WO1991019741A1 - Albumin binding proteins - Google Patents

Abstract

The invention concerns albumin binding proteins DNA sequences coding for the proteins, vectors and hosts incorporating such DNA, a process for preparing the proteins and reagent kit and a pharmaceutical composition containing the proteins.

Description

Albumin binding proteins

The present invention concerns albumin binding proteins and peptides, recombinant plasmids containing DNA sequences that codes for such proteins and peptides. host cells transformed with such plasmids, process for preparing the proteins and peptides, reagent kit and pharmaceutical compositions comprising them.

Human serum albumin is an important protein which is given to patients with protein deficiency e.g. patients suffering from kidney or liver insufficiency or burn injuries. Albumin is present in blood in high concentrations and is often present as a contaminant when other substances are to be purified from blood or cells or tissues that have been in contact with blood. The proteins according to the invention are useful tools for binding albumin either to concentrate albumin or to remove contaminating albumin. Many drugs are transported in the body bound to albumin. By coupling a substance which does not have affinity for albumin to the proteins or peptides of the invention such a substance could be bound and transported by albumin. This would prolong the halftime for the substance.

By comparing the binding characteristics and the amino acid sequence of protein H and protein M type 1 it was suggested that the albumin binding is located to the C and D domains of these proteins.

Thus, the present invention comprises an albumin binding protein with the amino acid sequence of claim 1. subfragments or variants of the protein or peptides specifically disclosed in the present application wherein the original amino acid sequence is modified or altered by insertion, addition substitution, inversion or deletion of one or more amino acids are within the scope of the present invention as far as they retain the essential binding specificity as mentioned above. Examples of such variants are the amino acid sequences Cl to D inclusive of protein M1 in Fig. 8 i.e. amino acids number 29 to 281 inclusive in Fig. 8 and the amino acid sequence Cl to D of protein H in Fig. 8 i.e. amino acid number 159 to number 376 inclusive.

The invention also concerns DNA sequences, that codes for the proteins mentioned above according to claim 2 or coding for the Cl-D parts of protein M1 and H respectively in Fig. 8.

The gene coding for the albumin binding protein can be isolated from the chromosomal DNA of a protein H-producing strain such as the streptococcus sp group strain API based on the information on the DNA sequence of the protein H shown in Fig. 8 or from the plasmids p16 or p26 (see Fig. 4). The isolation of the gene can also be carried out as follows:

The chromosomal DNA can be isolated from cells of the protein H-producing strain in accordance with known methods

(Fahnestock, J. Bacteriol. 167, 870(1986). The isolated chromosomal DNA is then segmented into fragments of suitable lengths by biochemical means such as digestion with a restriction enzyme or physical means.

The resulting fragments are then inserted at an suitable restriction site into an suitable cloning vector such as λgt11 (Young et al., Proc. Natl. Acad.Sci. USA 80. 1194

(1983)) or a plasmid vector such as pUCls (Messing et al.. Gene

33. 103 (1985)).

Alternatively the polymerase chain reaction technique may be used to amplify the 3'-part of the protein H gene, using primers, the region of the protein H gene coding for the signal, the C and D region or albumin binding parts thereof can be selectively amplified and cloned. Plasmid vector pKK233-2 can be used. The recombinants are then incorporated into suitable host in this case E. coli.

From the resulting transformants, the clones producing the protein which binds to albumin, are selected by a known method (Fahnestock et al., supra)

After the proteins capable of binding to albumin, are purified from the resulting positive clones according to conventional methods, the binding specificities of the proteins are determined to select the clones producing a protein binding albumin only.

After plasmid DNA of said clone is isolated by conventional methods, the DNA sequence of the insert is determined by known methods (Sanger et al., Proc.Natl. Acad. sci. USA 7_4, 5463 (1977); Choen et al.. DNA 4. 165 (1985)).

The invention also relates to DNA sequences that hybridize with said identified DNA sequence under conventional conditions and that encode a protein binding to albumin. In this context the term "conventional conditions" refers to hybridization conditions where stringent hybridization conditions are preferred.

