WO2014023315A1

WO2014023315A1 - Multimerization through beta-strand swapping in ccp domains

Info

Publication number: WO2014023315A1
Application number: PCT/DK2013/050261
Authority: WO
Inventors: Christian Brix Folsted ANDERSEN; Anders ETZERODT; Søren Kragh MOESTRUP; Morten TORVUND-JENSEN
Original assignee: Aarhus Universitet
Priority date: 2012-08-10
Filing date: 2013-08-08
Publication date: 2014-02-13

Abstract

The present invention relates to providing means for heterologous expression of dimer-/multimerizing proteins such as multimerizing single-chain antibodies in a bacterial system.

Description

Multimerization through beta-strand swapping in CCP domains

Technical field of the invention

The present invention relates to providing means for heterologous expression of dimer and multimerizing proteins such as multimerizing single-chain antibodies in a bacterial system.

Background of the invention

A protein domain is a part of protein sequence and structure that can evolve, function, and exist independently of the rest of the protein chain. Each domain forms a compact three-dimensional structure and often can be independently stable and folded.

Many proteins consist of several structural domains. One domain may appear in a variety of different proteins. Molecular evolution uses domains as building blocks and these may be recombined in different arrangements to create proteins with different functions.

Domains vary in length from between a few amino acids up to 500 amino acids in length. The shortest domains such as zinc fingers are stabilized by metal ions or disulfide bridges.

Because they are independently stable, domains can be combined by genetic engineering between one protein and another to make chimeric or recombinant hybrid proteins.

Bacterial protein expression systems are popular because bacteria are easy to culture, grow fast and produce high yields of recombinant hybrid protein.

Bacteria are simple and cost effective hosts for producing recombinant proteins.

However, their physiological features may limit their use for obtaining in native form proteins of some specific structural classes, such as for instance polypeptides that undergo extensive post-translational modifications. Multi-domain eukaryotic proteins expressed in bacteria often are non-functional because the cells are not equipped to accomplish the required post-translational modifications or molecular folding.

In this respect, the production of proteins that depending on disulfide bridges for their stability has been considered difficult in E. coli.

Disulfide bridges can stabilize protein structure and are often present in high abundance in secreted proteins.

Haptoglobin is a haemoglobin-binding plasma protein, which, in humans, exists in multimer forms due to interactions between so-called complement control protein CCP/Sushi domains.

These interactions are thought to be stabilized by disulphide bridges and such bridges have hitherto been considered to be essential for the coupling of CCP domains. Heterologous expression of dimer-/multimerizing proteins (such as multimerizing single-chain antibodies) in a bacterial system for example through the use of haptoglobin multimerizing as a model is thus not realistic due to the dependency on disulfide bridges. WO 2008/143794 discloses a fusion protein complex including a fusion protein where a polypeptide is fused to an IL 15Ra polypeptide, which may comprise a Sushi domain. This particular Sushi is not capable of forming dimers/multimers through beta-strand swapping. WO 2010/037837 discloses a bispecific single chain antibody of comprising a Suchi/CCP domain of TEM1. The Suchi/CCP domain of TEM1 also does not have the capacity of forming dimers/multimers through beta-strand swapping.

US 2003/0082630A1 discloses a polypeptide comprising at least two monomers domains separated by a linker, where the polypeptide domain may be selected from a long list of domains including Sushi domains. However, US 2003/0082630A1 is silent about haptoglobin.

Thus, there is a need for providing means for heterologous expression of dimer and multimerizing proteins (such as multimerizing single-chain antibodies) that utilize beta-strand swapping in a bacterial system.

Summary of the invention

An aspect of the present invention relates to a recombinant hybrid polypeptide, said polypeptide comprising a first polypeptide amino acid sequence and a second heterologous amino acid sequence wherein said second amino acid sequence is a complement control protein domain (CCP), wherein said hybrid polypeptide is capable of forming a dimer or a multimer through intermolecular beta-strand swapping.

Another aspect of the present invention relates to a recombinant hybrid

polypeptide, said polypeptide comprising a first polypeptide amino acid sequence and a second heterologous amino acid sequence, wherein said second amino acid sequence is selected from the group consisting of an amino acid sequence selected from the group consisting of SEQ ID NO: 1 or SEQ ID NO: 2, b) an amino acid sequence having at least 75% sequence identity to a), and an amino acid sequence which is a sub-sequence hereof, where said sub-sequence is having a minimum length of 10 amino acids. One embodiment of the present invention relates to a recombinant hybrid polypeptide of the present invention, which is capable forming a dimer or a multimer through intermolecular beta-strand swapping without the involvement of disulphide bridges. Another embodiment of the present invention relates to a recombinant hybrid polypeptide of the present invention wherein the CCP domain is from haptoglobin or haptoglobin-related protein (Hpr). Yet another embodiment of the present invention relates to a recombinant hybrid polypeptide of the present invention, wherein the first polypeptide amino acid sequence is selected from the group consisting of a single-chain antibody, insulin, cytokine or other kinds of biological effector molecules.

A further embodiment of the present invention relates to a recombinant hybrid polypeptide of the present invention wherein the first polypeptide and the second heterologous amino acid sequence is joined by a linker. An aspect of the present invention relates to a nucleic acid molecule comprising a nucleic acid sequence encoding the recombinant hybrid polypeptide of the present invention.

Another aspect of the present invention relates to a vector comprising the nucleic acid molecule of the invention.

Yet another aspect of the present invention relates to a cell comprising the above mentioned molecules of the invention. Figure legends

Figure 1 - Crystal structure of porcine HpHb

a, Structure of porcine HpHb. Haem groups are shown as dark grey sticks. Dark spheres represent Fe ions. Glycosylations are shown as light grey sticks and disulfide bridges as gray sticks, b, Stick representation of the Bl/B2-strands of Hp CCP domains, c, Interface between Hp CCP and Hp SP. Hp SP is coloured dark gray and Hp CCP light gray. Dashed lines represent hydrogen bonds and numbers indicate bond lengths (A), d, Structure of the Hp SP loop 3 region. Residues important for interaction with the CD163 scavenger receptor3 are shown as sticks. Figure 2 -The Hp-bound Hb dimer is in its oxy-state

a, UV-vis spectra of a porcine HpHb crystal (black line), porcine HpHb in solution (drak gray lines) and porcine Hb in solution (light gray lines), b+c,

Superimposition of porcine Hp bound aHb (b) or Hb (c), oxygenated human aHb (b) or Hb (c) (PDB accession 2DN1), and deoxygenated aHb (b) or Hb (c) (light gray, PDB accession 2DN2). Haem groups, oxygen molecules, proximal histidines (H87 in aHb, H92 in βΙ-lb), and distal histidines (H58 in aHb, H63 in βΙ-lb) are shown as sticks. Fe ions are shown as spheres. Figure 3 - The Hb contact area overlaps with the Hb dimer contact area in Hb tetramers

a, Surface representation of Hb in the HpHb complex. Residues within 3.8 A of Hp SP are coloured blue, b, Surface representation of human a l 1Hb in oxygenated tetrameric Hb (2DN1). alHb residues within 3.8 A of 2Hb and 1Hb residues within 3.8 A of 2Hb. c-f, Selected interactions between aHb (c,d) or Hb (e,f) and Hp SP. Dashed lines represent electrostatic interactions or hydrogen bonds.

Figure 4 - Hb residues prone to oxidative modifications and small angle x-ray scattering (SAXS)

a, Human Hb residues prone to oxidative modifications indicated on the structure of porcine HpHb (spheres, corresponding porcine residues in parentheses), b, SAXS curves of porcine HpHb (pHpHb), human HpHb (hHpHb) and CD163 SRCR 1-5 bound to hHpHb. Circles represent experimental data and lines theoretical intensities. The structure of pHpHb with modelled glycosylations gives the best fit to the experimental data, c, Ab initio modelling of hHpHb bound to CD163 SRCR 1-5. The modelling based on imposed P2 symmetry (a, figure 15) was performed using the structure of pHpHb with modelled glycosylations and the structure of the M2BP (PDB accession 1BY2). Figure 5 - Comparison of porcine Hp SP and bovine Chymotrypsin A

a, Structural alignment of porcine Hp SP and bovine Chymotrypsin A. Secondary structure elements are shown above the alignment, b, Superimposition of porcine Hp SP (dark gray) and bovine Chymotrypsin A (PDB accession 4CHA, gray).

