down from the center of the interface along the symmetry axis, residues on one side are in contact with the same residues at another partner. Residues on the other side of this line are in contact with the reverse domain of the partner

Method: for example, residues on the N-terminal side of the first domain contact residues in the second domain that are on the C-terminal side, and vice versa.

[0089] According to one embodiment, the CH3 based SEED is designed to break symmetry, creating two different sides. For example, the exchange of strands will form one side of the dimer more like IgA1 and the other side of the dimer more like IgG. This approach will produce two different CH3-based SEEDs that are approximately complementary in their use for IgG and IgA derived amino acids. As shown in the figure 3A and 3B, linear polypeptide sequences run back and forth between the IgG and IgA sequences to produce one physical side of the IgA-like dimeric structure and the other side similar to the IgG. Thus, each terminal SEED sequence contains a wide variety of exchange points, in which each linear sequence changes from IgA to IgG or from IgG to IgA (see Figures 3A and 3B).

[0090] Generally, there are a wide variety of potential exchange points in a polypeptide sequence that can be selected to effect a change between IgA and IgG1 sequences. The important point is that the final structure should have good structural properties (e.g. stability, folding, expression, homology to the original). This can be achieved through control, simple modeling, extensive calculations, testing and errors, selection or other means. In the specific embodiment described in the present application, sequence homology between the CH3 domains of IgA and IgG1 was used to decide on exchange points. Alignment of CH3 crystal structures with IgA and IgG1 revealed approximately parallel lines of amino acids along an approximate plane tilted through the center of the domain. Residues in this plane were identical in both CH3 classes in all cases except for the two strands in the structural IgG1 / IgA alignment. In addition, structural alignment essentially showed the side chains of these amino acids in the same rotamer orientations, in particular in the hydrophobic core. Therefore, the hypothesis was formulated that these residues could be used as exchange points, and the residues on one side or the other could be changed without breaking the entire structure. Figures 3A and 3B show sequence alignment with exchange points marked

- in bold letters. Figures 5 and 6A-C show the molecular structure illustrating the three-dimensional locations of the exchange points.

[0091] In two cases where the residues in the junction region are not the same, the choice of exchange points was based on structural considerations. In one case, Pro395 and Pro396 in IgG1 structurally correspond to Ala399 and Ser400 in IgA1. Separation was made between these two residues. Another location is near the C-terminus, Leu441 and Ser442 in IgG1 structurally corresponds to Ile448 and Asp449 in IgA1. Again, separation between these residues.

[0092] Protein-protein interactions are mediated by the complementarity of two interactive surfaces. The dominant factor for interaction is the composition and shape of these surfaces. Because the underlying skeletal and hydrophobic structures of the IgA and IgG1 CH3 domains are similar, it was considered according to the principles of the present invention that only this surface would need to be changed while the rest of the domain could contain IgG sequences. In this case, the exchange points would be designed on threads that form the interface and would be very close together, allowing only the dimerization critical residues to be exchanged. Thus, as an alternative, it is possible that the rest of such a structure could help stabilize the structural domain, and SEED CH3 with a single exchange point in each of the seven strands could have some advantages.

[0093] Therefore, two types of SEED may be designed and designated as "Full" in relation to SEEDs, most or all of which residues in this domain were involved in thread exchange (according to Figure 1A) or "Surface" in relation to SEED in which only the changed residues are at the CH3 dimerization interface (according to Figure 1B).

[0094] Based on this Example, one of skill in the art will appreciate that many different strategies can be used to generate SEEDs based on immunoglobulin superfamily constant domains.

EXAMPLE 3: Designing the "Full" SEED AG and GA sequences [0095] As an example, the easiest way to produce a "Full" SEED would be to use pure IgA sequence in the first place of the exchange point, and pure

- the IgG1 sequence in second place of the exchange point. If the exchange point is appropriately selected, the result will be SEED, which should fold correctly and should have an IgA1-like dimerizing surface in one place (e.g., half) the domain, and an IgG-like dimerizing surface in another place. The "mirror image" SEED may be similarly produced, in which image the first site is composed of the IgG1 sequence and the second site is composed of the IgA sequence. In the event that both SEEDs are expressed together, they will preferentially form heterodimers, because only in the heterodimer each surface will contact a surface on a different domain that matches its class: this means that the first half of the first SEED, which is similar to IgA1, will contact the second half of the second SEED, which is also similar to IgA1, while the second half of the first SEED, which is similar to IgG1, will contact the first half a second SEED, which is also similar to IgG1. Since both sides of the contact surface are highly complementary, this connection should be strong. On the other hand, when any SEED attempts to form a homodimer, each half of the dimerization surface will contact the surface of the SEED partner that is from a different class; that is, the first half of one SEED that is similar to IgA will contact the other half of the partner domain that is similar to IgG; and the second half of the first SEED, which is similar to IgG1, will contact the first half of the partner domain that is similar to IgA. Because these surfaces are not highly complementary, their affinity will be reduced, which will result in thermodynamic favoring the formation of less homodimers and more heterodimers.

[0096] In this example, CH3 is the only part of the Fc or antibody that has been changed. The Fc or immunoglobulin residue is derived from human IgG1. Changing the amino acid sequence in which CH3 contacts or interacts with CH2 could potentially create problems with the interaction between CHED SEED and IgG1 CH2 domains. In addition, such an interface contains a binding site for FcRn that confers important properties to Fc that it is desirable to maintain. Therefore, structural information (Martin et al. (2001) Molec. Cell 7: 867) was used to identify CH3 residues involved in the interaction between CH3 and CH2, and between Fc and FcRn. Human IgG1 sequences were used for these residues in all SEEDs.