The expression of genes may be collected with vectors having the necessary expression control regions in which only the structural gene is inserted. For this purpose, the structural gene shown in Fig. 5 can preferably be used. The structural gene coding for the albumin binding protein can be obtained from the DNA sequence of Fig. 5 and 8 or synthesized by conventional methods based on the sequences given in the present specification.

As for the expression vectors, various host-vector systems have already been developed, from which the most suitable host-vector systems can be selected for the expression of the gene of the present invention. It has been known that, for each host cell, there is a particularly preferred codon usage for the expression of a given gene, in constructing a gene to be used for a given host-vector system, the codons preferred by the host should be used. Suitable sequences for the gene for the albumin binding protein to be used in a particular host-vector system can be designed based on the gene sequence given in Fig. 5 and 8 and synthesized by conventional synthetic methods.

The present invention further relates to a process for producing the albumin binding protein by culturing a host cell transformed with an expression vector into which the DNA encoding the albumin binding protein is inserted.

The process comprises steps of

I) inserting a DNA fragment coding for the albumin binding

peptide into a vector;

II) introducing the resulting vector into a suitable host cell;

III) culturing the resulting transformant cell to produce the albumin binding peptide ; and

IV) recovering the peptide from the culture. in the first step, the gene coding for the albumin binding protein, which is isolated from the streptococcal chromosomal DNA is inserted into a vector suitable for a host to be used for the expression of the corresponding protein . Insertion of the gene can be performed by digesting the vector with a suitable restriction enzyme and linking thereto the gene by a conventional method, such as ligation. in the second step, the resulting vector with the recombinant plasmid is introduced into host cells . The host cells may be Escherich coli. Bacillus subtilis or Saccharomyces cereviβiae or other cells. The introduction of the expression vector into the host cells can be performed in a conventional way and.

recombinants expressing IgG binding peptides are then selected.

In the third step, the resulting transformant cells are cultu red in a suitable medium to produce the albumin binding peptide by the expression of the DNA fragment. The cultivation can be conducted in a conventional manner.

In the fourth step, the produced albumin peptide protein is recovered from the culture and purified, which can be conducted by known methods. For example, the cells are disrupted by known methods such as ultrasonification, enzyme treatment or grinding. The albumin binding protein released by the cells or secreted into the medium is recovered and purified by conventional methods usually used in the field of biochemistry such as ion-exchange chromatography, gel filtration, affinity chromatography using albumin as ligand, hydrophobia chromatography or reversed phase chromatography, which may be used alone or in suitable combinations.

As mentioned above, the protein provided by the present invention can be used for binding identification or purification of albumin. It can also be bound to pharmaceuticals and used in sustained release formulations. For these purposes, the protein may be brought into a reagent kit or pharmaceutical composition by combining or mixing it with suitable reagents. additives or carriers.

As carriers there can be used Sepharose®activated with

CNBr, polyacrylamide beads and gold particles.

There can also be used other proteins as carriers. Such carrier proteins may be bound to or incorporate the proteins or peptides according to the invention. The M proteins such as M1, M3 and M6 and protein H mentioned above may be used as carrier protein incorporating the peptide or protein according to the invention. Such a carrier protein with the albumin binding sequence incorporated can then be bound to the carriers mentioned above.

As pharmaceutical additives there may be used the normally employed excipients, such as, for example, pharmaceutical grades of mannitol, lactose, starch, magnesium stearate, sodium saccharine, talcum, cellulose, glucose, gelatin, sucrose, magnesium carbonate and the like; diluents such as lactose, sucrose, dicalcium phosphate, and the like; a disintegrant such as starch or derivatives thereof; a lubricant such as magnesium stearate and the like; a binder such as starch, gum acacia, gum arabicum, polyvinylpyrrolidone, gelatin, cellulose and derivatives thereof, and the like may also be used.

