Human serum albumin binding compounds and fusion proteins thereof

Abstract

The present invention relates to a polypeptide binding to human serum albumin and comprising or consisting of an amino acid sequence selected from the group consisting of: (a) GVTLFVALYDY(X.sup.1)(X.sup.2)(X.sup.3)(X.sup.4)(X.sup.5) (X.sup.6)D(X.sup.7)SFHKGEKFQIL(X.sup.8)(X.sup.9)(X.sup.10)(X.sup.11)(X.sup.12)G(X.sup.13)(X.sup.14)W(X.sup.15)(X.sup.16)RSLTTG(X.sup.17)(X.sup.18)G(X.sup.19)IPSNYVAPVDSIQ (SEQ ID NO: 1), wherein (X.sup.1) is A, V, I, L, M, G, P, S, T, N, Q, C, R, H, K, D or E; (X.sup.2) is R, H, K, A, V, I, L, M, G, P, S, T, N, Q or C; (X.sup.3) is R, H, K, S, T, N, Q, C, F, Y, W, A, V, I, L, M, G or P; (X.sup.4) is S, T, N, Q, C, A, V, I, L, M, G, P, R, H, K, F, Y, W, D or E; (X.sup.5) is S, T, N, Q, C, D, E, F, Y, W, A, V, I, L, M, G, P, R, H or K; (X.sup.6) is F, Y, W, A, V, I, L, M, G, P, R, H, K, S, T, N, Q or C; (X.sup.7) is A, V, I, L, M, G, P, R, H or K; (X.sup.8) is S, T, N, Q, C, D or E; (X.sup.9) is S, T, N, Q, C, D, E, A, V, I, L, M, G, P, F, Y or W; (X.sup.10) is A, V, I, L, M, G or P; (X.sup.11) is F, Y, W, R, H or K; (X.sup.12) is S, T, N, Q, C, F, Y or W; (X.sup.13) is F, Y, W, R, H, K, S, T, N, Q, C, D, E, A, V, I, L, M, G or P; (X.sup.14) is F, Y, W, A, V, I, L, M, G or P; (X.sup.15) is D, E, A, V, I, L, M, G or P; (X.sup.16) is A, V, I, L, M, G or P; (X.sup.17) is D, E, A, V, I, L, M, G, P, R, H or K; (X.sup.18) is S, T, N, Q, C, A, V, I, L, M, G or P; (X.sup.19) is F, Y, W, S, T, N, Q or C; and (b) an amino acid sequence which is at least 90% identical to the amino acid sequence of (a), wherein the identity determination excludes amino acid positions (X.sup.1) to (X.sup.19).

Claims

1. A polypeptide binding to human serum albumin, wherein the polypeptide comprises an amino acid sequence selected from the group consisting of: SEQ ID Nos: 4 to 40.

2. The polypeptide of claim 1, wherein the amino acid sequence is selected from the amino acid sequences of SEQ ID NOs: 4, 5 or 6 to 32.

3. A fusion protein comprising the polypeptide of any one of claims 1 or 2 fused to a pharmaceutically and/or diagnostically active protein or peptide.

4. The fusion protein of claim 3, wherein the polypeptide is directly fused to the pharmaceutically and/or diagnostically active protein or peptide.

5. The fusion protein of claim 3, wherein the polypeptide is fused to the pharmaceutically and/or diagnostically active protein or peptide via a linker.

6. The fusion protein of claim 3, wherein the pharmaceutically and/or diagnostically active protein or peptide is selected from the group consisting of a recombinant protein, an antibody, a blood factor, a hormone, an anticoagulans, a thrombolytic, a cytokine, a chemokine, and an interferon.

7. The fusion protein of claim 5, wherein the antibody is a bispecific T-cell engaging antibody.

8. A pharmaceutical and/or diagnostic composition comprising the fusion protein of claim 3.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) FIG. 1 shows the SDS-PAGE characterization of albumin-binding polypeptides of the invention: Lane M: molecular weight standard; Lane A: Fynomer® C1 (SEQ ID NO: 4); Lane B: Fynomer® 17H (SEQ ID NO: 5); Lane C: WT Fyn-SH3 (SEQ ID NO: 3).