- EP 1 999 154 B1

Molecular modeling has also been used to support the selection of adjacent residues to avoid changing the surface structure of the FcRn interaction. The part of CH3 that interacts with CH2 and FcRn is not part of the dimerization interface, so these changes were not expected to interfere with the formation of heterodimers.

[0097] Figure 3B has the sequences of "Full" SEED aligned with the sequences of IgG1 and IgA in structural alignment. Residues at exchange points are shown in bold. Residues that remained unchanged due to their importance in maintaining interaction with CH2 and / or FcRn are underlined.

EXAMPLE 4: Construction of heterodimeric Fc and CH3-based SEED containing antibody molecules [0098] The following general approach was used to generate HuFc and HuFc-IL2 constructs, as well as antibody and antibody-IL2 constructs containing CH3 SEED domains at the site of IgG1 CH3 domains. The Ig3 CH3 domain is almost always contained approximately in the 0.4kb Ngo MIV / Sma I genomic DNA fragment that is present in pdCs or pdHL expression plasmids that express the IgG1 heavy chain constant region. Examples of expression plasmids are, for example, pdCs-huFc-IL2 (see, for example, Lo et al., Protein Engineering [1998] 11: 495) or pdHL7-KS-IL2 (see, for example, US Patent Application Publication No. 6,696,517). The Ngo MIV site is inside the intron sequence directly at the 5 'exon coding for IgG1 CH3, and the Sma I site is in the Ser444Pro445Gly446 coding sequence near the C-terminus of IgG1 (Kabat EU Index). An exemplary DNA sequence of mature human IgG1 Fc expressed from the pdCs vector is SEQ ID NO: 27. Replacement of the parent Ngo MIV / Sma I fragment with the Ngo MIV / Sma I fragment encoding the CH3 SEED of the invention generates expression of the polypeptide containing the constant region of CH3 SEED.

SEQ ID NO: 27 [0099] DNA sequence in pdCs coding for mature human IgG1 Fc including terminal lysine residue.

- 44. EP 1 999 154 B1

GAGCCCAAATCTTCTGACAAAACTCACACATGCCCACCGTGCCCAGGTAAGCCAGCCCAGGCCTC

GCCCTCCAGCTCAAGGCGGGACAGGTGCCCTAGAGTAGCCTGCATCCAGGGACAGGCCCCAGCCG

GGTGCTGACACGTCCACCTCCATCTCTTCCTCAGCACCTGAACTCCTGGGGGGACCGTCAGTCTT

CCTCTTCCCCCCAAAACCCAAGGACACCCTCATGATCTCCCGGACCCCTGAGGTCACATGCGTGG

TGGTGGACGTGAGCCACGAAGACCCTGAGGTCAAGTTCAACTGGTACGTGGACGGCGTGGAGGTG

CATAATGCCAAGACAAAGCCGCGGGAGGAGCAGTACAACAGCACGTACCGTGTGGTCAGCGTCCT

CACCGTCCTGCACCAGGACTGGCTGAATGGCAAGGAGTACAAGTGCAAGGTCTCCAACAAAGCCC

TCCCAGCCCCCATCGAGAAAACCATCTCCAAAGCCAAAGGTGGGACCCGTGGGGTGCGAGGGCCA

CATGGACAGAGGCCGGCTCGGCCCACCCTCTGCCCTGAGAGTGACCGCTGTACCAACCTCTGTCC

CTACAGGGCAGCCCCGAGAACCACAGGTGTACACCCTGCCCCCATCACGGGAGGAGATGACCAAG

AACCAGGTCAGCCTGACCTGCCTGGTCAAAGGCTTCTATCCCAGCGACATCGCCGTGGAGTGGGA

GAGCAATGGGCAGCCGGAGAACAACTACAAGACCACGCCTCCCGTGCTGGACTCCGACGGCTCCT

TCTTCCTCTATAGCAAGCTCACCGTGGACAAGAGCAGGTGGCAGCAGGGGAACGTCTTCTCATGC

TCCGTGATGCATGAGGCTCTGCACAACCACTACACGCAGAAGAGCCTCTCCCTGTCCCCGGGTAA

ATGA [0100] Standard techniques were used to obtain DNA sequences encoding CH3 SEEDs according to the invention. For example, DNA molecules with the following sequences designated SEQ ID NO: 28, SEQ ID NO: 29, SEQ ID NO: 30,

SEQ ID NO: 31, SEQ ID NO: 32 and SEQ ID NO: 53 were re-synthesized and propagated in a pUC derived plasmid (Blue Heron Biotechnology, Bothell, WA).