The following figures are attached:

Figure 1: Dot binding to M1 protein (isolated from culture supernatants) applied to strips of nitrocellulose in different amounts. The strips were probed with HRPO-labelled proteins. Human serum albumin (HSA) , fibrinogen (Fbg), human polyclonal immunoglobulin G (IgG), a monoclonal human IgG of subclass 4 (lgG4) and Fc fragments of human IgG (IgG-Fc) all bound to M1 protein, whereas F(ab') fragments of IgG (IgG-F(ab') ), human polyclonal immunoglobulin M (IgM), vitronectin (Vn) and fibronectin (Fn) did not bind.

Figure 2: Dot binding to M3 protein (extracted from strain 4/55 streptococci with C-phage associated lysin) applied to strips of nitrocellulose in different eoncentrations. The strips were probed with HRPO-labelled proteins (for explanation of abbreviations see legend to Fig. 1) . The M3 protein was found to bind human serun albumin (HSA) and fibrinogen (Fbg) strongly, and showed weak affinity for human polyclonal

IgG(lgG) and fibronectin (Fn).

Figure 3: SDS-polyacrylamide electrophoresis (SDS-PAGE) in a 10% gel of 100 μg samples of extracellular M1 protein was followed by transfer of the separated protein samples to stripe of nitrocellulose. Strips were then probed with different

HRPO-labelled proteins. The two right lanes show (after staining with amidoblack) the separation of M1 protein and molecular weight markers on 10% SDS-PAGE.

The figure shows binding of albumin (HSA), fibrinogen (FbG). polyclonal IgG (IgG), monoclonal IgG's(IgGl - IgG4) and Fc fragments of IgG(lgG-Fe) to M1 protein.

Figure 4: Restriction maps of lambda EMBL3 clone 7:1 (a) and plasmid pK 18 clone p26 (b). Clone p16 is the indicated Sspl fragment cloned into pUCls.

Figure 5: Nucleic acid and protein sequence of the albuminbinding fragment of streptocoecal M protein type l encoded by pi6.

Figure 6: p16 encoded fragment of M1 protein and p26 encoded protein H were applied separately to strips of nitrocellulose. The strips were then incubated with HRPO-labelled human serum albumin (HSA), human fibrinogen or polyclonal human IgG (IgG) . The f igure shows binding of HSA to the f ragment of M1 and to protein H. Protein H also binds IgG.

Figure 7: Comparison of homologous regions in protein H, a fragment of M protein type 1 (p16) and M protein type 6 (M6). The amino acid sequence of encoded p16M1 was compared to M6 and protein H, and the homology is indicated as percentage identical amino acids. The three sequences were aligned to achieve maximal homology. All three peptides bind albumin, which maps the albumin-binding activity to the C and/or D regions.

Figure 8: Nucleic acid sequence of the region of p26 involved in encoding the albumin binding portion of M protein type l and protein H. The corresponding amino acid sequences are given.

Figure 9: Physical organization of the genes for p16M1 and protein H. The thin line represents the known DNA sequences. The open reading frames corresponding to plsM1 and protein H genes are indicated on this line. The division of the genes into subregions is indicated. The base coordinates are under the line whereas the figures above the line are the amino acid coordinates. The sequences corresponding to the insert DNA in P16 and pPH-1 are represented, respectively, by the two horizontal broken lines. Figure 10: schematic representation of protein H and two subfragments and the ability of the corresponding polypeptides to bind albumin. Amino acid residue numbers are indicated and so are the subregions of the peptides.

Figure 11: Schematic representation of the cloned protein H gene. Indicated are S, signal sequence; A and B unique regions; C1, C2, C3 repeated sequence; D, D region; w, wall spanning region; M membrane spanning region; T, charged C-terminal residues. Horizontal arrows labelled I, II, III, iv and V indicate the position of the primers.

Figure 12: Physicalmap of the expression vector pKK233-2

The present invention will more precisely be described by the following examples. But, they are not intended to limit the scope of the present invention.

Example 1.

isolation of albumin-binding proteins from Streptococcus pyoqenes.

