Streptococcal protein G. Gene structure and protein binding properties

Protein G was solubilized from 31 human group C and G streptococcal strains with the muralytic enzyme mutanolysin. As judged by the mobility in sodium dodecyl sulfate-polyacrylamide gel electrophoresis and the binding patterns of the solubilized protein G molecules in Western blot experiments, the strains could be divided into three groups, represented by the group G streptococcal strains G148 and G43 and the group C streptococcal strain C40. The 65-kDa G148 protein G and the 58-kDa C40 protein G showed affinity for both immunoglobulin G (IgG) and human serum albumin (HSA), whereas the 40-kDa G43 protein G bound only IgG. Despite the different molecular patterns, the three protein G species had identical NH2-terminal amino acid sequences. Apart from the 65-kDa peptide, digestion of G148 streptococci with mutanolysin also produced a 52-kDa IgG- and HSA-binding peptide and a 14-kDa HSA-binding peptide. It was demonstrated that these peptides resulted from cleavage of 65-kDa protein G by proteolytic components in the mutanolysin preparation. The protein G genes of the C40 and G43 strains were cloned and sequenced, and their structure was compared to the previously published sequence of the G148 protein G gene. As compared to G148, both the C40 and G43 genes lacked a 210-base pair fragment in the IgG-binding region, accounting for the 10-fold lower affinity of these proteins for IgG. The G43 gene also lacked a 450-base pair fragment in the 5'-end of the gene, explaining why the G43 protein G did not bind HSA. The differences in protein G structure did not correlate with the clinical origin of the strains used in this study. The IgG-binding region of protein G was further mapped. Thus, a peptide corresponding to a single IgG-binding unit was obtained by the cloning and expression of a 303-base pair polymerase chain reaction-generated DNA fragment. The affinity of this 11.5-kDa peptide for human IgG was 8.0 x 10(7) M-1, as determined by Scatchard plots. Finally, a 55-amino acid-long synthetic peptide, corresponding to one of the three repeated domains in the COOH-terminal half of strain G148 protein G, effectively blocked binding of protein G to IgG.     

Analysis and use of the serum albumin binding domains of streptococcal protein G

Streptococcal protein G is an IgG-binding receptor with a molecular weight of 63 kDa as predicted from the sequence of the corresponding gene. Here we show that a truncated recombinant protein of 23 kDa still has IgG-binding capacity and also interacts specifically with human serum albumin (HSA). This demonstrates that protein G is a bifunctional receptor. To investigate the structures needed for IgG- and albumin-binding, different parts of the receptor molecule were produced in E. coli using a coupled expression/secretion system. Affinity chromatography, using IgG or HSA immobilized on Sepharose, showed that the two binding activities are structurally separated. From these experiments, it was concluded that a region of 64 amino acid residues is sufficient for albumin-binding. The structure of this part of the protein suggests either a divalent or a trivalent binding capacity. The specific interaction to albumin was used to purify a heterologous protein by affinity chromatography to yield a pure fusion protein in a one-step procedure. The implication of this novel affinity system as a tool to facilitate protein immobilization and purification is discussed.     

Identification,characterization, and cloning of an immunoglobulin degrading enzyme in the cat flea,Ctenocephalides felis

The degradation of cat immunoglobulin G (IgG) in blood-fed adult C. felis midguts was examined. SDS-PAGE analysis of dissected midgut extracts obtained from C. felis that had been blood fed for various times between 0 to 44 h revealed that by 24 h most of the high molecular weight proteins, including the heavy chain of IgG, were digested. A 31-kDa serine protease with IgG degrading activity was purified from fed C. felis midguts by benzamidine affinity chromatography, hydrophobic interaction chromatography, and cation exchange chromatography. Three primary cleavage products between 30- and 40-kDa were observed when the purified protease was incubated with protein A purified cat IgG. N-terminal amino acid sequence analysis of the products revealed that the IgG degrading protease cleaves after specific cysteine and lysine residues within the hinge region of IgG. The enzyme is also capable of degrading other immunoglobulins, serum albumin, and hemoglobin, suggesting that it may have roles in both combating the host's immune system and providing nutrients for the flea. A cDNA clone encoding the 265 amino acid IgG degrading protease proenzyme was isolated. When expressed in a baculovirus/insect cell expression system, the recombinant protein had the same N-terminus as the processed 237 amino acid mature native protein and possessed IgG degrading activity indistinguishable from the native protein. Arch. Insect Biochem.     