Numbers and letters indicate the surface loop nomenclature described previously. "Catalytic triad" residues are marked by black boxes (a) or shown as sticks (b).

Figure 6 - Sequence alignment of porcine Hp, human Hpl and human Hpr

Secondary structure elements are shown above the alignment. Residues involved in Hp-Hp dimerisation and Hp-Hb interactions are marked (Hp-Hp) and dark (Hp- Hb) bars above the alignment. Glycosylation sites are marked by black boxes and residues important for CD163 interaction by boxes (261, 262, 264).

Figure 7 - Sequence alignment of porcine Hb and human Hb (top: aHb, bottom : bHb).

Secondary structure elements from porcine Hb in complex with Hp are shown above the alignment. Residues involved Hp-Hb interactions are marked by bars above the alignment. Figure 8 - Structure of Hp CCP and comparison with Clr CCP

a, Structure of a CCP domain from Clr (PDB accession 2QY0). The domain is stabilised by two disulphide bonds (C389/C430 and C359/C412). b, Structure of the Hp CCP fusion domain in the porcine HpHb complex. Individual Hp polypeptide chains are coloured in shades of gray. An intermolecular disulphide bond between two Cys33 residues compensates the lack of a cysteine at residue 68. c+d, Electron density maps (2mFo-DFc, omit maps contoured at 1σ) of the B1/B2 β- strands (c) and the entire Hp CCP fusion domain (d).

Figure 9

a, Crystal of porcine HpHb. b, Resonance Raman spectroscopy (Aex = 410 nm) on porcine HpHb in crystals show a v4 mode at 1377 cm-1, which is indicative of Hb in its oxygenated state as also observed for the porcine Hb frozen solution.

Figure 10

a+b, Electron density map (2mFo-DFc, omit map contoured at 1σ) of haem groups bound to aHb (a) or Hb (b) in the HpHb complex, c+d, Superimposition of porcine Hp bound aHb (c) or Hb (d) (dark gray), oxygenated human aHb (c) or Hb (d) (gray, PDB accession 2DN1), and deoxygenated aHb (c) or Hb (d) (light gray, PDB accession 2DN2). Haem groups, oxygen molecules, proximal histidines (H87 in aHb, H92 in 3Hb), and distal histidines (H58 in aHb, H63 in Hb) are shown as sticks. Fe ions are shown as spheres. Figure 11

Comparison of the enzyme-product complex of Clr with the HpHb complex a, Enzyme-product complex of Clr (PDB accession 2QY0). Clr is shown as grey ribbons, b, Position of the C-terminus of aHb in the porcine HpHb complex. aHb is show as orange ribbons and Hp SP in dark ribbons.

Figure 12 - Comparison of the porcine HpHb complex with human oxyHb a, Structure of porcine a&Hb binding to Hp SP. Hp SP is dark, aHb and Hb gray. b, Structure of human oxyHb (PDB accession 2DN1). alHb and 1Hb is gray.

a2Hb and 2Hb is dark. Haem groups are shown as light grey sticks. Spheres represent Fe ions. aHb/ Hb (a) and alHb/ 1Hb (b) are shown in the same orientation.

Figure 13

a, Schematic overview of Hp-Hb interactions. Hb residues are left and Hp residues right. Structural regions as defined in figures 6+7 are indicated by letters or numbers next to the boxes. Grey lines represent van der Waals contacts, dashed black lines represent hydrogen bonds and black lines represent salt bridges. Hb residues involved in alHb- 2Hb/ 1Hb-a2Hb contacts in human tetrameric oxy/deoxy Hb (PDB accession 2DN1/2DN2) are marked by asterisks. The electron density map suggests that Hb GlulOl occurs in two conformations (ElOla and ElOlb).

Figure 14 - Superimposition of human deoxygenated a^Hb on porcine Hp bound oxygenated a&Hb

Selected aHb (a) or Hb (b) residues interacting with Hp are shown as gray sticks. Corresponding residues of deoxygenated human Hb are shown as light grey sticks. Selected Hp residues are shown as dark gray sticks. Dashed lines represent hydrogen bonds or salt bridges.

Figure 15 - Models of human multimeric Hp in complex with Hb

a, Schematic representation of selected human HpHb multimers in individuals with different Hp genotypes2. Numbers indicate Hpl (1) and Hp2 (2) gene products. Split circles represent fusion Hp CCP domains, b-d, Models of human trimeric Hpl/Hp2 (b), trimeric Hp2 (c) and tetrameric Hp2 (d) based on the structure of dimeric porcine HpHb. Arrows indicate Hp loop 3 involved in CD163 binding.

Figure 16 - Small angle x-ray scattering of human HpHb-CD163 SRCR 1-5 a+b, Ab initio modelling of human dimeric HpHb bound to CD163 SRCR 1-5. The modelling was performed using the model for glycosylated human HpHb

(described in Fig. 4d) and assuming one receptor (a, HpHb-lxCD163) or two receptors per complex (b, HpHb-2xCD163) without P2 symmetry, c, SAXS scattering curves of human HpHb bound to CD163 SRCR 1-5. Circles represent experimental data and lines theoretical intensities calculated from models. The experimental data set has been repeated for better viewing of the model curves.

Figure 17 - Models of dimeric human HpHb binding to CD163 on the surface of macrophages

The SAXS studies show that CD163 SRCR domains 1-5 bind HpHb along the longitudinal axis. However, the orientation CD163 SRCR 1-5 in relation to HpHb remains unknown. Therefore, two different models (shown in a and b) of HpHb induced CD163 cross-linking are proposed. Hp SP is shown as large spheres. The fused Hp CCP domain is shown as a split sphere. Hb subunits are shown as light gray spheres and the SRCR domains of CD163 are shown as dark smaller spheres. CD163 contains an extended linker region between SRCR 6 and 7, as indicated by a black line. Horizontal grey lines indicate the plasma membrane

Figure 18 - Schematic representation of KN2 scFv-CCP constructs.

The design of CCP fusion platform consists of the dimeric CCP domain found in Hpl-1 (CCP1) and the multimeric CCP domain found in Hp2-2 (CCP2). CCP2 corresponding to aa 19-148 in Hp2-2 and CCP1 corresponding to aa 19-89 in Hpl- 1 are N-terminally fused to the KN2 scFv fragment. The constructs include a cysteine to alanine mutation (C->A) to avoid crosslinking and precipitation of the expressed fusion proteins. Moreover, the possibility to add a free C-terminal cysteine allows for subsequent chemical modification with i.e. toxic drugs. Figure 19 - Western blot of KN2-CCP1/CCP2 expression on picha pastoris

Western blot of growth media from single colonies of Pichia Pastoris transformed with plasmid for KN2 scFv-CCPl/CCP2 expression. Immunereactive bands were visualized using HRP-conjugated anti-human IgG antibody (Dako). Clone KN2- CCP1#1, KN2-CCP1#2, KN2-CCP1#3 and KN2-CCP1#4 are positive for KN2 scFv expression. Clone KN2-CCP2#2 is positive for KN2 scFv-CCP2 expression.

Figure 20 - Purification strategy for KN2 scFv-CCPl

Purification strategy of KN2 scFv-CCPl from KN2-CCP1 positive clones. Media containing KN2 scFv-CCPl is obtained by culturing Clone KN2-CCP1-#1 for 72 hours in BMMY supplemented with 2% methanol every 12 hours.