SEQ ID NO: 28 [0101] Ngo MIV / Sma I DNA fragment, containing the SEED 10 AG (f0) coding sequence (the underlined fragment corresponds to Figure 3B, "Full SEED AG"):

gccggctcggcccaccctctgccctgagagtgaccgctgtaccaacctctgtccctacaGGGCAG

CCCTTCCGGCCAGAGGTCCACCTGCTGCCCCCATCACGGGAGGAGATGACCAAGAACCAGGTCAG

CCTGACCTGCCTGGCACGCGGCTTCTATCCCAAGGACATCGCCGTGGAGTGGGAGAGCAATGGGC

AGCCGGAGAACAACTACAAGACCACGCCTTCCCGGCAGGAGCCCAGCCAGGGCACCACCACCTTC

GCTGTGACCTCGAAGCTCACCGTGGACAAGAGCAGATGGCAGCAGGGGAACGTCTTCTCATGCTC

CGTGATGCATGAGGCTCTGCACAACCACTACACGCAGAAGACCATCTCCCTGtccccggg

SEQ ID NO: 29

[0102] Ngo MIV / Sma I DNA fragment, comprising the SEED AG (s0) coding sequence (the underlined fragment corresponds to Figure 3A, "Surface SEED AG"):

gccggctcggcccaccctctgccctgagagtgaccgctgtaccaacctctgtccctacaGGGCAG

CCCTTCGAACCAGAGGTCCACACCCTGCCCCCATCACGGGAGGAGATGACCAAGAACCAGGTCAG

CCTGACCTGCCTGGTCCGCGGCTTCTATCCCAGCGACATCGCCGTGGAGTGGGAGAGCAATGGGC

AGCCGGAGAACAACTACAAGACCACGCCTTCCCGGCTGGAGCCCAGCCAGGGCACCACCACCTTC

GCTGTGACCTCGAAGCTCACCGTGGACAAGAGCAGATGGCAGCAGGGGAACGTCTTCTCATGCTC

CGTGATGCATGAGGCTCTGCACAACCACTACACGCAGAAGAGCCTCTCCCTGtCCcęggg

SEQ ID NO: 30 [0103] Ngo MIV / Sma I DNA fragment containing the SEED GA (f0) coding sequence (the underlined fragment corresponds to Figure 3B, "Full GA SEED"):

gccggctcggcccaccctctgccctgagagtgaccgctgtaccaacctctgtccctacaGGGCAG

CCCCGAGAACCACAGGTGTACACCCTGCCCCCACCGTCGGAGGAGCTGGCCCTGAACGAGCTGGT

GACGCTGACCTGCCTGGTCAAAGGCTTCTATCCCAGCGACATCGCCGTGGAGTGGCTGCAGGGGT

CCCAGGAGCTGCCCCGCGAGAAGTACCTGACTTGGGCACCCGTGCTGGACTCCGACGGCTCCTTC

TTCCTCTATAGTATACTGCGCGTGGCAGCCGAGGACTGGAAGAAGGGGGACACCTTCTCATGCTC CGTGATGCATGAGGCTCTGCACAACCACTACACGCAGAAGAGCCTCGACCGCtccccgg g

SEQ ID NO: 31 [0104] Ngo MIV / Sma I DNA fragment containing the SEED coding sequence

GA (s0) (the underlined fragment corresponds to Figure 3A, "Surface SEED GA"):

gccggctcggcccaccctctgccctgagagtgaccgctgtaccaacctctgtccctacaGGGCAG

CCCCGAGAACCACAGGTGTACACCCTGCCCCCACCGTCGGAGGAGCTGGCCCTGAACAACCAGGT

GACGCTGACCTGCCTGGTCAAAGGCTTCTATCCCAGCGACATCGCCGTGGAGTGGGAGAGCAATG

GGCAGCCGGAGCCCCGCGAGAAGTACCTGACTTGGGCACCCGTGCTGGACTCCGACGGCTCCTTC

TTCCTCTATTCGATACTGCGCGTGGACGCAAGCAGGTGGCAGCAGGGGAACGTCTTCTGATGCTC

CGTGATGCATGAGGCTCTGCACAACCACTACACGCAGAAGAGCCTCTCCCTGtCCCęggg

SEQ ID NO: 32

[0105] Ngo MIV / Sma I DNA fragment, comprising the SEED GA (f1) coding sequence (underlined):

gccggctcggcccaccctctgccctgagagtgaccgctgtaccaacctctgtccctaca-GGGCAG

CCCCGAGAACCACAGGTGTACACCCTGCCCCCACCGTCGGAGGAGCTGGCCCTGAACGAGCaGGT

GACGCTGACCTGCCTGGTCAAAGGCTTCTATCCCAGCGACATCGCCGTGGAGTGGCTGCAGGGGT

CCCAGGAGCTGCCCCGCGAGAAGTACCTGACTTGGaCcCCCGTGgTGGACTCCGACGGCTCCTTC

TTCCTCTATAGTATACTGCGCGTGaCAGCCGAtGACTGGAAGAAGGGGGACACCTTCTCATGCTC

CGTGATGCATGAGGCTCTGCACAACCACTACACGCAGAAGAGCCTCGACCGCtccccggg

SEQ ID NO: 53 [0106] Ngo MIV / Sma I DNA fragment containing the SEED coding sequence

GA (f2) (underlined):

gccggctcggcccaccctctgccctgagagtgaccgctgtaccaacctctgtccctacaGGGCAG

CCCCGAGAACCACAGGTGTACACCCTGCCCCCACCGTCGGAGGAGCTGGCCCTGAACGAGCaGGT

GACGCTGACCTGCCTGGTCAAAGGCTTGTATCCGAGCGAGATCGGCGTGGAGTGGCTGCAGGGGT

CCCAGGAGCTGCCCCGCGAGAAGTACGTGACTTGGgCaCCCGTGgacGACTCCGACGGCTCCcaC

TTCCTCTATAGTATACTGCGCGTGaCAGCCGAtGACTGGAAGAAGGGGGACACCTTCTCATGCTC

CGTGATGCATGAGGCTCTGCACAACCACTACACGCAGAAGAGCCCCACCGCtC-CCCggg [0107] In addition, a SEED GA (f3) containing polypeptide can be encoded by the following DNA sequence:

SEQ ID NO: 54

Ngo MIV / Sma I DNA fragment containing the SEED GA (f3) coding sequence (underlined):

gccggctcggcccaccctctgccctgagagtgaccgctgtaccaacctctgtccctacaGGGCAG

CCCCGAGAACCACAGGTGTACACCCTGCCGCCACCGTCGGAGGAGCTGGCCCTGAACGAGCaGGT

GACGCTGACCTGCCTGGTCAAAGGCTTGTATCCCAGCGACATCGCCGTGGAGTGGCTGCAGGGGT

CCCAGGAGCTGCCCCGCGAGAAGTACCTGACTTGGaCcCCCGTGaccGACTCCGACGGCTCCgac

TTCCTCTATAGTATACTGCGCGTGaCAGCCGAtGACTGGAAGAAGGGGGACACCTTCTCATGCTC

CGTGATGCATGAGGCTCTGCACAACCACTACACGCAGAAGAGCCTCGACCGCtccccggg

These synthetic sequences have been further extended at their 3 'end by a 15 DNA random section of approximately 50 base pairs so that

- allow easy separation of the excised Ngo MIV / Sma I insertion fragment and similar plasmid vector fragment sizes during fragment purification. Gel purification of Ngo MIV / Sma I fragments were then ligated into a similarly treated pdCs vector containing either an Fc unit or Fc-IL2 unit, or alternatively to a similarly treated vector containing a DI-KS unit or a DI-KS-IL2 unit. Thus, for example, pdCs-HuFc (AG (f0)) - IL2 was obtained, containing the Ngo MIV / Sma I fragment for SEED AG (f0) (SEQ ID NO: 28), and pdCs-HuFc (GA (f0)), containing the Ngo MIV / Sma I fragment for SEED GA (f0) (SEQ ID NO: 30). pdCsHuFc (AG (f0)) - IL2 and pdCs-HuFc (GA (f0)) encode the Fc (AG (f0) SEED) -IL-2 polypeptide chain and the Fc (GA (f0) SEED) polypeptide chain, respectively. Exemplary Fc (AG (f0) SEED) -IL-2 and Fc (GA (f0) SEED) sequences are shown as SEQ ID NO: 33 and SEQ ID NO: 34, respectively. The resulting heterodimeric protein diagram is shown in Figure 11A. To obtain simultaneous expression of both polypeptide chains from the host cell, transcription units for such Fc polypeptides were combined on a single expression vector as described in Example 5.

SEQ ID NO: 33

Fc (AG (f0) SEED) -IL2 polypeptide sequence:

EPKSSDKTHTCPPCPAPELLGGPSVFLFPPKPKDTLMISRTPEVTCVWDVSHEDPEVKFNWYVD gvevhnaktkpreeqynstyrwsvltvlhqdwlngkeykckvsnkalpapiektiskakgqpfr

PEVHLLPPSREEMTKNQVSLTCLARGFYPKDIAVEWESNGQPENNYKTTPSRQEPSQGTTTFAVT

SKLTVDKSRWQQGNVFSCSVMHEALHNHYTQKTISLSPGKAPTSSSTKKTQLQLEHLLLDLQMIL

NGINNYKNPKLTRMLTFKFYMPKKATELKHLQCLEEELKPLEEVLNLAQSKNFHLRPRDLISNIN

VIVLELKGSETTFMCEYADETATIVEFLNRWITFCQSIISTLT

SEQ ID NO: 34

Fc (GA (f0) SEED) polypeptide sequence:

EPKSSDKTHTCPPCPAPELLGGPSVFLFPPKPKDTLMISRTPEVTCWVDVSHEDPEVKFNWYVD

GVEVHNAKTKPREEQYNSTYRWSVLTVLHQDWLNGKEYKCKVSNKALPAPIEKTISKAKGQPRE

PQVYTLPPPSEELALNELVTLTCLVKGFYPSDIAVEWŁQGSQELPREKYLTWAPVLDSDGSFFLY

SILRVAAEDWKKGDTFSCSVMHEALHNHYTQKSLDR5PGK

- 48 EP 1 999 154 B1

Similarly, pdHL-DI-KS (AG (f0)) - IL2 was obtained, containing the Ngo MIV / Sma I fragment for SEED AG (f0) (SEQ ID NO: 28), and pdHL-DI-KS (GA (f0)) , containing the Ngo MIV / Sma I fragment for SEED GA (f0) (SEQ ID NO: 30). pdHL-DI-KS (AG (fD)) - IL2 and pdHL-DI-KS (GA (f0)) encode the DI-KS (AG (f0) SEED) -IL-2 heavy chain, respectively (SEQ ID NO: 35) and DI-KS (GA (f0) SEED) heavy chain (SEQ ID NO: 36). Both of these expression vectors also encode the DI-KS light chain (SEQ ID NO: 37).