Purification and characterization of a 52-kilodalton immunoglobulin G-binding protein from Streptococcus suis capsular type 2.

We previously reported that group D streptococci exhibited immunoglobulin G (IgG)-binding activity and that a 52-kDa IgG-binding protein was present in all Streptococcus suis strains examined (B. Serhir, R. Higgins, B. Foiry, and M. Jacques, J. Gen. Microbiol. 139:2953-2958, 1993). The objective of the present study was to purify and characterize this protein. Pig IgG were immobilized through their Fab fragments to ECH-Sepharose 4B, and the protein was purified by affinity chromatography. Electron microscopy observations of the purified material showed filamentous structures with a diameter of approximately 4 nm; these structures were not observed when the material was treated with either urea or ethanolamine. Electrophoretic and Western immunoblot analyses showed that the 52-kDa protein constituted the bulk of the recovered material. This protein was stained with either Coomassie brilliant blue or silver nitrate; it reacted with a large variety of mammalian IgG, human IgG (Fc) fragments, human IgA, and other human plasma proteins. The 52-kDa protein exhibited lower IgG-binding affinities than protein A and protein G. However, it was able to compete with protein A and protein G for binding to human IgG. In addition, it bound chicken IgG with high affinity. This last property differentiated the 52-kDa protein of S. suis from the six IgG-binding proteins described to date. The 52-kDa protein displayed similar affinities for untreated and deglycosylated pig IgG. The N-terminal amino acid sequence (SIITDVYAXEVLDSXGNPTLEV) revealed no homology with any bacterial proteins in the Swiss-Prot database. Its isoelectric point of approximately 4.6 and its amino acid composition, rich in aspartic and glutamic acids, showed that it had some similarities with other IgG-binding proteins. In this report, we have purified and characterized a 52-kDa IgG-binding protein from S. suis capsular type 2. Although this protein shares some similarities with other IgG- and/or IgA-binding proteins, it is unique in reacting with chicken IgG.     

Characterization of Group A Streptococcal M23 Protein and Comparison of the M3 and M23 Protein’s Ligand-Binding Domains

The present study concerns the properties for binding of human plasma and extracellular matrix proteins and the relationship between M3 and M23 molecules. Here, it is demonstrated that M23 protein shows a multiple binding to fibrinogen (FG), fibronectin (FN), human serum albumin (HSA), immunoglobulin G (IgG), kininogen, and collagen type I (CI) in Western blot analysis. Some sets of truncated-recombinant M3 or M23 protein fragments were assayed for their capacity to bind FN, FG, IgG, HSA, and CI. The HSA binding activity resided in the C-repeat region of M3 protein, whereas fibrinogen-binding activity resided in the A-repeat region. The FG, FN, and IgG binding sites were mapped to the N-terminal portion of M23 protein, whereas HSA binding was localized in the B-repeat domain, which has homology with C-repeat domain in M3 molecule. Therefore, it is concluded that the FN, FG, and IgG binding regions in the M3 and M23 proteins are quite dissimilar at the amino acid sequence level, whereas HSA binding is localized to the conserved C-repeat domain in the M3 and M23 proteins.     