Figure 21 - Coomasie stain of SDS-PAGE loaded with samples from purification strategy

Coomasie stain of SDS loaded wtih samples from the above presented purification strategy. KN2 scFv-CCPl is clearly precipitated at 40 % ammonium sulfate. The precipitate is reconstituted in 1 ml 20 mM Tris pH 7.6 and subjected to Size exclusion chromatography for further purification. Figure 22 - Size exclusion chromatography of 40% AMS precipitate containing KN2 scFv-CCPl

Chromatogram of 40% AMS precipitate from culture media containing KN2 scFv- CCP1. 0,5 ml reconstituted precipitate is fractionated using a Superdex200 10/30 column. Fraction are analyzed by SDS-PAGE to identify peak containing KN2 scFv- CCP1.

Figure 23 - Coomasie stain of SDS-PAGE loaded with fractionated KN2 scFv-CCPl Protein on fraction B8 to B15 corresponds to KN2 scFv-CCPl. Fractions are pooled and concentrated using Amicon concentrators.

Figure 24 - Biacore analysis of purified KN2 scFv-CCPl and KN2 scFv

Binding of KN2 scFv and KN2 scFv-CCPl to immobilized human CD163 were analyzed at 25 pg/ml using the Biacore3000. Kd values were calculated based on the molecular mass of monomeric KN2 scFv (26 kDa) and dimeric KN2 scFv-CCPl (67 kDa). Both proteins show similar on-rate whereas the off-rate of KN2 scFv- CCP1 is slower resulting in a lower Kd value. This corresponds to dimeric formation via the CCP1 unit.

Figure 25 - Purification and analysis of the multimeric state of Hp-CCP2. (A) Western blot of growth media from a single colonies of Pichia Pastoris transformed with mock plasmid or plasmid for expression of Hp-CCP2. Immunereactive bands were visualized using HRP-conjugated anti-human Haptoglobin antibody (Dako). Bands corresponding to monomeric (20 kDa) and multimeric Hp-CCP are highlighted with arrows. (B) SDS-PAGE of purified Hp-CCP2. Hp-CCP2 was purified by hydrophobic interaction chromatography from growth supernatant adjusted to 40% ammonium sulfate. Protein was eluted using a gradient from 40% to 80% ammonium sulfate with 20 mM Tris pH 7.6. Protein was visualized by coomasie staining. (C) Native-PAGE of purified Hp-CCP2 fractionated by size-exclusion chromatography using a superdex 200 column (GE Healthcare). Protein was visualized by coomasie staining.

Figure 26 - Expression, purification and binding of Hp-CCPl-RAP. (A) Western blot of growth media from a single colony of Pichia Pastoris transformed with plasmid for expression of Hp-CCPl-RAP. For Western blot of Hp-CCPl-RAP protein was separated either by SDS-PAGE (left) or by NativePAGE (right). Immunereactive bands were visualized using HRP-conjugated anti-human Haptoglobin antibody (Dako). (B) Surface Plasmon resonance analysis of Hp-CCPl-RAP binding. Binding of RAP and Hp-CCPl-RAP to immobilized cubilin (top) or megalin (bottom) were analyzed at 10 pg/ml using the Biacore3000. Kd values were calculated based on the molecular mass of monomeric RAP (50 kDa) and dimeric Hp-CCPl-RAP (100 kDa). Dimerization of RAP changes both on and off rates compared to monomeric RAP resulting in a lowered KD and stronger binding.

Detailed description of the invention

The present invention relates to the expression of the protein of interest as a fusion protein with a complement control protein domain (CCP domain) capable of forming dimers or multimers through beta-strand swapping between the CCP domains.

The inventors have performed detailed experiments using human haptoglobin, SEQ ID NO: 1;

MSALGAVIALLLWGQLFAVDSGNDVTDIADDGCPKPPEIAHGYVEHSVRYQCKNYYKLRTE GDGVYTLNDKKQWINKAVGDKLPECEADDGCPKPPEIAHGYVEHSVRYQCKNYYKLRTEG DGVYTLNNEKQWINKAVGDKLPECEAVCGKPKNPANPVQRILGGHLDAKGSFPWQAKMVS HHNLTTGATLINEQWLLTTAKNLFLNHSENATAKDIAPTLTLYVGKKQLVEIEKVVLHPNYSQ VDIGLIKLKQKVSVNERVMPICLPSKDYAEVGRVGYVSGWGRNANFKFTDHLKYVMLPVAD QDQCIRHYEGSTVPEKKTPKSPVGVQPILNEHTFCAGMSKYQEDTCYGDAGSAFAVHDLEE DTWYATGILSFDKSCAVAEYGVYVKVTSIQDWVQKTIAEN, which shows that beta- strand swapping is possible without the need for disulphide bridges. The human CCP1 domain sequence is

CPKPPEIAHGYVEHSVRYQCKNYYKLRTEGDGVYTLNDKKQWINKAVGDKLPECE (SEQ ID NO: 2) and the human CCP2 domain sequence is

CPKPPEIAHGYVEHSVRYQCKNYYKLRTEGDGVYTLNNEKQWINKAVGDKLPECEA (SEQ ID NO: 3) while the mouse CCP domain sequence is

CPKPPEIANGYVEHLVRYRCRQFYRLRAEGDGVYTLNDEKQWVNTVAGEKLPECEA (SEQ ID NO: 4).

The part of the human CCP domain sequence that is directly involved with beta- strand swapping is lAHGYVEHSVRYQCK (SEQ ID NO: 5) and the part of the mouse CCP domain sequence that is involved with beta-strand swapping is

IANGYVEHLVRYRCR (SEQ ID NO: 6).

Haptoglobin-related protein (Hpr) has the sequence

MSDLGAVISLLLWGRQLFALYSGNDVTDISDDRFPKPPEIANGYVEHLFRYQCKNYYRLRTE GDGVYTLNDKKQWINKAVGDKLPECEAVCGKPKNPANPVQRILGGHLDAKGSFPWQAK MVSHHNLTTGATLINEQWLLTTAKNLFLNHSENATAKDIAPTLTLYVGKKQLVEIEKVVL HPNYHQVDIGLIKLKQKVLVNERVMPICLPSKNYAEVGRVGYVSGWGQSDNFKLTDHLKY VMLPVADQYDCITHYEGSTCPKWKAPKSPVGVQPILNEHTFCVGMSKYQEDTCYGDAGSA FAVHDLEEDTWYAAGILSFDKSCAVAEYGVYVKVTSIQHWVQKTIAEN (SEQ ID NO: 7). The CCP domain of haptoglobin-related protein is

FPKPPEIANGYVEHLFRYQCKNYYRLRTEGDGVYTLNDKKQWINKAVGDKLPEC (SEQ ID NO: 8). Examples of duplicated CCP domain includes human Haptoglobin CCP1+CCP2 with mutated disufide (C>S) :

VDSGNDVTDIADDGSPKPPEIAHGYVEHSVRYQCKNYYKLRTEGDGVYTLNDKKQWINKA VGDKLPECEADDGSPKPPEIAHGYVEHSVRYQCKNYYKLRTEGDGVYTLNNEKQWINKAVG DKLPECEA (SEQ ID NO: 9) and the duplicated human Hpr CCP:

LYSGNDVTDISDDRFPKPPEIANGYVEHLFRYQCKNYYRLRTEGDGVYTLNDKKQWINKAV GDKLPECSDDRFPKPPEIANGYVEHLFRYQCKNYYRLRTEGDGVYTLNDKKQWINKAVGDK LPECEA (SEQ ID NO: 10).

The present inventors have also found that mutations that restrain the geometry of the linker could affect the distribution of CCP multimers. Sites for such mutation are positions 87-91 of SEQ ID NO: 1 corresponding to ADDG.

Thus, in one embodiment of the present invention can one or more of positions 87-91 of SEQ ID NO: 1 be mutated to optimize the distribution of CCP multimers.