SEQ ID NO: 35

DI-KS (AG (f0) SEED) -IL2 heavy chain polypeptide sequence:

QIQLVQSGPELKKPGSSVKISCKASGYTFTNYGMNWVRQAPGKGLKWMGWINTYTGEPTYADDFK

GRFTITAETSTSTLYLQLNNLRSEDTATYFCVRFISKGDYWGQGTTVTVSSASTKGPSVFPLAPS

SKSTSGGTAALGCLVKDYFPEPVTVSWNSGALTSGVHTFPAVLQSSGLYSLSSWTVPSSSLGTQ

TYICNVNHKPSNTKVDKRVEPKSCDKTHTCPPCPAPELLGGPSVFLFPPKPKDTLMISRTPEVTC

WVDVSHEDPEVKFNWYVDGVEVHNAKTKPREEQYNSTYRWSVLTVŁHQDWLNGKEYKCKVSNK

ALPAPIEKTISKAKGQPFRPEVHLLPPSREEMTKNQVSLTCLARGFYPKDIAVEWESNGQPENNY

KTTPSRQEPSQGTTTFAVTSKLTVDKSRWQQGNVFSCSVMHEALHNHYTQKTISLSPGAAPTSSS

TKKTQLQLEHLLLDLQMILNGINNYKNPKLTRMLTFKFYMPKKATELKHLQCLEEELKPLEEVLN

LAQSKNFHLRPRDLISNINVIVLELKGSETTFMCEYADETATIVEFLNRWITFCQSIISTLT

SEQ ID NO: 36

DI-KS (GA (f0) SEED) heavy chain polypeptide sequence:

QIQLVQSGPELKKPGSSVKISCKASGYTFTNYGMNWVRQAPGKGLKWMGWINTYTGEPTYAODFK GRFTITAETSTSTLYLQLNNLRSEDTATYFCVRFISKGDYWGQGTTVTVSSASTKGPSVFPLAPS SKSTSGGTAALGCLVKDYFPEPVTVSWNSGALTSGVHTFPAVLQSSGLYSLSSWTVPSSSLGTQ TYICNVNHKPSNTKVDKRVEPKSCDKTHTCPPCPAPELLGGPSVFLFPPKPKDTLMISRTPEVTC VWDVSHEDPEVKFNWYVDGVEVHNAKTKPREEQYNSTYRVVSVLTVLHQDWLNGKEYKCKVSNK ALPAPIEKTISKAKGQPREPQVYTLPPPSEELALNELVTLTCLVKGFYPSDIAVEWLQGSQELPR EKYLTWAPVLDS DGS FFLYSILRVAAEDWKKGDTFSCSVMHEALHNHYTQKSLDRS PGK

SEQ ID NO: 37

DI-KS light chain polypeptide sequence:

- 49 EP 1 999 154 B1

QIVLTQSPASLAVSPGQRATITCSASSSVSYILWYQQKPGQPPKPWIFDTSNLASGFPSRFSGSG

SGTSYTLTINSLEAEDAATYYCHQRSGYPYTFGGGTKVEIKRTVAAPSVFIFPPSDEQLKSGTAS

WCLLNNFYP REAKVQWKVDNALQSGNSQESVTEQDSKDSTYSLSSTLTLSKADYEKHKVYACEV

THQGLSSPVTKSFNRGEC [0108] To obtain a single expression vector expressing both the heavy chain transcription units DI-KS (AG (f0) SEED) -IL-2 and DI-KS (GA (f0) SEED) as well as the usual light chain transcription units, construct essentially as follows: containing approximately 3.9 thousand base pairs the Sal I / Mfe I fragment containing the KS (AG (f0) SEED) -IL-2 coding sequence was excised from the pdHL-10 expression construct (pdHL-10 is a pdHL late generation expression vector containing a single Sal I site outside the transcription unit) and ligated into Sal I / Bam HI digested pBS plasmid, along with a fragment of the Bam HI / Mfe I duplex linker. This fragment of the duplex connector consists of Oligo11 (SEQ ID NO: 38) and Oligo12 (SEQ ID NO: 39) and contains the internal site of Sal I.

Oligo11 (SEQ ID NO: 38)

AATTGCCGGGTCGACATACG

Oligo12 (SEQ ID NO: 39)

GATCCGTATGTCGACCCGGC [0109] This 3.9 thousand base pair fragment was then excised from pBS as a Sal I fragment and inserted into the unique Sal I site of the pdHL-10 expression construct already containing transcription units encoding DI-KS heavy chain (GA (f0) SEED) and the DI-KS light chain.

EXAMPLE 5: Test for determining heterodimeric Fc molecules containing

CH3 based SEED [0110] Examples described above include CH3 dimerization, which is an important step in the nucleation of the formation of Fc dimers and immunoglobulin heavy chains. In theory, if two distinct Fc units (for example, designated A and B) containing CH3 domains are expressed simultaneously in the cell, they could pair and form dimer Fc molecules in

- the following configurations: A: A, A: B and B: B. If the CH3 domains and the hinge domains are identical, then the A: A, A: B and B: B configurations are expected to be in a 1: 2: 1 ratio if A and B are expressed in equal amounts. The relative amounts, kinetics and thermodynamics of AA, AB and BB interactions are important control factors for the observed ratio of these three end species, which would be expression levels. Basically, when protein A and protein B are expressed in relative amounts [A] and [B], where [A] + [B] = 1, and homodimers and heterodimers are produced in relative concentrations [AA], [AB] and [ BB], if there is an unencumbered relationship, these dimeric species will occur in the ratio [A] 2: 2 * [A] * [B]: [B] respectively². If the relative concentration is [AB]&gt; 2 * [A] * [B], then heterodimerization is favored, while if the relative concentration is [AB] &lt;2 * [A] * [B], then homodimerization is favored . For preferred SEED pairs, the ratio [AB] / 2 * [A] * [B] is greater than 2, preferably greater than 3, and more preferably greater than 5.