(2) FIG. 2 shows the SDS-PAGE characterization of the unmodified BITE® polypeptide (SEQ ID NO: 53, lane 1) and the Fynomer®-BITE® fusion protein COVA406 (SEQ ID NO: 54, lane 2), consisting of the albumin binding Fynomer® 17H and the BITE® molecule. The molecular weight standard is shown in Lane M.

(3) FIG. 3 shows the size exclusion chromatogram (SEC) of COVA406 (SEQ ID NO: 54).

(4) FIG. 4 depicts a FACS binding experiment with COVA406 (SEQ ID NO: 54) using cells expressing CD3 (Jurkat E6-1), cells expressing PSMA (22Rv1 cells) and an irrelevant cell line expressing neither CD3 nor PSMA (LS174T cells). Binding is expressed as mean fluorescence intensity. CD3+: CD3-positive cells; CD3−: CD3-negative cells; PSMA+: PSMA-positive cells; PSMA−: PSMA-negative cells; COVA406: COVA406 is used as binding reagent; PBS: negative control, phosphate buffered saline is added instead of COVA406. COVA406 recognizes both antigens CD3 and PSMA expressed on cells.

(5) FIG. 5 depicts the analysis of redirected cell lysis of PSMA positive cells (22Rv1 cells) or PSMA negative cells (HT29 cells) by COVA406 (SEQ ID NO: 54) using human PBMCs as effector cells. The target cells 22Rv1 and HT29 were pre-labeled with Calcein AM and then incubated with human PBMCs (at effector cell to target cell ratio (E:T) of 25:1) and different concentrations of COVA406 (SEQ ID NO: 54) for 5 hours. The percentage of specific tumor cell lysis was measured by detection of calcein-release into the supernatant. Triplicates of 3 wells are shown ±SEM.

(6) FIG. 6 shows the serum concentrations of COVA406 (SEQ ID NO: 54) and the BITE® protein (SEQ ID NO: 53) at different time-points after a single i.v. injection into C57BL/6 mice. The concentration in serum was determined by ELISA. Mean values of 5 mice are shown ±SD

(7) FIG. 7 Specificity ELISA of prior art Fyn SH3 variants isolated after affinity selections. MSA=mouse serum albumin, HSA=human serum albumin, RSA=rat serum albumin, BSA=bovine serum albumin

(8) FIG. 8 Sequence alignment of SEQ ID NOs 4 to 40.

(9) The examples illustrate the invention.

EXAMPLE 1: FYN SH3 DERIVED POLYPEPTIDES BIND TO HUMAN SERUM ALBUMIN

(10) Methods

(11) 1) Lysate ELISA on Human Serum Albumin Protein

(12) Using the Fynomer® phage libraries described in Schlatter et al. (Schlatter et al. (2012) mAbs, 4(4) p. 497-50) Fyn-SH3 derived binding proteins specific to human serum albumin were isolated using human serum albumin (Sigma-Aldrich, cat. no A3782) and serum albumin from a rodent species (rat serum albumin, Sigma-Aldrich, cat. no A6414) as antigens and standard phage display as selection technology (Grabulovski D. et al., (2007) J Biol Chem 282, p. 3196-3204, Viti, F. et al. (2000) Methods Enzymol. 326, 480-505).