Strains of streptococcus pyogenes M type 1 (40/58) and M 3 (4/55) of the strain collection of the WHO streptococcal

Reference Laboratory, Prague, Czechoslovakia were grown in 5 1 of Todd Hewitt Broth (Difco) by using a 7 1 fermentor. The culture was inoculated with 500 ml of an over night stand culture and fermenation was performed for 18 hours at 37°C at pH 7.2 and with regulation of the glucose concentration at 0.2%. The culture supernatants and the cell pellets were saved separately and used for purification of albumin-binding proteins.

1.

Preparation of albumin-binding proteins from culture

supernatants

Proteins in the culture supernatants were precipitated with ammonium sulphate at 70% saturation. The precipitate was dialyzed in the presence of 50 mM phenylmethylsulfonyl fluoride against distilled water and finally against 50 mM sodium acetate, pH 4.5. This crude preparation was applied to 200 ml of CM-Sepharose C1 4B equilibrated with 50 mM Na-acetate, pH 4.5, and washed with the same buffer. The column was then rinsed with 200 mM sodium phosphate buffer, pH 6.0. Fractions showing binding activity to fibrinogen and human serum albumin (HSA) in dot binding to nitrocellulose filter were collected and directly applied to affinity chromatography on fibrinogen-sepharose (50 ml). The column was washed with PBS and bound proteins were eluted with 2 M KSCN buffer. Fractions which were positive in dot binding on nitrocellulose with fibrinogen and albumin were collected, dialyzed against 50 mM ammonium hydrogen carbonate and applied to albumin-Sepharose (20 ml). The column was washed and eluted as described above and fractions eluted with 2 M KSCN buffer, which showed fibrinogenand albumin-binding were collected and dialyzed against 50 mM ammonium hydrogen carbonate. The albuminand fibrinogen-binding material was stored in alliquots or lyophized.

2.

Preparation of albumin-binding protein from cell extracts obtained with C-phage associated lysin.

Streptococcal cell pellets of the two strains (about 60 g of bacteria wet weight) were each suspended in 300 ml PBS containing 10 mg Penicillin and l ml 2-mercaptoethanol. Then 10 ml of phage lysin prepared according to Cohen et al. (Applied Microbiol. 29, 175-178, 1975) was added and cells incubated for 6 hours at 37 c with stirring. The resulting lysate was centrifuged and the pellet washed with 100 ml of PBS. The resulting supernatants were pooled, dialyzed against water containing 50 mM sodium acetate, pH 4.5, and fractionated on CM-Sepharose as described above. Albumin- and fibrinogen-binding fractions were further purified on fibrinogenand albumin-Sepharose, also as described above. Example 2

Characterization of the isolated albumin-binding streptococcal proteins

Streptococcal M proteins (for a review on M protein see

Fischetti 1989. Clin. Microbiol. Rev. 2,285) are known to bind fibrinogen and the albumin-binding proteins from both streptococcal strains (M protein type l and type 3) did indeed show affinity for both fibrinogen and albumin (purified first on fibrinogen-Sepharose and then on albumin-Sepharose). Further analysis of the proteins demonstrated serological identity with M1 and M3 protein, respectively. Thus, it was concluded that the albumin-binding proteins of the two strains represented M1 and M3 proteins.

On SDS-polyacrylamide electrophoresis (SDS-PAGE) the molecular weight of M1 was 49 kDa, whereas the M3 protein was more size heterogeneous with several bands around 60 kDa. In Figs. 1 and 2, M1 and M3 proteins were applied to nitrocellulose filters and probed with some. human plasma proteins labelled with horseradish peroxidase (HRPO) as described by Tijssen (In: Laboratory techniques in Biochemistry and Molecular Biology. Eds.