[0111] To determine the relationships of different species, a test discrimination method is needed. A simple way to accomplish this task is to attach the fusion partner to one of the Fc subunits (for example, "A"), which could result in each of the three species having a significantly different molecular weight. Accordingly, constructs were prepared for expressing both human Fc (HuFc) and human Fc attached to human IL-2 (HuFc-IL-2) in one cell. These constructs were obtained as follows: the HuFc gene was excised from the vector containing the Fc unit (see, for example, Lo et al., Protein Engineering [1998] 11: 495) by enzymatic restriction at the 5 'XbaI site and the 3' XhoI site. The 1.4 thousand base pair fragment containing the HuFc gene was gel purified and subcloned into the second vector, pdCS-MuFc-KS-kappa, replacing its muFc with HuFc. The HuFc gene was flanked by two Sall sites outside the promoter region.

[0112] A third vector containing the gene coding for HuFc-IL-2 and a single SalI site was selected to accept the HuFc gene. This vector was cut with Sall, and processed with Calf Intestinal.

Phosphatase - CIP) and gel purified. The second vector was digested with

Sall a, containing 2.5 thousand base pairs, was purified by gel.

- 51 EP 1 999 154 B1

This fragment contained the HuFc gene and promoter and was inserted into the gel purified third vector. The resulting vector contained two different transcription units with duplicate versions of the same regulatory elements, one transcription unit controlling wild-type HuFc expression and another wild-type HuFc-IL-2 expression unit. Expression constructs containing HuFc based on SEED and HuFc-IL-2 based on SEED were produced analogously.

[0113] The final vector was extended using Qiagen maxi-prep. 10 mg DNA was used for transient transfection of baby hamster kidney (BHK) cells using the Lipofectamine TM2000 kit (Invitrogen). The cells were separated, one half grew in a regular center and the other half in a serum-free center for two days. Supernatants (for example, 100 μθ were collected. 10 microliters of protein-A grains were added and stirring was carried out overnight at 4^aboutC for protein binding. After washing three times with PBS containing 1% Triton-X100, the samples were loaded on Nu-Page (Invitrogen) with a 4-12% Bis-Tris gel gradient, both under reducing and non-reducing conditions. Gels were stained with colloid dye (Invitrogen) for direct visualization of protein.

[0114] Typical control results are shown in lines 8-10 on the gels shown in Figure 12. The reducing gel in Figure 12C shows the ratio of HuFc and HuFc-IL-2 subunits. The non-reducing gel in Figure 12B shows that HuFc and HuFc-IL2 molecules dimerize randomly, with no preference for heterodimerization over homodimerization.

[0115] Gels were transferred to nitrocellulose membranes for Western blot analysis. In Western blot analyzes, proteins were detected in two ways to measure both Fc and IL-2. Antibodies to human IgG Fc (Jackson Immunolabs) conjugated to horseradish peroxidase (HRP) were used to detect Fc. Blots were detected with ECL substrate and film exposure. Biotinylated anti-human IL-2 antibody (R&amp;D systems) was used to detect IL-2 and the signal was extracted by adding HRP-conjugated avidin and detection from ECL and exposure

- film. These experiments confirmed the identity of the bands shown in Figure 12.

[0116] To measure levels of heterodimers and homodimers produced during the expression of white "Full" SEED GA / SEED AG and "Surface" SEED GA / SEED AG, similar experiments were performed. Single expression vector constructs expressing the AG-IL2 SEED fusion protein and the GA SEED protein were constructed as described above for Fc / Fc-IL2 expression. As shown on belts 2-4 of Figs. 12B and 12C, when the "Full" GA SEED (Fc (GA (f1)) SEED) and "Full" SEED AG-IL-2 (Fc (AG (f0)) SEED) -IL2) co-expressed in NS / 0, heterodimerization was strongly preferred, with no detectable Fc homodimers (AG SEED) -IL2 when only very small amounts of Fc homodimers (GA SEED) were detected. Similarly, as shown in belts 5-7 in Figures 12B and 12C, when the "Surface" SEED GA (Fc (GA (s0) SEED) and "Surface" SEED AG-IL-2 (Fc (AG (s0) SEED) -IL2) proteins were co-expressed in NS / 0 cells, heterodimerization was strongly favored, with undetectable Fc homodimers (AG SEED) -IL-2 and Fc homodimer (GA SEED) only detectable in small amounts. It was estimated that heterodimers accounted for approximately more than 90% of the total amount of proteins deposited in the cell.

EXAMPLE 6. Construction, expression and heterodimerization properties of SEED molecules with reduced immunogenicity [0117] Because the SEED AG and GA protein sequences are hybrids between two naturally occurring human sequences, such sequences contain peptide segments that are not found in normal human proteins, and which can be processed into both MHC class II T cell epitopes. Therefore, the following sequences were designed to reduce the number of potential foreign T cell epitopes in the SEED AG and SEED GA sequences, depicted by the polypeptide sequences shown in SEQ ID NO: 1 and SEQ ID NO: 2, respectively, wherein X1, X2, X3, X4, X5 or X6 can be any amino acid. In some embodiments, in the sequence of SEQ ID NO: 1, X1 is S, X2 is V or T, and X3 is S. In some embodiments, in the sequence SEQ ID NO: 2, X1 is Q, X2 is A or T, X3 is L, V, D

Or T, X4 is F, A, D, E, G, H, K, N, P, Q, R, S or T, X5 is T and X6 is

D.