(13) After naïve and affinity maturation selections, enriched Fyn SH3-derived polypeptides were screened for binding to human serum albumin and/or serum albumin from a rodent species (mouse/rat) by lysate ELISA. DNA encoding the Fyn SH3-derived binding proteins was cloned into the bacterial expression vector pQE12 (Qiagen) so that the resulting constructs carried a C-terminal myc-hexahistidine tag as described in Grabulovski et al. (Grabulovski et al. (2007) JBC, 282, p. 3196-3204). The polypeptides were expressed in the cytosol of E. coli bacteria in a 96-well format and 200 μl of cleared lysate per well was prepared essentially as described in Bertschinger et al. (Bertschinger et al. (2007) Protein Eng Des Sel 20(2): p. 57-68). Briefly, transformed bacterial colonies were picked from agar plates and grown in a round bottom 96-well plate (Nunc, cat. no. 163320) in 200 μl 2×YT medium containing 100 μg/ml ampicillin and 0.1% (w/v) glucose. Protein expression was induced after growth for 3 h at 37° C. and 200 r.p.m. by adding 1 mM IPTG (Applichem, Germany). Proteins were expressed overnight in a rotary shaker (200 r.p.m., 30° C.). Subsequently, the 96-well plate was centrifuged at 1800 g for 10 min and the supernatant was discarded. The bacterial pellets were resuspended in 65 μl Bugbuster containing Benzonase Nuclease (VWR, cat. No. 70750-3) and incubated at RT for 30 minutes. Afterwards, the monoclonal bacterial lysates were cleared by centrifugation (1800 g for 10 min), diluted with 170 μL PBS and filtered using a multiscreen filter plate (0.45 μm pore size; Millipore cat. No. MSHVN4510). Monoclonal bacterial lysates were used for ELISA: human serum albumin was immobilized on maxisorp F96 wells (Nunc, cat. no 439454) overnight at room temperature. Plates were then blocked with PBS, 4% (w/v) milk (Rapilait, Migros, Switzerland). Subsequently, 20 μl of PBS, 10% milk containing 25 μg/ml anti-myc antibody 9E10 and 80 μl of bacterial lysate were applied (resulting in a final anti-myc antibody concentration of 5 μg/ml). After incubating for 1 h and washing, bound Fyn SH3-derived polypeptides were detected with anti-mouse-HRP antibody conjugate (Sigma) at a final concentration of 5 μg/ml. The detection of peroxidase activity was done by adding 100 μL per well BM blue POD substrate (Roche) and the reaction was stopped by adding 50 μl 1 M H.sub.2SO.sub.4. The DNA sequence of the specific binders was verified by DNA sequencing. Cross-reactivity towards serum albumin from a rodent species was detected by monoclonal lysate ELISA using mouse serum albumin (Sigma-Aldrich, cat. no A3139) as an antigen and the protocol described above. Alternatively, cross-reactivity towards mouse and rat serum albumin was confirmed surface plasmon resonance experiments (see below).

(14) 2) Expression and Purification of Fyn SH3-Derived Polypeptides in E. coli

(15) Fyn SH3-derived albumin-binding polypeptides were expressed in the cytosol of TG1 E. coli bacteria as well as purified as described in Grabulovski et al. (Grabulovski et al. (2007) JBC, 282, p. 3196-3204).

(16) 3) Affinity Measurements

(17) Affinity measurements were performed using a Biacore T200 instrument (GE Healthcare). For the interaction analysis between serum albumin, derived from mouse, rat or human, and Fyn SH3-derived albumin-binding polypeptides, a Series S CM5 chip (GE Healthcare) was used with albumin proteins immobilized using the amine coupling kit (GE healthcare). Serum albumin proteins from different species (mouse, rat or human) were immobilized (2000-3000 RU) on different flow cells of the chip whereas a blank-immobilized flow cell served as a reference flow cell. The running buffer was PBS containing 0.05% Tween 20 at pH 7.4. The interactions were measured at a flow of 30 μl/min and 25° C. and different concentrations of Fyn SH3-derived albumin-binding polypeptides were injected. All kinetic data of the interaction was evaluated using Biacore T200 evaluation software.

(18) Results

(19) 1) The amino acid sequences of ELISA positive Fyn SH3-derived polypeptides binding to human serum albumin is presented in SEQ ID NOs: 4 to 40 as appended in the sequence listing. In addition, Fyn SH3-derived polypeptides (SEQ ID NOs: 4 to 32) also showed binding to mouse serum albumin as confirmed by lysate ELISA and/or Biacore affinity measurements.

(20) 2) The expression yields of two selected Fyn SH3-derived albumin-binding polypeptides of the invention from bacterial cultures under non-optimized conditions in shake flasks is depicted in Table 1. The yield was in the same range as the expression yield of the WT Fyn-SH3 polypeptide. High protein-purity was confirmed by SDS-PAGE analysis and the gel is depicted in FIG. 1.

(21) TABLE-US-00002 TABLE 1 Expression yields of Fyn SH3-derived albumin-binding polypeptides produced in TG1 E. coli bacteria Fynomer ® SEQ ID NO. Yield (mg/l) 17H 5 10 C1 4 25 WT Fyn-SH3 3 10

(22) 3) The binding properties were analyzed by real-time interaction analysis on a Biacore chip revealing the following dissociation constants (K.sub.D) for selected albumin-binding polypeptides against albumin derived from either rat (RSA), mouse (MSA) or human (HSA) (depicted in Table 2).