Burdon and van Knippenberg. Elsevier, Amsterdam-New York-Oxford 1985).Both M proteins bound albumin. Both M proteins bound albumin. HRPO-labelled M1 and M3 proteins were also both found to react with albumin applied to nitrocellulose filters. As shown in figure 1, M1 protein also showed affinity for HRPO-labelled fibrinogen and immunoglobulin G (IgG). In the case of IgG the interaction was mediated through the Fc region. M3 protein showed a much weaker affinity for IgG but reacted strongly with HRPO-labelled albumin and fibrinogen (figure 2). The binding properties of M1 protein was also studied by Western blot analysis. As shown in Fig. 3, M1 separated by SDS-PAGE and blotted onto nitrocellulose filters, reacted with albumin, fibrinogen and IgG.

Example 3

Identification and characterization of the albumin-binding part of M proteins and protein H. A genomic library of the streptococcus M type 1 strain 40/58 was made in the lambda replacement vector λ EMBL3. The DNA was partially cleaved with the restriction enzyme Sau3A and ligated, without size fractionation. to vector DNA cleaved with Bam HI. Recombinants expressing M protein type 1 (M1) were detected using specific anti-M1 serum. Three reacting clones were selected for further study. Fig. 4 is a physical map on one of these clones, named λ 7:1. A.4.4 kb Xbal fragment of ω 7:1 was ligated to the plasmid vector pK18, and the ligation mix used to transform E.coli TBl. Transformants harbouring the correct plasmid were initially selected using plasmid screening and confirmed by restriction enzyme digest analysis. One strain carrying the correct plasmid, named p26 was selected for further study. p26 (Fig. 4b) was shown to contain the entire gene sequence for protein H an IgG binding protein of streptococcus. and a truncated M1 gene sequence (P. Akesson, J. Cooney, F. Kishimoto and L. Björck. 1990. Protein H - a novel IgG-binding bacterial protein. Molec. Immunol. 27 523-531; H. Gomi, T. Hozumi, s. Hattori, c. Tagawa, F. Kishimoto and Lδ Björck. 1990. The gene sequence and some proterties of protein H - a novel IgG-binding protein. J. Immunol. 144, 4046-4052). p16 was generated by subcloning the 1.6 kb Sβp I frag- ment of p26. This fragment is indicated in Fig. 4b, and is upstream of the protein H gene which is located on the smaller Sβp I fragment. The entire insert of p16 was sequenced and within this sequence an open reading frame corresponding to 281 amino acids was identified (Fig. 5).

Lysates of E. coli harbouring either p26 or p16 were applied to IgG-Sepharose and albumin-Sepharose, respectively. Proteins bound were eluted with glycine buffer pH2 .0 and 0.15 to 0.63 mg of IgG or albumin binding proteins were isolated per 1 of culture of the p26 and p16 clones, respectively. The N-terminal amino acid sequences of p26 encoded protein H and p16 encoded M1 were determined (10 amino acids each) and found to be fully compatible with the gene sequences of protein H (Gomi et

al.1990, supra ) and the truncated M1 sequence shown in Fig . 5 . Eluted proteins were also analyzed in binding experiments. As expected protein H (from the p26 lysate) reacted with HRPO- labelled human IgG. However, also HRPO-labelled albumin reacted with this IgG-binding protein (Fig. 6), whereas the protein did not react with albumin when it was labelled with 125-1 using lactoperoxidase or chloramin-T, suggesting that this labelling procedure destroys the albumin-binding activity of protein H. The albumin-binding property of p26 encoded protein H was further demonstrated when the protein was coupled to sepharose. Thus, p26 encoded protein H coupled to Sepharose bound albumin (and IgG) very effectively. Finally, protein H did not bind fibrinogen (Fig. 6). The fragment of M1 protein expressed by p16, also showed affinity for albumin. Fibrinogen-binding was lost in the p16 encoded M1 fragment as compared to entire M1 (figure 6). Also IgG-binding seemed to be lost, but additional experiments showed that in solution p16 M1 could block the binding of entire M1 protein to IgG, suggesting that part of the IgG-binding site is present on the p16 M1 protein fragment. To summarize: protein H and M proteins (exemplified with M type 1 and 3 proteins) all show affinity for albumin and IgG whereas M proteins, but not protein H, also bind fibrinogen.