SEQ ID NO: 1 [0118] Polypeptide sequence for SEED AG, with different amino acids X1 X3:

GQPFRPEVHLLPPSREEMTKNQVSLTCLARGFYPX1 DIAVEWESNGQPENNYKTT PSRQEPSQGTT TFAVTSKLTX2DKSRWQQGNVFSCSVMHEALHNHYTQKX3ISL

SEQ ID NO: 2 [0119] SEED GA polypeptide sequence with different amino acids X1 X6:

GQPREPQVYTLPPPSEELALNEX₁VTL'TCLVKGFYPSDIAVEWLQGSQEŁPREKYLTWX; .PVX-, DS DGSX, FLYSILRVX₅AX₆DWKKGDTFSCSVMHEALHNHYTQKSŁDR.

[0120] A DNA molecule (SEQ ID NO: 32) coding for an exemplary variant of SEED GA (f1) (SEQ ID NO: 7) was produced using de novo synthesis and was introduced into the pdCs expression vector as described in Example 4, producing an Fc polypeptide (GA (f1) SEED).

SEQ ID NO: 7 [0121] Polypeptide sequence for SEED GA (f1):

GQPREPQVYTLPPPSEELALNEQVTŁTCLVKGFYPSDIAVEWLQGSQELPREKYLTWTPWDSDG sfflysilrvtaddwkkgdtfscsvmhealhnhytqksldr [0122] Mutations were introduced to the example SE ID SEED SE ID NO SEED SE 8), using a two-step PCR approach in which two mutagenized, partially overlapping PCR fragments from the first round of PCT amplification are combined in a second round of PCR amplification to generate the full-length final fragment using standard methods known to those skilled in the art. Basically two

PCR reactions are carried out in the first round, with each PCR primer incorporating a mutation sequence paired with the appropriate flanking primer containing the respective restriction sites, Ngo MIV for the upstream primer and Sma I for the primer Down and DNA template encoding the respective parent SEED. The same flanking PCR primers were used in the second PCR amplification reaction, using the first PCR amplification products as templates. The resulting fragment was cloned into pCR2.1 vector (Invitrogen) and its sequence was verified. Finally, the 0.4 kb Ngo MIV / Sma I DNA fragment was excised from this vector, gel purified and ligated into a similarly treated expression plasmid as described in Example 4.

[0123] Specifically, for SEED AG (f1), the primer pair Oligo1 (SEQ ID NO: 40) / Oligo2 (SEQ ID NO: 41) and Oligo3 (SEQ ID NO: 42) / Oligo4 (SEQ ID NO: 43) with The pdCs-Fc (AG (f0) SEED) -IL2 template was used in the first round of the PCR reaction. Oligo1 (SEQ ID NO: 40) / Oligo4 (SEQ ID NO: 43) was used in the second round of the PCR reaction, generating the DNA fragment shown in the sequence SEQ ID NO: 44, which was introduced into pdCs-Fc (AG (f0) SEED ) IL2. For SEED AG (f1), the primer pair Oligo1 (SEQ ID NO: 40) / Oligo5 (SEQ ID NO: 45) and Oligo6 (SEQ ID NO: 46) / Oligo4. (SEQ ID NO: 43) with the template pCR2.1 containing the sequence shown in SEQ ID NO: 44 was used in the first round of PCR. Oligo1 (SEQ ID NO: 40) / Oligo4 (SEQ ID NO: 43) were used in the second round of the PCR reaction, generating the DNA fragment shown in the SEQ ID NO: 47 sequence that was introduced into pdCs-Fc (AG (f0) SEED ) IL2. For SEED GA (f2), the primer pair Oligo1 (SEQ ID NO: 40) / Oligo10 (SEQ ID NO: 48) and Oligo7 (SEQ ID NO: 49) / Oligo9 (SEQ ID NO: 50) with the template plasmid pUC containing the sequence shown in SEQ ID NO: 32 was used in the first round of PCR. Oligo1 (SEQ ID NO: 40) / Oligo9 (SEQ ID NO: 50) were used in the second round of the PCR reaction, generating the DNA fragment shown in the SEQ ID NO: 47 sequence that was introduced into pdCs-Fc (GA (f2) SEED ). The following are all the sequences referenced above.

- EP 1 999 154 B1

Oligo1 (SEQ ID NO: 40)

GCCGGCTCGGCCCACCCTCT Oligo2 (SEQ ID NO: 41)

CGGCGATGTCGCTGGGATAGAA

Oligo3 (SEQ ID NO: 42)

TTCTATCCCAGCGACATCGCCG

Oligo4 (SEQ ID NO: 43)

CCCGGGGACAGGGAGATGGACTTCTGCGTGT

Oligo5 (SEQ ID NO: 45)

GCTCTTGTCTGTGGTGAGCTT

Oligo6 (SEQ ID NO: 46)

AAGCTCACCACAGACAAGAGC

Oligo7 (SEQ ID NO: 49)

CCTGACTTGGGCACCCGTGGACGACTCCGACGGCTCCCACTTCCTCTATA

Oligo9 (SEQ ID NO: 50)

CCCGGGGAGCGGTCGAGGCTC

Oligo10 (SEQ ID NO: 48)

TATAGAGGAAGTGGGAGCCGTCGGAGTCGTCCAGGGGTGCCCAAGTCAGG

SEQ ID NO: 44 [0124] Ngo MIV / Sma I DNA fragment containing the SEED coding sequence