The mutated amino acids can, for example, be by changing one or more of positions 87-91 to proline. Such change will limit rotational freedom. In one embodiment of the present invention are all 5 amino acids in positions 87- 91 mutated.

In another embodiment of the present invention are 4 of the 5 amino acids in positions 87-91 mutated.

In another embodiment of the present invention are 3 of the 5 amino acids in positions 87-91 mutated.

In another embodiment of the present invention are 2 of the 5 amino acids in positions 87-91 mutated. In another embodiment of the present invention is 1 of the 5 amino acids in positions 87-91 mutated. These findings and the above sequences can therefore be utilized for designing recombinant hybrid polypeptides that are capable of forming a dimer or a multimer through intermolecular beta-strand swapping.

One hallmark of such recombinant hybrid polypeptides is that they are capable of forming a dimer or a multimer through intermolecular beta-strand swapping without the involvement of disulphide bridges.

Thus, one aspect of the present invention relates to a recombinant hybrid polypeptide, said polypeptide comprising a first polypeptide amino acid sequence and a second heterologous amino acid sequence wherein said second amino acid sequence is a complement control protein domain (CCP), wherein said hybrid polypeptide is capable of forming a dimer or a multimer through intermolecular beta-strand swapping. A further aspect of the present invention relates to a recombinant hybrid polypeptide, said polypeptide comprising a first polypeptide amino acid sequence and a second heterologous amino acid sequence, wherein said second amino acid sequence is selected from the group consisting of an amino acid sequence selected from the group consisting of SEQ ID NO: 2 or SEQ ID NO: 3, and an amino acid sequence having at least 75% sequence identity to SEQ ID NO: 2 or SEQ ID NO: 3, and an amino acid sequence which is a sub-sequence of the above, where said sub-sequence is having a minimum length of 10 amino acids.

This sub-sequence may be full length or have a minimum length of 60 amino acids, such as 55 amino acids, such as 45 amino acids, such as 35 amino acids, such as 30 amino acids, such as 25 amino acids, such as 20 amino acids, such as 18 amino acids, such as 14, such as 13, such as 12, such as 11, such as 10, such as 9, such as 8 amino acids of SEQ ID NO: 2 or SEQ ID NO: 3. In one embodiment of comprises the second amino acid sequence at least the part of the CCP domain that is involved with beta-strand swapping i.e. SEQ ID NO: 5 or 6. In one embodiment of the present invention is the hybrid a dimer.

In another embodiment of the present invention is the hybrid a trimer in which the dimer described above is continued with a third amino acid sequence. The hybrid may also by a multimer of more than three units. Sequence identity

As commonly defined "identity" is here defined as sequence identity between genes or proteins at the nucleotide or amino acid level, respectively.

Thus, in the present context "sequence identity" is a measure of identity between proteins at the amino acid level and a measure of identity between nucleic acids at nucleotide level. The protein sequence identity may be determined by comparing the amino acid sequence in a given position in each sequence when the sequences are aligned.

Similarly, the nucleic acid sequence identity may be determined by comparing the nucleotide sequence in a given position in each sequence when the sequences are aligned.

To determine the percent identity of two nucleic acid sequences or of two amino acids, the sequences are aligned for optimal comparison purposes (e.g., gaps may be introduced in the sequence of a first amino acid or nucleic acid sequence for optimal alignment with a second amino or nucleic acid sequence). The amino acid residues or nucleotides at corresponding amino acid positions or nucleotide positions are then compared.

When a position in the first sequence is occupied by the same amino acid residue or nucleotide as the corresponding position in the second sequence, then the molecules are identical at that position. The percent identity between the two sequences is a function of the number of identical positions shared by the sequences (i.e., % identity = # of identical positions/total # of positions (e.g., overlapping positions) x 100). In one embodiment the two sequences are the same length.

One may manually align the sequences and count the number of identical nucleic acids or amino acids. Alternatively, alignment of two sequences for the

determination of percent identity may be accomplished using a mathematical algorithm. Such an algorithm is incorporated into the NBLAST and XBLAST programs. BLAST nucleotide searches may be performed with the NBLAST program, score = 100, wordlength = 12, to obtain nucleotide sequences homologous to a nucleic acid molecules of the invention. BLAST protein searches may be performed with the XBLAST program, score = 50, wordlength = 3 to obtain amino acid sequences homologous to a protein molecule of the invention.

To obtain gapped alignments for comparison purposes, Gapped BLAST may be utilised. Alternatively, PSI-Blast may be used to perform an iterated search which detects distant relationships between molecules. When utilising the NBLAST, XBLAST, and Gapped BLAST programs, the default parameters of the respective programs may be used. See http://www.ncbi.nlm.nih.gov. Alternatively, sequence identity may be calculated after the sequences have been aligned e.g. by the BLAST program in the EMBL database (http://blast.ncbi.nlm.nih.gov/). Generally, the default settings with respect to e.g. "scoring matrix" and "gap penalty" may be used for alignment. In the context of the present invention, the BLASTN and PSI BLAST default settings may be advantageous.

The percent identity between two sequences may be determined using techniques similar to those described above, with or without allowing gaps. In calculating percent identity, only exact matches are counted.

A comparison of SEQ ID NO: 5 and SEQ ID NO: 6 show that they have 11 out of 15 amino acids in common. This corresponds to around 73 % sequence identity. Thus, one embodiment of the present invention relates to the recombinant hybrid polypeptide of the present invention wherein the second amino acid sequence is an amino acid sequence with at least 60 % sequence identity to SEQ ID NO: 2, such as 65 % sequence identity, such as 70 % sequence identity, such as 75 % sequence identity, such as 80 % sequence identity, such as 85 % sequence identity, such as 90 % sequence identity, such as 95 % sequence identity, such as 97 % sequence identity, such as 99 % sequence identity.

Another embodiment of the present invention relates to the recombinant hybrid polypeptide of the present invention wherein the second amino acid sequence is an amino acid sequence with at least 60 % sequence identity to SEQ ID NO: 3, such as 65 % sequence identity, such as 70 % sequence identity, such as 75 % sequence identity, such as 80 % sequence identity, such as 85 % sequence identity, such as 90 % sequence identity, such as 95 % sequence identity, such as 97 % sequence identity, such as 99 % sequence identity.

Yet another embodiment of the present invention relates to the recombinant hybrid polypeptide of the present invention wherein the second amino acid sequence is an amino acid sequence with at least 60 % sequence identity to SEQ ID NO: 4, such as 65 % sequence identity, such as 70 % sequence identity, such as 75 % sequence identity, such as 80 % sequence identity, such as 85 % sequence identity, such as 90 % sequence identity, such as 95 % sequence identity, such as 97 % sequence identity, such as 99 % sequence identity. A further embodiment of the present invention relates to the recombinant hybrid polypeptide of the present invention wherein the second amino acid sequence is an amino acid sequence with at least 60 % sequence identity to SEQ ID NO: 5 or 6, such as 65 % sequence identity, such as 70 % sequence identity, such as 75 % sequence identity, such as 80 % sequence identity, such as 85 % sequence identity, such as 90 % sequence identity, such as 95 % sequence identity, such as 97 % sequence identity, such as 99 % sequence identity.

Another embodiment of the present invention relates to the recombinant hybrid polypeptide of the present invention wherein the second amino acid sequence is an amino acid sequence with at least 60 % sequence identity to SEQ ID NO: 8, such as 65 % sequence identity, such as 70 % sequence identity, such as 75 % sequence identity, such as 80 % sequence identity, such as 85 % sequence identity, such as 90 % sequence identity, such as 95 % sequence identity, such as 97 % sequence identity, such as 99 % sequence identity.

The above mentioned recombinant hybrid polypeptides will retain the ability to form dimer or multimers while being having one or more alterations (variations) in the sequence that does not have an influence on the functionality. In some cases can the variation of the sequences have an influence on the function, which sometimes can generate stronger or weaker mulitmerization depending on the purpose.