(23) TABLE-US-00003 TABLE 2 Dissociation constants of Fyn SH3-derived serum albumin-binding polypeptides to RSA, MSA and HSA. SEQ ID K.sub.D (nM) K.sub.D (nM) K.sub.D (nM) Fynomer ® NO. RSA MSA HSA C1 4 72 408 1290 17H 5 17 96 455

EXAMPLE 2: ALBUMIN-BINDING FYN SH3 DERIVED POLYPEPTIDES HAVE A PROLONGED SERUM HALF-LIFE IN MICE

(24) Methods

(25) The pharmacokinetic profile of albumin-binding Fyn-SH3 derived polypeptides was investigated in BALB/c mice (Charles River) and compared to the WT Fyn-SH3 molecule. Fynomer® C1 (SEQ ID NO: 4), Fynomer® 17H (SEQ ID NO: 5) and WT Fyn-SH3 (SEQ ID NO: 3) were radiolabeled using Iodine-125 (Perkin Elmer cat no. NEZ033A001MC) and Chloramine T (Sigma-Aldrich cat NO 31224). The labeling reaction was carried out for two minutes at room temperature before removal of labeling reagents using PD MiniTrap G-25 columns (GE Healthcare cat. no 28-9180-07). Three BALB/c mice were injected i.v. with 13.5 μg of either radiolabeled Fynomer® C1 (SEQ ID NO: 4), Fynomer® 17H (SEQ ID NO: 5) or WT Fyn-SH3 (SEQ ID NO: 3). After 10 minutes, 2.5, 4, 6, 9, 25, 35 hours, blood was collected into EDTA coated microvettes (Sarstedt) and centrifuged for 10 min at 9300 g. Radioactivity was counted by mixing the serum with Supermix Perkin Elmer Scintillation Fluid and quantification of beta-emission of each sample with a 1450 MicroBeta Trilux scintillation counter and serum levels were calculated (results expressed as % injected dose (ID)/ml of blood). From the serum levels of Fynomer® C1, Fynomer® 17H and WT Fyn-SH3 determined in serum at different time points and the resulting slope k of the elimination phase (plotted in a semi-logarithmic scale), the half-lives were calculated using the formula t.sub.1/2=ln 2/−k.

(26) Results

(27) As depicted in Table 3, Fynomer® C1 (SEQ ID NO: 4) and Fynomer® 17H (SEQ ID NO: 5) show a significantly better terminal half-life as the WT Fyn-SH3 protein (SEQ ID NO: 3). Time-points used for half-life calculation: Fynomer® C1 and Fynomer® 17H: 2.5-35 h; WT Fyn-SH3: 2.5-25 h)

(28) TABLE-US-00004 TABLE 3 Terminal half-life of Fyn SH3-derived serum albumin-binding polypeptides in mice compared to the WT Fyn-SH3 protein. Fynomer ® SEQ ID NO: t.sub.1/2 (h) C1 4 10.5 17H 5 21.3 WT Fyn-SH3 3 4.4

EXAMPLE 3: ALBUMIN-BINDING FYN SH3 DERIVED POLYPEPTIDES CAN EXTEND SERUM HALF-LIFE OF BITE® Molecules

(29) Methods:

(30) 1) Expression and Purification of an Albumin Binding Fyn-SH3 Fusion Protein

(31) The Fynomer® 17H (Seq ID NO: 5) has been genetically fused to the N-terminus of the CD3-PSMA specific BITE® (Seq ID NO: 53) via a 15 amino acid linker (linker SEQ ID NO: 52) yielding the trispecific anti-albumin/PSMA/CD3 protein COVA406 (SEQ ID NO: 54). The BITE® protein (SEQ ID NO: 53) and the fusion molecule of the invention COVA406 (SEQ ID NO: 54) carrying a C-terminal penta-his-tag were transiently transfected into FreeStyle CHO—S cells and expressed in serum-free/animal component-free media for 6 days. The proteins were purified from the supernatants by Protein L affinity chromatography (Thermo Scientific, cat. No. 89928) with an ÄKTA Purifier instrument (GE Healthcare). Concentrations were determined by absorbance measurement at 280 nm. Yields are listed in Table 4. The SDS PAGE of both proteins is shown in FIG. 2.