Production and immunohistochemical application of antiserum against Tyr-D-Ala-Phe, a N-terminal tripeptide common to dermorphin/deltorphin family

Tyr-D-Ala-Phe is a N-terminal sequence commonly found in a peptide family including dermorphin and deltorphin. The tripeptide was synthesized and conjugated with poly L-lysine. Nuclear magnetic resonance (NMR) indicated that approximately 38 molecules of the tripeptide were bound to each molecule of poly L-lysine. The conjugate was used to immunize rabbits, and high titer antisera were obtained. An IgG fraction was purified by protein G affinity chromatography. A specific antibody to the tripeptide was then obtained by affinity chromatography using formylcellulofine conjugated with Tyr-D-Ala-Phe. On immunospot assay, the best IgG antibody was capable of detecting 125 ng of Tyr-D-Ala-Phe but failed to react even with 2.0 microg of Tyr-L-Ala-Phe or poly L-lysine. Our immunohistochemical examination selectively localized the secretory glands of frog skin.     

Chromatographic and protein chemical analysis of the ubiquinol-cytochrome c2 oxidoreductase isolated from Rhodobacter sphaeroides

The ubiquinol-cytochrome c2 oxidoreductase (cytochrome bc1 complex) purified from chromatophores of Rhodobacter sphaeroides consists of four polypeptide subunits corresponding to cytochrome b, c1, and the Rieske iron-sulfur protein, as well as a 14-kDa polypeptide of unknown function, respectively. In contrast, the complex isolated from Rhodospirillum rubrum by the same procedure lacked a polypeptide corresponding to the 14-kDa subunit. Gel-permeation chromatography of the R. sphaeroides cytochrome bc1 complex in the presence of 200 mM NaCl removed the iron-sulfur protein, while the 14-kDa polypeptide remained tightly bound to the cytochromes; this is consistent with the possibility that the latter protein is an authentic component of the complex rather than an artifact of the isolation procedure. The individual polypeptides of the R. sphaeroides complex were purified to homogeneity by gel-permeation chromatography in the presence of 50% aqueous formic acid and their amino acid compositions determined. The 14-kDa polypeptide was found to be rich in charged and polar residues. Edman degradation analysis indicated that its N terminus is blocked and not rendered accessible by de-blocking procedures. Cyanogen bromide cleavage gave rise to a blocked N-terminal fragment as well as a C-terminal peptide comprising more than one-third of the protein. Gas-phase sequence analysis of this peptide established a sequence of 48 residues and identified a putative trans-membrane segment near the C terminus. The blocked N-terminal fragment was cleaved at tryptophan with BNPS-skatole. The resulting peptides, together with tryptic fragments derived from the intact protein, yielded additional sequence information; however, none of the sequences exhibited significant homologies to any known proteins. Tryptic fragments were also used to generate sequence information for cytochrome c1.     

In vitro interaction of a polypeptide homologous to human Ro/SS-A antigen (calreticulin) with a highly conserved amino acid sequence in the cytoplasmic domain of integrin alpha subunits

We endeavored to identify proteins interacting with KLGFFKR, a highly conserved motif in the cytoplasmic domain adjacent to the transmembrane domain of the alpha subunit of integrins. We found that affinity chromatography of cell extracts with this peptide followed by elution with EDTA resulted in the isolation of a 60-kDa protein (p60). The N-terminal amino acid sequence of this 60-kDa polypeptide was found to be highly homologous to the Ro/SS-A antigen, a 60-kDa protein homologous to calreticulin and Aplysia "memory molecule". The binding of p60 was found to be specific for the KLGFFKR sequence since this polypeptide did not bind to a peptide with a scrambled amino acid sequence (KLRFGFK), and it was also specifically eluted from the KLGFFKR affinity matrix ith soluble KLGFFKR peptide but not with the scrambled peptide. Solid phase in vitro binding assays demonstrated specific interaction of p60 with integrin alpha 3 and alpha 5 subunits but not with the beta 1 subunit. Furthermore, p60 could be copurified with alpha 3 beta 1 following coincubation in vitro. These interactions could be inhibited by KLGFFKR peptide and also by EDTA, indicating sequence-specific and divalent cation dependent binding. Despite the fact that calreticulin is thought to be localized in the endoplasmic reticulum, a pool of Ro/SS A antigen homologous 60-kDa polypeptide was found to be present in the soluble cytoplasm, indicating the feasibility of an interaction of p60 with the integrin alpha subunits. Our data suggest that p60 (Ro/SS-A Ag) can specifically bind to integrin alpha subunits via the highly conserved KLGFFKR amino acid sequence.     