The amino acid sequence of p16 M1 was then compared to the sequences of protein H (Gomi et al. 1990, supra) and M protein type 6 (Hollingshead et al. 1986. J. Biol. Chem. 261 1677).

As shown in Fig. 7 there is a striking homology between the p16 encoded M1 sequence and the C-terminal sequences of protein H and M6, indicating that the albumin-binding region corresponds to the C and D domains of these proteins. Thus, any protein sequence homoloqous to the sequence encoded by p16 is likely to bind albumin. It is also interesting to note that M1 and M3 proteins bind IgG. Fibrinogen-binding is a well-known property of M proteins which was also demonstrated for the M1 and M3 protein used in this study, whereas protein H did not bind fibrinogen.

CONCLUSIONS:

1. Proteins showing homology with the p16 sequence of Fig. 5 will most probably bind albumin. In the present work this was exemplified with the binding of albumin to M proteins 1 and 3, a fragment of M1 corresponding to the p16 sequence of Fig. 5, and protein H.

2. M proteins type 1 and 3 were also found to bind IgG and at least in the case of M1 this binding was mediated through the Fc region.

Example 4

Production of an albumin binding protein which binds albumin only.

Association of albumin binding of protein H with the C repeat region and D region is inferred by the homology of this sequence with the C-terminal region of M1 protein, which was demonstrated to bind albumin. By deletion of the A-B part of the protein H gene it was possible to express a peptide which binds albumin but not IgG.

Using polymerase chain reaction (PCR) the 3' part of protein H gene was amplified. The PCR technique is an in vitro method in which genomic or cloned target sequences are specifically enzymatically amplified as directed by a pair of oligonucleotide primers. (J.F. Williams. 1989. Optimization strategies for the polymerase chain reaction. Biotechniques 7 : 762-768). Thus, the region of the protein H gene coding for the C and D regions can be selectively amplified and. cloned.

To generate a form of protein H which binds only albumin primers which amplify the 3' section of the gene were employed. The map position of these is indicated in Fig. 11. The sequence of these is tabulated below.

Primer IV

AAC CAT GGA AGC TAG TCG TAA GAG CCT AAG CCG TGA CC Primer V

GAG TAA GCT TGT TTG TGA CCT CTOCTT AAC CTC

Primer IV includes an Nco I site which generates an in frame ATG start codon. Primer V incorporates a Hind III site and will allow amplification of DNA to a position just downstream of the presumed natural termination structure. PCR was performed as follows:

0.25 ng of p26 plasmid DNA was used as the target for the PCR. The primers were added to a final concentration of 0.1 μM by dilution from a l -μM stock. The reaction buffer contained 250μm dATP, dCTP. dGTP and dTTP in 10 mM Tris-HCl pH 8.3, 50 mM KCl, 1.5 mM MgCl 0.001% gelantine. 100 y.1 of mineral oil was layered over the reaction mix. The DNA was denatured by heating to 98°C and then cooled to 53°C after which 2 units of Tag polymerase enzyme was added.

30 cycles of the following steps were performed.

Step 1:

1 min at 92°C.

Step 2 :

1 min at 53 c.

Step 3:

2 min at 72 C.

A Cetus intellegent heating block, set on mode 4, was used to control the temperature changes.

After completion of the 30 cycles the sample was cooled to room temperature. A 10 μl aliquot of the reaction mixture was electrophoresed on a 3% agarose gel. As there was only one DNA product of this reaction the remaining sample could be used directly for further steps without size fractionataion.

The amplified fragment can be cloned directionally as an Ncol - Hind III fragment into an expression vector, such as pKK233-2 (Amann and J Brosius. 1985. "ATG" vectors for the regulated high level expression of cloned genes in E. coli Gene 40 :

183-190). This vector has a composite trp-lac promoter which is inducable by IPTG, and an Nco I cloning site 7 base pairs downstream from a ribosome binding site (RBS). As this fragment of DNA could not include its own RBS it was necessary to supply such sequences in the vector.

The Hind III site further downstream facilitated the cloning of the amplified DNA in the correct orientation with respect to the promoter and RBS.