(32) After purification size exclusion chromatography has been performed with COVA406 using an ÄKTA FPLC system and a Superdex G200, 30/100 GL column (GE Healthcare) (see FIG. 3).

(33) 2) FACS Binding Experiment with a BITE® Fusion Protein of the Invention

(34) The polypeptide COVA406 (SEQ ID NO: 54, final concentration 300 nM) was mixed with 100 μl cell suspension containing either (i) 1×10.sup.5 Jurkat E6-1 cells (CD3 positive cells), (ii) 1×10.sup.5 22 Rv1 prostate carcinoma cells (PSMA positive cells) or (iii) 1×10.sup.5 LS174T colorectal adenocarcinoma cells (PSMA and CD3 negative, ATCC cat. No. CL-188) in PBS/1% BSA/0.2% sodium azide. As a negative control, the same cells were incubated with PBS/1% BSA/0.2% sodium azide instead of COVA406 (PBS control). After 60 min incubation on ice, cells were washed, and bound protein was detected by incubation with 10 μg/ml mouse anti tetra-HIS antibody (Qiagen, cat no. 34670), followed by incubation with anti-mouse IgG-Alexa488 conjugate (Invitrogen) at a concentration of 10 ug/mL. Finally cells were washed three times and stained cells were then analyzed on a Guava EasyCyte™ (Millipore) flow cytometer.

(35) 3) Redirected T-Cell Mediated Cell Cytotoxicity Analysis

(36) The polypeptide COVA406 (SEQ ID NO: 54) was tested in a redirected T-cell mediated cell cytotoxicity assay using a protocol adapted from Dreier et. al. (2002) Int. J. Cancer: 100, 690-697.

(37) Human PBMCs were used as effector cells. On the day before the experiment PBMCs were isolated from fresh buffy coat preparations by Ficoll Plaque plus (GE Healthcare) and density gradient centrifugation using standard procedures. Isolated PBMCs were then incubated over night at a cell concentration of 4×10.sup.6 cells/ml in 10% FCS, RPMI and 37° C., 5% CO.sub.2.

(38) For the cell kill experiment PBMCs were centrifuged and resuspended in 10% FCS, RPMI at a cell concentration of 2.5×10.sup.7 cells/ml.

(39) Target cells were labeled with Calcein AM by incubating cells at a final Calcein AM concentration of 10 μM for 30 min at 37° C., 5% CO.sub.2. Subsequently excess dye was removed by washing cells twice with approx. 15 mL Medium. Finally target cell number was adjusted to 1*10.sup.6 cells/ml. Target tumor cells were either 22Rv1 cells (PSMA positive, ATCC cat. No. CRL-2505) or HT29 colon carcinoma cells (PSMA negative, DSMZ cat. No. ACC-299).

(40) Effector molecules were diluted in 10% FCS, RPMI to a maximum concentration of 1200 ng/mL. A dilution series of 1/10 dilutions was prepared.

(41) Finally target cell suspension, effector cell suspension and the different concentrations of the polypeptide COVA406 (SEQ ID NO: 54) were then mixed in equal amounts. A total of 50000 target cells were added per well and the effector to target cell ratio was 25/1, The final maximal concentration of effector molecules was 400 ng/μl. Cell lysis was measured after 5 hours incubation at 37° C. and 5% CO.sub.2. After incubation, the cell suspension was centrifuged and cell lysis was quantified by detection of Calcein AM fluorescence in the supernatant using a fluorescence reader.

(42) The amount of redirected cell lysis was normalized to the maximum lysis control (cells lysed by the addition of 1% Triton X-100) and spontaneous lysis (target cells incubated with PMBCs in the absence of effector molecules). Percentage of cell lysis was calculated according to the following formula:
% lysis=(((fluorescence sample)−(fluorescence spontaneous lysis control))/((fluorescence maximum lysis control)−(fluorescence spontaneous lysis control)))×100

(43) All measurements were done in triplicates. Specific cell lysis was plotted versus the concentration of COVA406 and evaluated using Prism 5 (GraphPad Software) by fitting a sigmoidal dose-response.