Disulphide bridges

The present inventors have surprisingly shown that the recombinant hybrid polypeptides of the present invention are capable of forming a dimer or a multimer through intermolecular beta-strand swapping without the involvement of disulphide bridges. Thus, in one embodiment of the present invention the recombinant hybrid polypeptide is capable forming a dimer or a multimer through intermolecular beta- strand swapping without the involvement of disulphide bridges.

The CCP domains of the present invention, including SEQ ID NOs. 2, 3, 4 and 8, can be with or without mutated cysteins.

Haptoglobin

In one embodiment of the present invention, the CCP domain of the present invention is from haptoglobin (SEQ ID NO: 1) or haptoglobin-related protein (SEQ ID NO: 7).

Origin

One embodiment of the present invention relates to the origin of the recombinant hybrid polypeptide of the invention. In one embodiment the CCP originates from the group consisting of mouse, human, pig, dog, animal, mammal, rat, hamster, primate, and ape. In another embodiment of the present invention the first polypeptide amino acid sequence originates from the group consisting of mouse, human, animal, mammal, rat, hamster, primate, and ape.

In yet another embodiment of the present invention the first and the second polypeptides are of the same origin.

In another embodiment the first and the second polypeptides are of different origin. Trimers or mulitmers can be made of polypeptides originating from the same or different species.

Polypeptides for multimerization

In one embodiment of the present invention the first polypeptide amino acid sequence is selected from the group consisting of a single-chain antibody, insulin, cytokine or other kinds of biological effector molecules.

Antibodies that can be used for dimerization or multimerization includes anti-TNF- alpha, anti IL6 or/and anti-ILl.

The dimers or multimers of the present invention can also be bi- or multispecific single chain antibodies.

The dimers or multimers of the present invention can be any single-chain fragments of therapeutic or diagnostic antibodies.

The dimers or multimers can also be soluble protein receptor fragments or receptor antagonists. Linkers

The recombinant hybrid polypeptide of the present invention can comprise a linker that joins the first polypeptide and the second heterologous amino acid sequence.

Such linker can be 1-30 amino acids long, such as 27 amino acids, such as 25 amino acids, such as 22 amino acids, such as 20 amino acids, such as 18 amino acids, such as 14 amino acids, such as 12 amino acids, such as 10 amino acids, such as 8 amino acids, such as 8 amino acids, such as 7 amino acids, such as 6 amino acids, such as 5 amino acids, such as 4 amino acids, such as 3 amino acids, such as 2 amino acids, such as 1 amino acid.

In a preferred embodiment the linker is one that can be used for ScFv

(GGSSRSSSSGGGGSGGGG) or a variation hereof.

Nucleic acids

An aspect of the invention is a nucleic acid encoding the fusion protein of the invention. The nucleic acid of the invention may be part of a plasmid or a vector. Thus, one aspect of the present invention relates to a nucleic acid molecule comprising a nucleic acid sequence encoding the recombinant hybrid polypeptide of the present invention.

Another aspect of the present invention relates to a vector comprising the nucleic acid molecule of the present invention.

In such vectors will the nucleic acid molecules usually be fused to a promoter or regions that can regulate the transcripton. Cells and cell types

One aspect of the present invention relates to a cell comprising the recombinant hybrid polypeptide, the nucleic acid molecule or the vector of the present invention. In one embodiment of the present invention the cell according is a bacterial or a yeast cell.

The cell line can be any cell line that is capable of promoting expression of the recombinant hybrid polypeptides of the present invention.

Such cell types can originate from the group consisting of E. coli, S. cerevisiae, T. thermophiles, Leishmania major, Spodoptera frugiperda Sf9 cells, Drosophila melanogaster S2 cells, Chinese Hamster Ovaries (CHO), and Human Hek293,

In a preferred embodiment of the present invention is the cell type the yeast pichia pastoris.

The dimers and multimers of the present invention can also be produced in cell free protein production.

Example 1- Structure of the Haptoglobin-Haemoglobin Complex

Abstract

Red cell haemoglobin (Hb) is the fundamental oxygen-transporting molecule in blood but also a potentially tissue damaging compound due to its highly reactive haem groups.

During intravascular haemolysis, such as in malaria and haemoglobinopathies, Hb is released into plasma, where it is captured by the protective acute-phase protein haptoglobin (Hp), leading to formation of the HpHb complex that represents a virtually irreversible non-covalent protein-protein interaction. Here we present the crystal structure of the dimeric porcine HpHb complex determined at 2.9 A resolution.

This structure reveals that Hp molecules dimerise through an unexpected β-strand swap between two complement control protein (CCP) domains defining a new fusion CCP domain structure. The Hp serine protease domain forms extensive interactions with both the a- and β-subunits of Hb, explaining the tight binding between Hp and Hb.

The Hp interacting region in the Hb αβ-dimer is highly overlapping with the interface between the two αβ-dimers that constitute the native Hb tetramer.

Several Hb residues prone to oxidative modification upon exposure to haem- induced reactive oxygen species are buried in the HpHb interface thus showing a direct protective role of Hp.

The Hp loop previously shown to be essential for binding of HpHb to the macrophage Hb scavenger receptor CD163 protrudes from the surface of the distal end of the complex adjacent to the associated aHb subunit. Small-angle x- ray scattering measurements of human HpHb bound to the ligand-binding fragment of CD163 confirm receptor-binding in this area and show that the rigid dimeric HpHb complex can bind two receptors.

Such receptor cross-linkage may facilitate scavenging and explain the increased functional affinity of multimeric HpHb for CD163.

Experiments

The release of haemoglobin (Hb) during haemolysis and tissue damage is potentially hazardous due to the reactive properties of haem, which can engage in chemical reactions and generate free radicals.

After release into plasma, tetrameric Hb dissociates into αβ-dimers (c^Hb) and is instantly captured by the acute-phase protein haptoglobin (Hp) via a virtually irreversible interaction. In humans, Hp mediates rapid clearance of Hb through high affinity-binding to the macrophage scavenger receptor CD163. In addition, Hp protects tissues and cells from Hb-induced oxidative damage and preserves the structural integrity of Hb whereby its clearance is retained. Hp-Hb complex formation also prevents renal filtration of Hb and toxic effects in the kidneys. A complement control protein (CCP) domain and a serine protease (SP) domain form the structural entities of Hp. The CCP-SP assembly is a feature of several serine proteases, including complement factors Cls and Clr. However, Hp is not an active protease due to an incomplete 'catalytic triad' (figure 5. Furthermore, a unique feature of Hp is the dimerisation of CCP domains.

In order to elucidate the structural basis for Hp-mediated recognition and protection of Hb, we determined the crystal structure of HpHb purified from porcine blood to 2.9 A resolution (Table 1). Whereas the first crystal structure of Hb was reported more than five decades ago, previous attempts on structure determination of human Hp or the HpHb complex have proved unsuccessful.

Porcine HpHb exhibits 82% sequence identity with its human counterpart (figures 6 and 7) and has an overall shape resembling a barbell (180A x 65A x 5θΑ) with a two-fold rotational symmetry around its centre (Fig. la). The CCP domains connect two Hp molecules, whereas the Hp SP domains are responsible for Hb binding.

The Hp CCP domain (residue 33-90) has a β-sandwich arrangement similar to CCP domains in Clr/Cls (Fig. la). CCP domains usually contain four cysteine residues forming two disulfide bridges, but the Hp CCP domain lacks a cysteine at position 68. Instead Cys33 engages in an interchain disulphide bridge linking two CCP domains (Fig. lb). Our structure reveals that the two Hp CCP domains dimerise through a β-strand swap not previously observed for CCP domains (Fig. la, Figure 8). Instead of forming an anti-parallel β-sheet, strands Bl and B2 combine into a single B1/B2 strand forming an anti-parallel β-sheet with B1/B2 of the opposing CCP domain (Fig. lb). This results in a hitherto unknown fusion CCP domain structure containing a central six-stranded β-sheet.