The reaction mix was deproteinized by extraction with phenol saturated with 10 mM Tris-HCl pH 8.0, 1 mM EDTA. The extracted material was treated with cholorform-isoamylalcohol (mixed in a 24:1 ratio and saturated with water) to remove excess phenol. The DNA was then precipitated by addition of a 1/10 volume of 8 M LiCl and 2.5 volumes of ice-cold ethanol.

The recovered DNA was dried under vacuum in a Speedy-Vac and resuspended in water to a final concentration of 200 ng/μi 400 ng of product was ligated to 200 ng of pKC30 (M. Rosenberg, Y.S. Ho and A. Shatzman. 1989). The use of pKC30 and its deriva- tives for controlled expression of genes. pp, 429-444 In R. Wu, L. Grossman and K. Moldave (eds.). Recombinant DNA methodology. Academic Press new York) which had been previously cleaved with Hpal deproteinized and precipitated as described for the PCR product. The molar ratio of insert to vector was approx. 5:1. The ligation was performed using T4 ligase (Gibro BRL) and the buffering system supplied in accordance with the manufacturer's instructions.

After ligation the DNA was transformed into E. coli JM109 (a strain which carries multiple copies of lac I^q).

Colony hybridization experiments (Grundstein and Hogness, 1975) were performed on the resulting transformants using the generated PCR fragment as a probe. 10 colonies which appeared to hybridize were selected and plasmid DNA prepared from these according to the method of Birnboim and Doly (1979). The DNA was further purified by treatment with RNase, deproteinized and precipitated. Further hybridization studies and restriction enzyme digest analysis of the plasmids confirmed the presence of the desired DNA fragment. The correct orientation with respect to the PI promoter was established for several of the plasmids.

The presence of the correct fragment was confirmed by the ability of the purified plasmids to hybridize in southern blots to digoxiging labelled amplified fragment, and by digestion of the plasmids with the restriction enzymes Tag I - Real in a double digest.

Selected strains were grown at 37°c to A of 0.6 in LB

broth containing ampicillin (100 mg/ml). Expression was then induced by addition of IPTG to a final concentration of l mM. After 6 hrs further incubation at 37 C, the cells were harvested and the polypeptide, with 215 amino acids and amolecular weight of 23.463, was purified from cell lysates by affinity chromatograph or Sepharose coupled with human serum albumin.

Bound protein was eluted with clycine buffer pH 2.0 and a single band of the expected size was obtained. The purified peptide was radiolabelled with 125-1 using the Bolton-Hunter reagent and was found to bind to albumin both in solution and to albumin-Sepharose. The peptyde did not bind to IgG-Sepharose,

Claims

C L A I M S

1. An albumin binding protein with the following anino acid sequence and subfragments and variants thereof binding to albumin: v v v v