(44) 4) Comparison of the Pharmacokinetic Profiles of COVA406 and the BITE® Molecule

(45) The pharmacokinetic profile of COVA406 in C57BL/6 mice (Charles River) was investigated and compared to the parental BITE® molecule. Five C57BL/6 mice were injected i.v. each with 48 μg COVA406 (SEQ ID NO: 54) or BITE® (SEQ ID NO: 53). After 10 and 30 min, 1, 3, 5, 7, 9, 12, 24, 28, 33 and 48 hours, blood was collected into EDTA coated microvettes (Sarstedt), centrifuged for 10 min at 9300 g and the serum levels of COVA406 or BITE® were determined by ELISA. Briefly, black maxisorp microtiter plates (Nunc) were coated with 10 μg/ml of a peptide derived from CD3 (Sequence: QDGNEEMGGITQTPYKVSISGTTVILT; SEQ ID NO: 55) (expressed as Fc-fusion) and incubated over night at 4° C. After blocking with 4% milk (Rapilait, Migros, Switzerland) in PBS, serum samples at appropriate dilutions were applied, resulting in a final buffer concentration of 2% mouse serum (Sigma) and 4% milk. After incubation for 1 hr, wells were washed with PBS, and bound COVA406 or BITE® were detected with Penta-His-biotin (Qiagen) followed by Streptavidin-HRP conjugate (Sigma). The assay was developed with QuantaRed fluorogenic substrate (Pierce). The reaction was stopped after 3 min incubation and the fluorescence intensity was measured at 544 nm (excitation) and 590 nm (emission). The serum levels of COVA406 and BITE® were determined using a standard curve of COVA406 and BITE® (diluted to 333—0.5 ng/ml each). From the concentrations of COVA406 and BITE® determined in serum at different time points and the resulting slope k of the elimination phase (plotted in a semi-logarithmic scale), the half-lives were calculated using the formula t.sub.1/2=ln 2/−k. Timepoints used for half-life calculation: COVA406: 1-48 h; BITE®: 1-12 h.

(46) Result:

(47) COVA406 (SEQ ID NO: 54) expressed with a similar yield as the BITE® molecule (SEQ ID NO: 53) (Table 4).

(48) TABLE-US-00005 TABLE 4 Purification yields of the BITE ® and Fyn-SH3 derived albumin-binding polypeptide fusions produced in transiently transfected CHO-S cells. SEQ ID NO: Yield (mg/l) BITE ® 53 8.1 COVA406 54 5.0

(49) The size exclusion chromatography (SEC) profile after purification demonstrated that COVA406 eluted as a single, monomeric peak showing that the fusion protein has excellent biophysical properties (FIG. 3). Specific binding to PSMA-positive cells (22Rv1 cells) and CD3-positive cells (Jurkat E6-1, CD3 positive) was validated in a FACS experiment. Mean fluorescence intensities (MFI) of the stainings are depicted in FIG. 4. Redirected T-cell mediated cell cytotoxicity was validated in a Calcein release assay using PBMCs as effector cells. Specific redirected cell-lysis of PSMA-positive cells with COVA406 (EC.sub.50=4.35 ng/ml) is shown in FIG. 5. Cells with no PSMA expression (HT29 cells) were not lysed under the same conditions, showing that COVA406 is able to kill specifically PSMA positive cells. An improved pharmacokinetic profile of COVA406 (SEQ ID NO: 54) compared to the BITE® protein (SEQ ID NO: 53) was observed in mice. FIG. 6 shows the serum concentrations (ng/ml) and terminal elimination phase of COVA406 and the parental BITE®. COVA406 shows a significantly better half-life (14.3 hours) compared to the BITE® (1.5 hours). This example shows that serum albumin binding proteins of the invention are able to prolong the in vivo half-life of otherwise short-lived molecules, in particular of BITE® molecules.

EXAMPLE 4: PRIOR ART FYNOMERS® WHICH BIND TO SERUM ALBUMIN

(50) For Material and Methods, see Publications EP2054432 and “Grabulovski, Dragan: The SH3 domain of fyn kinase as a scaffold for the generation of new binding proteins. ETH Dissertation Nr 17216 (May 2007). www.dx.doi.org/10.3929/ethz-a-005407897”.

(51) FIG. 7 shows specificity ELISA of Fyn SH3 variants isolated after affinity selections. None of the Fynomers® binds to HSA or HSA/rodent serum albumin, except for C3. However, C3 cross-reacts also with the non-related ovalbumin (hen egg white albumin). Therefore, C3 is considered as an unspecific binding protein.