Electron micrographs showing HpHb as a rigid structure is supported by the presence of two nearly identical porcine HpHb dimers in the asymmetric unit of the crystals. The rigidity is most likely achieved by means of specific interactions between the CCP and SP domains of Hp. In particular, the side chains of Tyr56 and Glul35 form hydrogen bonds with each other and with the main chains of Ala88 and Vall99 (Fig. lc). In addition, Tyr55, Ala88, Vall99, and Glu201 are involved in van der Waals contacts. The Hp SP domain has the typical fold of chymotrypsin-like SPs with two anti- parallel β-barrel subdomains each containing six β-strands and two or three a- helices. Several surface loop regions differ in length and conformation compared to other SP domains (figure 5). In particular, the region designated loop 3

(residues 258-274) in SPs is extended (Fig. Id).

This region protruding from the surface is involved in CD163 interaction because mutation of Val259, Glu261, Lys262, and Thr264 in human Hp disrupts CD163 binding. These residues are conserved in porcine Hp and located at the tip of the loop. Loop 3 residues 267-271 interact directly with aHb, which may affect the conformation of the loop and influence CD163 binding.

The crystallised porcine HpHb complex contains haem in oxygenated ferrous (Fe(II)) form as evidenced by the bright red colour of the crystals (figure XX5), the distinct a and β absorption bands at 575 and 538 nm (Fig. 2a) and the v4 Raman mode at 1377 cm-1 (figure 9b). Identical spectra are observed for porcine HpHb and Hb in solution.

In agreement, Hp bound a^Hb is in a conformation similar to human oxygenated a Hb with the Fe atom positioned in the plane of the haem group (Fig. 2b+c, figure 10).

This indicates that macrophages metabolise the oxygenated HpHb complex, which may help fuelling the oxygen-dependent haem conversion by the haem

oxygenases.

The Hb binding site on Hp SP resides in surface exposed loops, including loop A (residues 121-127), loop D (226-234), loop 1 (residues 283-289), loop 2

(residues 318-327), and loop 3 (residues 253-277) (figure 5). In addition, the N- terminal region of Hp SP (residues 104-110) also contacts Hb. Although Hp is not an active protease, the Hb binding site in Hp is located in the region responsible for substrate specificity in other SPsl4. Furthermore, the C-terminus of aHb is in a position resembling the enzyme-product complex observed in Clr although aHb Argl41 is positioned outside the SI pocket (figure 11). These observations suggest that the Hp-Hb interaction originates from a product-like complex between the C-terminus of aHb and an active serine protease.

Hp interacts extensively with both Hb subunits. The binding site on aHb includes residues from helix G, helix H, and FG loop, while residues from Hb helix C, helix G and the FG loop contact Hp (figure 3a).

Remarkably, helix C, helix G and the FG loop also constitute the primary sites for interaction between alHb and 2Hb, and 1Hb and a2Hb in tetrameric Hb (Fig. 3b, figure 12). This overlap explains why Hp only binds αβ dimers. Furthermore, Hb tetramer formation from a^Hb buries only 1980 A2 compared to 2954 A2 for the Hp-Hb interaction. Thus, the interaction with Hp will push the equilibrium between Hb tetramers and dimers further towards dimers. An extensive network of electrostatic interactions combined with van der Waals contacts forms the interface between Hp and Hb. Selected interactions are shown in figure 3c-f and a complete list of contacts in figure 13. This comprehensive set of interactions fully explains the tight binding of Hp and Hb.

Tetrameric Hb is well known to undergo conformational changes that are important for regulating gas transport. Several Hb residues suggested to be involved are directly recognised by Hp (e.g. aVall, aVal96, al_ys99, aTyrl40, aArgl41, βΤφ37, Tyrl45, Fig. 3d-g). Hp-bound a&Hb dimers exhibit non- cooperative oxygen binding with a reported P₅₀ (half-saturation oxygen tension) value of 0.3 mmHg. Engineered a^Hb dimers also lack cooperative oxygen binding with a reported P50 value of 0.59 mmHg, indicating that dimeric a^Hb has a high oxygen affinity compared to tetrameric Hb (P₅₀ value of 25-30 mmHg in blood), irrespective of whether it is bound to Hp or not. The interactions of Hp with the FG loops and C- terminal regions of both aHb and Hb probably preserve the conformation of the F helix and consequently the position of the proximal histidine with respect to the haem group.

This may explain the maintained high oxygen affinity. The high oxygen affinity implies that deoxygenated α βΗ b dimers are only present under extremely low oxygen tension. Deoxygenated Hb exhibits slow binding to Hp, which has been suggested to be due to low dissociation rate of deoxygenated Hb tetramers.

However, if deoxygenated a Hb retains its conformation after dissociation into a Hb dimers, several residues are not favourably positioned for interaction with Hp (figure 14).

Hp protects the vascular system from damage by free Hb, but does not alter the reactive properties of the Hb haem-group. The ability of Hb to oxidise lipids and undergo structural modifications most likely stems from radical intermediates formed on the globin moieties. Hb residues that are specifically prone to oxidative modifications by hydrogen peroxide in the absence of Hp are displayed on the structure of HpHb in Fig. 4a. Several of these residues are located in the interface between Hp and Hb suggesting that Hp shields the radicals formed on these residues. The Tyr42 residue of aHb may play a key role in radical migration from a lHb to 2Hb (or a2Hb to 1Hb). Hp most likely blocks radical migration by forcing dissociation of Hb tetramers. Other effects of hydrogen peroxide exposure are subunit dissociation, globin cross-linking and haem release. These effects are probably also prevented by the tight interaction of Hp with Hb. Furthermore, Hp binds close to the haem group and it may stabilise this region of the globin moiety, which in turn may prevent haem release. Dimeric human Hp is the gene product of the Hpl gene. However, humans have two Hp allels and the Hp2 gene gives rise to Hp multimers (figure 15). The multimerisation is caused by duplication of the CCP domain and the ability of each domain to dimerise with another CCP domain from an Hpl or Hp2 protein. Duplication of Cys33 involved in CCP linkage has been the common explanation for this. However, these multimers are most likely also stabilised by the formation of multiple CCP fusion domains (figure 15). Interestingly, β-strand swapping may also occur in Hp-related protein (Hpr) (figure 6), which is present in old world primates including humans where it medicates innate immunity against trypanosome parasites. Although Hpr lacks the cysteine residue at position 33 forming an interchain disulfide bridge in Hp, it associates into non-covalently linked dimers. Small angle x-ray scattering (SAXS) on solutions of dimeric human and porcine HpHb shows almost identical scattering curves (Fig. 4b). Furthermore, scattering curves and ab initio modelling (Fig. 4b+c, figure 16) of dimeric human HpHb incubated with a soluble recombinant CD163 fragment (SRCR domains 1-5) harbouring the ligand-binding site, indicate that the receptor binding site is located in the protruding Hp loop 3 area.

The data also shows that the receptor fragments can bind simultaneously to each of the HpHb entities. Such receptor cross-linkage may be important for efficient CD163-mediated uptake and it explains the increased receptor avidity of the multimeric HpHb complex (figure 17).

Methods summary

HpHb was purified from porcine blood by a three-step chromatographic procedure (anion exchange, hydrophobic interaction, and size exclusion) and crystallised using sitting-drop vapour diffusion against a reservoir containing 18% PEG3350, 10% jeffamine M-600 and 200 mM ammonium citrate pH 7.0.