Proteolytic processing of the mosquitocidal toxin from Bacillus sphaericus SSII-1.

The 97-kDa protein Mtx21, derived from the 100-kDa mosquitocidal protein (Mtx) from Bacillus sphaericus SSII-1 by the deletion of the putative signal sequence, was expressed as a fusion protein with glutathione S-transferase in Escherichia coli, and the fusion protein was purified by affinity chromatography. The fusion protein bound to glutathione agarose was cleaved with thrombin to release the Mtx21 protein. The 97-kDa Mtx21 protein was found to be toxic to Culex quinquefasciatus larvae with a 50% lethal concentration of 15 ng/ml. Treating Mtx21 with crude mosquito larval gut extracts gave rise to two major peptides of 70 and 27 kDa. Treating the 97-kDa Mtx21 protein with trysin also gave rise to a similar proteolytic cleavage pattern. N-terminal sequencing showed that the 27-kDa peptide was derived from the N-terminal region of the 97-kDa protein and that the 70-kDa protein was from the C-terminal region of the 97-kDa protein. The 27-kDa peptide has all the previously identified regions of homology with the catalytic peptides of the ADP-ribosyltransferase toxins, such as pertussis toxin S1 peptide, while the 70-kDa peptide has three internal regions of homology.     

Site-directed lipid modification of IgG-binding protein by intracellular bacterial lipoprotein process.

IgG-binding protein was genetically expressed and lipid-modified in a site-directed manner in Escherichia coli. The DNA sequence encoding the signal peptide and the nine N-terminal amino acid residues of the major lipoprotein of E. coli (lpp) was fused to the sequence of B-domain which was one of the IgG binding domains of Staphylococcal Protein A (SpA). The N-terminal cysteine residue of the resulting protein was enzymatically linked with lipids in the bacterial membrane. The lipid-modified protein was translocated at the bacterial membrane in a manner similar to native bacterial lipoprotein, and it was purified with IgG-Sepharose by affinity chromatography. The lipid modified proteins (lppB1 and lppB5) showed a similar IgG binding activity to unmodified proteins, which was estimated by competitive ELISA. Proteoliposomes of lipid modified proteins were prepared in an elegant fashion so that the IgG binding site should be properly oriented on the surface of an individual liposome by anchoring the lipid-tail into the hydrophobic layer of the liposome membrane. As compared with the unmodified one, the lipid modified protein incorporated into the proteoliposome exhibited higher IgG binding activity.     

Characterization of the Co(2+) and Ni(2+) binding amino-acid residues of the N-terminus of human albumin. An insight into the mechanism of a new assay for myocardial ischemia.

Patients suffering from myocardial ischemia reportedly exhibit reduced in vitro binding of exogenous Co(2+) to the N-terminal of human serum albumin (HSA). The purpose of our investigation was to simulate changes in the N-terminus of HSA that may account for these ischemia-induced modifications to the cobalt binding site. HPLC, LC-MS and (1)H NMR analyses have shown that the N-terminal region of HSA Asp-Ala-His-Lys binds the transition metals Co(2+) and Ni(2+). Synthetic peptides with the first 2-12 amino acids of the HSA sequence demonstrated that the first three amino acids, Asp-Ala-His, are essential for strong binding of cobalt. Modification of the N-terminus peptide of HSA by way of N-acetylation or the deletion of one or more amino acid resulted in no binding of cobalt. Because the degradation of the susceptible, specific transition metal binding site of HSA may account for the decreased cobalt binding observed during ischemic events, an assay that detects this reduced binding could be useful in the diagnosis of ischemia.     