MetThrTrpLysIleValThrAspSerGlyCysAspGlnLeuSerSerGluLysGluGln 20 v v v v

LeuThrlleGluLysAlaLysLeuGluGluGluLysGlnlleSerAspAlaSerArgGln 40 v v v v

SerLeuArgArgAspLeuAspAlaSerArgGluAlaLysLysGlnValGluLysAspLeu 60 v v v v

AlaAsnLeuThrAlaGluLeuAspLysValLysGluAspLysGlnlleSerAspAlaSer 80 v v v v

ArgGlnArgLeuArgArgAspLeuAspAlaSerArgGluAlaLysLysGlnValGluLys 100 v v v v

AspLeuAlaAsnLeuThrAlaGluLeuAspLysvalLysGluGluLysGlnlleSerAsp 120 v v v v

AlaSerArgGlnArgLeuArgArgAspLeuAspAlaSerArgGluAlaLysLysGlnVal 140 v v v v

GluLysAlaLeuGluGluAlaAsnSerLysLeuAlaAlaLeuGluLysLeuAsnLysGlu 160 v v v v

LeuGluGluSerLysLysLeuThrGluLysGluLysAlaGluLeuGlnAlaLysLeuGlu 180 v v v v

AlaGluAlaLysAlaLeuLysGluGlnLeuAlaLysGlnAlaGluGluLeuAlaLysLeu 200 v v v v

ArgAlaGlyLysAlaSerAspSerGlnThrProAspThrLysProGlyAsnLysAlaVal 220 v v v v

ProGlyLysGlyGlnAlaProGlnAlaGlyThrLysProAsnGlnAsnLysAlaProMet 240 v v v v

LysGluThrLysArgGlnLeuProSerThrGlyGluThrAlaAsnProPhePheThrAla 260 v v v v

AlaArgValThrValMetAlaThrAlaGlyValAlaAlaValValLysArgLysGluGlu 230

Asn 281

2. A DNA sequence coding for the protein according to Claim 1:

v v v v v v

ATGACTTGGAAAATTGTAACAGATTCAGGATGTGATCAACTTTCATCTGAAAAAGAGCAG 60 v v v v v v

CTAACGATCGAAAAAGCAAAACTTGAGGAAGAAAAACAAATCTCAGACGCAAGTCGTCAA 120 v v v v v v

AGCCTTCGTCGTGACTTGGACGCATCACGTGAAGCTAAGAAACAGGTTGAAAAAGATTTA 180 v v v v v v

GCAAACTTGACTGCΓGAACTTGATAAGGTTAAAGAAGACAAACAAATCTCAGACGCAAGC 240 v v v v v v

CGTCAACGGCTTCGCCGTGACTTGGACGCATCACGTGAAGCTAAGAAACAGGTTGAAAAA 300 v v v v v v

GATTTAGCAAACTTGACTGCTGAACTTGATAAGGTTAAAGAAGAAAAACAAATCTCAGAC 360 v v v v v v

GCAAGCCGTCAACGGCTTCGCCGTGACTTGGACGCATCACGTGAAGCTAAGAAACAAGTT 420 v v v v v v

GAAAAAGCTTTAGAAGAAGCAAACAGCAAATTAGCTGCTCTTGAAAAACTTAACAAAGAG 480 v v v v v v

CTTGAAGAAAGCTVAGAAATTAAJAGAAAAAAGAAAAAGCrGAACTACAAGCAAAACTTGAA 540 v v v v v v

GCAGAAGC.AAAAGCACCαϋ^GAACAATTAGCGAAACAAGCTGAAGAACTCGCAAAACTA 600 v v v v v v

AGAGCTGGAAAAGCATCAGACTCACAAACCCCTGATACAAAACCAGGAAACAAAGCTGTT 660 v v v v v v

CCAGGTAAAGGTCAAGCACCACAAGCAGGTACAAAACCTAACCAAAACAAAGCACCAATG 720 v v v v v v

AAGGAAACTAAGAGACAGTTACCATCAACAGGTGAAACAGCTAACCCATTCTTCACAGCG 780 v v v v v v

GCACGCGTTACTGTTATGGCAACAGCTGGAGTAGCAGCAGTTGTAAAACGCAAAGAAGAA 840

AACTAA 846

3. A DNA sequence hybridizing to a DNA sequence of Claim 2 under conventional conditions and encoding a protein displaying the same binding properties as the proteins or peptides of Claim l.

4. A recombinant plasmid containing a DNA sequence of Claim 3.

5. A host cell transformed with the recombinant plasmid of Claim 4.

6. A host cell according to Claim 5 which belongs to the species Escherichia coli.

7. A process for producing an albumin binding protein or peptide comprising cultivating a host cell according to Claims 5 or 6 under suitable conditions, accumulating the protein or peptide in the culture or lysing the host and recovering it therefrom.

8. A reagent kit for binding, separation and identification of albumin, chararcterized in that it comprises a protein or peptide according to Claim l.

9. A composition comprising a protein or peptide according to claim l and optionally additives or carriers.

10. A pharmaceutical composition comprising a protein or peptide according to Claim l and optionally pharmaceutically acceptable addiitives or carriers.