Before freezing crystals were exchanged into cryoprotection buffer containing reservoir solution with 25% PEG3350. Diffraction data were collected at 100 K at SLS X06SA. The structure was determined by molecular replacement with porcine Hb (PDB accession 1QPW) and human Clr (PDB accession 2QY0) as search models. UV-vis spectra of HpHb or Hb in solution were measured on an Agilent 8453 Diode Array spectrophotometer and HpHb crystals or frozen solution were measured on a XSPECTRA micro spectrophotometer. Raman spectra were recorded at 100 K on a Jobin-Yvon Horiba T64000

instrument. Small angle x-ray scattering data were collected on a pinhole camera using a rotating anode x-ray source. Methods

Purification of HpHb from porcine blood

Anti-coagulant (trisodium citrate and EDTA) was added to fresh porcine blood to a final concentration of 15 mM (trisodium citrate) and 0.15 mM (EDTA). Plasma and blood cells were separated by centrifugation at 4,000 g for 20 minutes. Clotting factors were removed from the plasma fraction by addition of 25 mM BaCI2, followed by incubation on ice overnight and centrifugation at 8,000 g for 15 minutes. The blood cell fraction was lysed by addition of water (1 : 1 ratio) and cell debris was removed by centrifugation at 8,000 g for 15 minutes. Serum and blood cell fractions were stored at -80°C.

Thawed serum and blood cell fractions were mixed in a ratio of 25: 1 and incubated at 4°C overnight. The sample was diluted 1 : 5 in 20 mM HAc pH 5.3, 10% glycerol and loaded on a Q Sepharose Fast Flow column (GE Healthcare) equilibrated in buffer Q-A (20 mM KCI, 20 mM HAc pH 5.3, 10% glycerol). A gradient from 5 to 55% buffer Q-B (500 mM KCI, 20 mM HAc pH 5.3, 10% glycerol) was applied. The pH of the eluted fractions was adjusted by addition of Tris-HCI pH 7.6 to a final concentration of 50 mM. Source Q fractions containing HpHb were pooled and ammonium sulphate added to 60% saturation. The sample was centrifuged at 27,000 g for 20 minutes and the supernatant was loaded on a Source 15 Iso column (GE Healthcare) equilibrated in buffer Iso-A (60% ammonium sulphate, 20 mM Tris-HCI pH 7.6). A gradient from 0-60% buffer Iso-B (20 mM Tris-HCI pH 7.6) was applied and HpHb-containing fractions were pooled and concentrated using an Amicon Ultra centrifugal filter (10 kDa MWCO, Millipore).

The sample was further purified using a Superdex 200 column (GE Healthcare) equilibrated in 75 mM KCI, 20 mM Tris-HCI pH 7.6, 0.5 mM EDTA. Fractions containing >98% pure HpHb were pooled and concentrated to 10 mg/ml using a Vivaspin 500 centrifugal filter (10 kDa MWCO, GE Healthcare).

Crystallisation and data collection

Crystals were obtained at 4°C using the sitting-drop vapour diffusion method by mixing 2 μΙ protein solution (10 mg/ml) with 2 μΙ reservoir solution containing 18% PEG 3350, 10% jeffamine M-600 and 200 mM ammonium citrate pH 7.0. Prior to flash-freezing in liquid nitrogen, crystals were exchanged into cryo- protection buffer containing 25% PEG 3350, 10% jeffamine M-600 and 200 mM ammonium citrate pH 7.0. X-ray data were collected at the X06SA beamline (Swiss Light Source, Villigen, Switzerland) using a wavelength of 1.0 A and at a temperature of 100K.

Structure determination

Data were indexed, integrated, and scaled with the XDS-package30. Initial phases were calculated by molecular replacement using the program PHASER31. The structures of porcine Hb (PDB accession 1QPW) and human Clr (PDB accession 2QY0) were used as search models. The model was refined using iterative cycles of refinement in PHENIX32 followed by model building using the program Ό'33. Both B-factors and atomic coordinates were restrained by tight four-fold (Hb and Hp SP) or two-fold (Hp CCP) non-crystallographic symmetry with one B-factor group per residue throughout the refinement procedure. PISA34 and

PROCHECK35 were used for structure analysis and validation. The

Ramanchandran plot statistics shows 90.7% residues in the most favoured region, 9.0% in the additionally allowed region and 0.3% in the generously allowed region. Figures were prepared with PYMOL and ALINE.

UV-vis and Raman spectroscopy

The UV-vis spectra of porcine Hb and HpHb in solution at room temperature were measured in a 1 cm quartz cuvette on an Agilent 8453 Diode Array UV-vis

Spectrophotometer. The UV-vis spectra of porcine HpHb as frozen solution and as crystals were measured using a microspectrophotometer model XSPECTRA (4DX System AB, Uppsala, Sweden) equipped with a halogen lamp and iDUS CCD detector with a Shamrock monochromator (Andor Technology). These

measurements were performed at 100 K in a nitrogen cold stream (Oxford Cryosystems). The frozen solution sample was generated by mixing the HpHb solution with ~50% glycerol. The spectra of HpHb, both as frozen solution and as crystals, were recorded with the sample placed in a nylon loop (Hampton

Research). Raman spectra were recorded at 100 K on a Jobyn Yvon Horiba

T64000 instrument with the laser generated through a Millennia Pro 12sJS

Nd :YV04 a Matisse TR ring laser and a Wavetrain™ frequency doubler.

Small Angle X-ray Scattering :

SAXS data were collected on a pinhole camera using a rotating anode as x-ray source and Gobel mirrors as optics. Human Hp (Hpl/Hpl) was purchased from Sigma and CD163 SRCR 1-5 was expressed and purified as described previously. Samples were measured at 20°C in re-usable quartz capillaries with a diameter of 1.5 mm. The data are displayed as a function of the modulus of the scattering vector, q = (4 π/λ) sin(6, where λ = 1.54 A is the x-ray wavelength and 2Θ is the angle between the incident and scattered x-rays. Background subtraction and normalizations were made using the SUPERSAXS package (C. L. P. Oliveira and J.S. Pedersen, unpublished). The data were normalized to absolute scale using a pure water sample as primary standard. In all cases a concentration series (from 2 up to 8 mg/ml) were investigated to check for concentration effects. The initial analysis of the data was performed using the Indirect Fourier Transformation (IFT) procedure. Ab initio structure determination was performed using the programs DAMMIN and GASBOR. Modelling using known atomic resolution structures was done using the programs CRYSOL, BUNCH, SASREF and CORAL. For the set of generated models both the average model and the most probable model were determined using the program DAMAVER.

Example 2 - Expression and binding analysis of the multimerization module CCP fused to KN2 single chain fragment This experiment describes the design of a multimerization domain utilizing the CCP/Sushi domain found in human haptoglobin. The design includes the CCP domain found in Hpl-1 (CCP1) which is anticipated to form dimers and the CCP domain found in Hp2-2 (CCP2) which is anticipated to form trimers and higher multimers. CCP2 corresponding to aa 19-148 in Hp2-2 and CCP1 corresponding to aa 19-89 in Hpl-1 is N-terminally fused to the anti-human CD163 (KN2) scFv fragment.

To facilitate expression the constructs are designed with the cysteine involved in inter domain disulfides between CCP's mutated (C->A).

All constructs are being optimized for expression in Pichia Pastoris and cloned into the pJexpress912 vector (see figure 18). To obtain single yeast colonies expressing the KN2-CCP1 and KN2-CCP2 construct yeast cells (P. pastoris) were transformed with linearized plasmid and grown on zeocin (25 pg/ml) containing YPD plates. Single colonies were select and scaled for protein expression in shaking flask using appropriate growth media and methanol induction.

Following 72 h of incubation, obtained clones were screened for expression of the CCP constructs by western blotting. Using a HRP-conjugated anti-human IgG antibody positive expression were observed in four out of five clones for KN2- CCP1 while only one clone expressed KN2-CCP2 (see figure 19).

To analyse the effect of the CCP fusion partner on KN2 scFv structure and function, the fusion protein was purified from growth media of KN2-CCP1 clone#l using a simple ammonium sulphate (AMS) precipitation strategy followed by size- exclusion chromatography (SEC) (see figure 20).

The KN2-CCP1 fusion protein readily precipitated at 40% AMS and subsequent separation of the 40% AMS precipitate using SEC provided a more than 90% pure protein corresponding to the KN2-CCP1 fusion protein (see figure 21 and 23).