AG (f) (underline):

gccggctcggcccaccctctgccctgagagtgaccgctgtaccaacctctgtccctacaGGGCAGCCCTTCCGGC

CAGAGGTCCACCTGCTGCCCCCATCACGGGAGGAGATGACCAAGAACCAGOTCAGCCTGACCTGC

CTGGCACGCGGCTTCTATCCCAgcGACATCGCCGTGGAGTGGGAGAGCAATGGGCAGCCGGAGAA

CAACTACAAGACCACGCCTTCCCGGCAGGAGCCCAGCCAGGGCACCACCACCTTCGCTGTGACCT

CGAAGCTCACCGTGGACAAGAGCAGATGGCAGCAGGGGAACGTCTTCTCATGCTCCGTGATGCAT GAGGCTCTGCACAACCACTACACGCAGAAGtCCATCTCCCTGt ccccggg

SEQ ID NO: 47 [0125] Ngo MIV / Sma I DNA fragment containing the coding sequence of SEED 25 AG (f2) (underline):

GccggctcggcccaccctctgccctgagagtgaccgctgtaccaacctctgtccctacaGGGCAGCCCTTCC

GGCCAGAGGTCCACCTGCTGCCCCCATCACGGGAGGAGATGACCAAGAACCAGGTCAGCCTGACCTGCCTGG

CACGCGGCTTCTATCCCAgcGACATCGCCGTGGAGTGGGAGAGCAATGGGCAGCCGGAGAACAACTACAAGA

CCACGCCTTCCCGGCAGGAGCCCAGCCAGGGCACCACCACCTTCGCTGTGACCTCGAAGCTCACCacaGACA

AGAGCAGATGGCAGCAGGGGAACGTCTTCTCATCCTCCGTGATGCATGAGGCTCTGCACAACCACTACACGC

AGAAGtCCATCTCCCTGtccccggg [0126] Fc (AG (f1) SEED), Fc (AG (f2) SEED), Fc (GA (f1) SEED) -IL2 and Fc (GA (f2) SEED) -IL2 sequences were expressed individually and in combinations in HEK 293T cells and the resulting secreted proteins were partially purified based on Fc binding to Staphylococcus protein A and characterized by SDS-PAGE. When the samples were loaded on a reducing SDS gel, it became apparent that Fc (AG (f1) SEED) and Fc proteins (AG (f2) SEED) were expressed very weakly without help, which is similar to the parent Fc protein (AG (f0) SEED). Without being limited to theoretical considerations, poor expression most likely results from the proteolysis of a monomeric protein that has no dimerization partner. The Fc (GA (f1) SEED) -IL2 protein was expressed at a high level, while the Fc (GA (f2) SEED) -IL2 protein, differing by an additional Va175Thr amino acid substitution, was expressed at a very low level. Again, without being limited to theoretical considerations, poor expression may result from proteolysis of a monomeric protein that has no dimerization partner. Combinations Fc (AG (f1) SEED) plus Fc (GA (f1) SEED) -IL2, Fc (AG (f2) SEED) plus Fc (GA (f1) SEED) -IL2, Fc (AG (f1) SEED) plus Fc (GA (f2) SEED) -IL2 and Fc (AG (f2) SEED) plus Fc (GA (f2) SEED) -IL2 were tested and all were expressed at high levels. The same samples were loaded onto the non-reducing gel and these results were confirmed. This analysis indicated that for these combinations essentially all expressed proteins were heterodimeric. These results indicate that certain GA and AG variant SEED proteins with reduced immunogenicity retain their heterodimerization ability.

EXAMPLE 7. Expression of an antibody-cytokine fusion protein using SEED Fc regions [0127] To further demonstrate the universality of SEED-based Fc regions, an intact antibody with a single IL-2 unit was constructed as described in Example A. A diagram of this protein is shown in

Fig. 11B. In particular, this protein contained V antibody regions that bind to EpCAM, and which have the sequences described in US Patent Application Publication US 6,696,517, CH1 and CH2 domains of human IgG1, human Ckappa, SEED GA and AG domains, and human IL-2 combined fused to the C-terminus of the heavy chain containing SEED AG.

[0128] This protein was expressed in mammalian cells according to standard techniques for producing protein from the polypeptide chains shown in SEQ ID NO: 37, SEQ ID NO: 36 and SEQ ID NO: 35.

[0129] The obtained protein was characterized to determine the extent to which heterodimeric forms were secreted from mammalian cells. For example, the secreted protein was characterized by non-reducing gel electrophoresis SDSpoliacrylamide. In principle, three bands corresponding to antibodies without any, with one or two IL-2 molecules, could be identified. The actual non-reducing gel showed a predominantly single band with the molecular weight corresponding to the antibody with a single IL-2 molecule. A much lower intensity band with a molecular weight not corresponding to any IL-2 molecule was observed, and a band with a molecular weight corresponding to two molecules was not detected. When the samples were reduced before loading on the gel, approximately equal amounts of protein corresponding to the heavy chain of the antibody and heavy chain-IL2 were detected.

[0130] The above description of the present invention is an illustration and description thereof, but is not intended to be an exhaustive description of it, nor does it limit the scope of the invention strictly to the scope of the disclosed examples of its implementation. Various modifications and variations of the invention may be developed by practicing the present invention. Therefore, it should be emphasized that the scope of the present invention is defined by the claims and equivalents for the solutions defined therein.

Merck Patent GmbH Representative:

- 58 EP 1 999 154 B1