The binding of the purified KN2-CCP1 fusion protein to human CD163 was analyzed using surface plasmon resonance analysis. When compared to the monomeric KN2 scFv (Kd 96 nM) a stronger binding of the KN2-CCP1 fusion protein was observed (Kd 14 nM) (see figure 24). Expression of multimerized binding proteins or ligands would in theory provide fusion proteins with increased avidity resulting in a stronger interaction with the ligand or antigen. In the present experiments the difference in binding strength observed for non- fused KN2 scFv and for KN2-CCP1 was mostly due to a slower off-rate (kd) for KN2-CCP1, which in turn resulted in a stronger binding constant.

When translated to avidity, this increase in binding strength, suggest that expression of the KN2 scFv as a N-terminal fusion protein with the CCPl module results in a multimerized fusion protein with improved binding characteristics.

Example 3 - Multimerization of protein-binding domains using the CCP domain found in haptoglobin.

The following describes the design of a multimerization domain utilizing the CCP/Sushi domain found in human haptoglobin. The design includes the CCP domain found in Hpl-1 (Hp-CCPl) which is anticipated to form dimers and the CCP domain found in Hp2-2 (Hp-CCP2) which is anticipated to form trimers and higher multimers. CCP2 corresponding to aa 19-148 in Hp2-2 and CCPl

corresponding to aa 19-89 in Hpl-1 are available for N-terminal and/or C-terminal fusion to proteins of interest. To facilitate expression a cysteine involved in inter-domain disulfides bonding is mutated (C->A). All constructs are being optimized for expression in Pichia

Pastoris and cloned into the pPICZaA vector (Life Technologies).

To obtain single yeast colonies expressing the KN2-CCP1 and KN2-CCP2 construct yeast cells (P. pastoris) were transformed with linearized plasmid and grown on zeocin (25 pg/ml) containing YPD plates. Single colonies were select and scaled for protein expression in shaking flask using appropriate growth media and methanol induction. Following 72 h of incubation obtained clones were screened for expression of the CCP constructs by western blotting using rabbit polyclonal anti-human haptoglobin antibody and HRP-conjugated goat anti-rabbit IgG.

To clarify if expression of the CCP domain alone also facilitates multimerization as seen for full-length haptoglobin a construct consisting of Hp-CCP2 was

transformed into Pichia Pastoris and single expressing colony was selected

(Fig.25A). Expressed Hp-CCP2 was purified by hydrophobic interaction

chromatography (Fig.25B) and subjected to size exclusion chromatography. As expected for a protein with several multimeric states the purified protein eluted as a wide peak consisting of protein species with decreasing size when analyzed by nativePAGE (Fig.25C).

To analyze the effect of the CCP fusion partner on a protein-binding domain, the Hp-CCPl dimerization domain was C-terminally fused with the receptor-associated protein (RAP). The fusion protein was expressed in Pichia Pastoris as described and dimerization of the fusion protein was verified by haptoglobin specific immunoblotting of growth media separated by SDS-PAGE and nativePAGE.

Subsequently the protein was purified using anion-exchange chromatography and subjected to binding analysis using surface plasmon resonance analysis. When analyzing the binding of the monomeric RAP to immobilized cubilin (Kd 30 nM) and megalin (Kd 10 nM) a stronger binding of the Hp-CCPl-RAP fusion protein was observed (cubilin : Kd 7 nM, Megalin 3.5 nM). Expression of multimerized binding proteins or ligands would in theory provide fusion proteins with increased avidity resulting in a stronger interaction with the ligand or antigen. In the present example the difference in binding strength observed for non-fused RAP and for Hp-CCPl-RAP was due to changes in both the on-rate and off-rate which in turn resulted in a stronger binding constant for the dimeric RAP. When translated to avidity, the increase in binding strength, and the migration on nativePAGE shows that expression of the RAP as a C-terminal fusion protein with the CCP1 module results in a dimeric fusion protein with increased size and improved binding characteristics (see figure 26).

In conclusion our data shows that the CCP multimerization platform provides a method for large-scale recombinant expression of multimeric fusion proteins yeast and potentially also in bacteria. Multimerization is, as shown, an attractive method for obtaining increased avidity, sensitivity and size of protein binders including single chain antibody fragments, natural ligands and receptor fragments. The use of both the Hp-CCPl and Hp-CCP2 multimerization domains allows for the selection of the best suited multimeric structure ranging from dimeric and up while keeping the cost low do utilization of the cost-effective yeast or bacterial expression systems. Most pronounced, the multimerization platform allows for a cost-effective production of multimeric single chain variants with tailored avidity and plasma half-lifes of already well-established biologies such as TNF-a inhibitors.

Tables

Table 1 - Data collection and refinement statistics (Molecular replacement)

Porcine HpHb

Data collection

Space group Ρ2ι2_!2_!

Cell dimensions

a, b, c (A) 72.88, 197.78, 322.07

A, β, Y (°) 90,90,90

Resolution (A) 30-2.9(3.0-2.9) *

^sym Or /Emerge 7.5(63.5)

Completeness (%) 99.0(94.6)

Redundancy 4.36(4.27)

Refinement

Resolution (A) 30-2.9

No. reflections 103,264

^work/ ^free 21.1 / 22.9

No. atoms

Protein 18,508

Ligand/ion 652

B-factors (Chains A-F)

Protein 78.78

Ligand/ion 103.02

B-factors (Chains G-H)

Protein 105.82

Ligand/ion 136.81

R.m.s deviations

Bond lengths (A) 0.014

Bond angles (°) 1.63

Highest resolution shell is shown in parenthesis.

Claims

1. A recombinant hybrid polypeptide, said polypeptide comprising a first polypeptide amino acid sequence and a second heterologous amino acid sequence wherein said second amino acid sequence is a complement control protein domain (CCP), wherein said hybrid polypeptide is capable of forming a dimer or a multimer through intermolecular beta-strand swapping.

2. A recombinant hybrid polypeptide, said polypeptide comprising a first polypeptide amino acid sequence and a second heterologous amino acid sequence, wherein said second amino acid sequence is selected from the group consisting of a) an amino acid sequence selected from the group consisting of SEQ ID NO: 2 or

3. and b) an amino acid sequence having at least 75% sequence identity to a), and c) an amino acid sequence which is a sub-sequence of a) and b), where said subsequence is having a minimum length of 16 amino acids.

3. The recombinant hybrid polypeptide of claim 1-2, which are capable forming a dimer or a multimer through intermolecular beta-strand swapping without the involvement of disulphide bridges.

4. The recombinant hybrid polypeptide of claims 1-3, wherein the CCP domain is from haptoglobin or a haptoglobin related protein.

5. The recombinant hybrid polypeptide of claims 1-4, wherein the CCP originates from the group consisting of mouse, human, pig, dog, animal, rat, hamster, primate, and ape.

6. The recombinant hybrid polypeptide of claims 1-5, wherein the first polypeptide amino acid sequence originates from the group consisting of mouse, human, animal, rat, hamster, primate, and ape.

7. The recombinant hybrid polypeptide of claims 1-6, wherein the first polypeptide amino acid sequence is selected from the group consisting of a single-chain antibody, insulin, a cytokine or other kind of biological effector molecules.

8. The recombinant hybrid polypeptide of claims 1-7, wherein the first polypeptide and the second heterologous amino acid sequence is joined by a linker.

9. The recombinant hybrid polypeptide of claim 8, wherein the linker has a size selected from the group consisting of 1-20 amino acids.

10. A nucleic acid molecule comprising a nucleic acid sequence encoding the recombinant hybrid polypeptide of claims 1-9.

11. A vector comprising the nucleic acid molecule of claim 10.

12. A cell comprising the recombinant hybrid polypeptide of claims 1-9, the nucleic acid molecule of claim 10 or the vector of claim 11.

13. The cell according to claim 12 which is a bacterial or a yeast