Primary structure of human plasma fibronectin. Characterization of a 31,000-dalton fragment from the COOH-terminal region containing a free sulfhydryl group and a fibrin-binding site

The 31-kDa domain of human plasma fibronectin has been completely characterized. This fragment is located at the COOH-terminal end of the molecule immediately preceding the 3-kDa interchain disulfide-containing peptide. The 31-kDa domain was obtained after trypsin digestion of fibronectin and purified by affinity chromatography on gelatin- and heparin-Sepharose columns. The fragment eluted in the heparin-unbound fraction and was further purified by DEAE-cellulose and high performance liquid chromatography. The 31-kDa fragment contained a fibrin-binding site (fibrin II site) which was only active at physiological NaCl concentrations and therefore differed from that located in the NH2-terminal domain which also bound at lower NaCl concentrations. The 31-kDa domain bound to thiopropyl-Sepharose and was shown to contain a free sulfhydryl group located at position 35 in the sequence. To determine the complete amino acid sequence of this fragment, a trypsin digestion was performed on the reduced and alkylated 31-kDa domain, and the 17 resulting peptides were isolated by high performance liquid chromatography; their amino acid compositions and amino acid sequences have been determined, and the arrangement of peptides was achieved by comparison with the sequences deduced from human and rat cDNA clones and with a related plasmic fragment from bovine fibronectin. Comparison of these three sequences showed 23 amino acid differences between human and rat fibronectin and 16 between human and bovine fibronectin. This represents a 91 and 94% homology, respectively. An interesting finding is that the 31-kDa fragment contains a deletion of 31 residues when compared to the rat cDNA sequence. This deletion appears to represent a species difference since it is due to a shorter mRNA in the case of human fibronectin.     

Novel lectin-like proteins on the surface of human monocytic leukemia cell line THP-1 cells that recognize oxidized cells

Presence of lectin-like receptors on the membranes of human monocytic leukemia cell line THP-1 cells for clustered sialylated poly-N-acetyllactosaminyl sugar chains on the membranes of oxidized erythrocytes and T-lympoid cells was investigated. Membranes of THP-1 cells differentiated into macrophages were solubilized, and the membrane proteins obtained by affinity chromatographies using lactoferrin-Sepharose and band 3-Sepharose were purified by successive DE column chromatography and sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Proteins of 50, 60, and 80 kDa with specificity to bind to sialylated poly-N-acetyllactosaminyl sugar chains were detected in the chromatographic fractions. A 50-kDa protein was isolated in a pure form. N-Terminal amino acid sequence of the protein was Lys-Gln-Lys-Val-Ala-Gly-Lys-Gln-Pro-Val-, which has not been found in the N-terminal regions of the hitherto known proteins. The antibody, raised against the chemially synthesized peptide composed of the N-terminal amino acid sequence, bound to 50-, 60-, and 80-kDa proteins as analyzed by immunoblotting and immunoprecipitation, indicating that these proteins had the same N-terminal amino acid sequence. The results demonstrate that THP-1 cells have novel 50-, 60-, and 80-kDa lectin-like proteins with the same N-terminal amino acid sequence on the cell surface which would bind to clustered sialylated poly-N-acetyllactosaminyl sugar chains generated on oxidized erythrocytes and T-lymphoid cells.     

A gene fusion system for generating antibodies against short peptides

A novel method to obtain specific antibodies against short peptides is described, involving synthesis of the corresponding oligodeoxynucleotides followed by cloning into a new set of fusion vectors, pEZZ8 and pEZZ18, based on two synthetic IgG-binding domains (ZZ) of Staphylococcus aureus protein A. The soluble gene fusion product thus obtained, can be collected from the culture medium of Escherichia coli and rapidly recovered in a one-step procedure by IgG affinity chromatography. The system was used to express a fusion protein consisting of the two Z fragments and the C-terminal part [amino acids (aa) 57-70] of human insulin-like growth factor I (IGF-I). This 16-kDa protein was purified by affinity chromatography on IgG Sepharose and antibodies were raised in rabbits. The fusion protein elicited peptide-specific antibodies, as measured by solid-phase radioimmuno assay and Western blotting, reactive with both synthetic C-terminal peptide and the native human IGF-I protein. The results suggests that the gene fusion system can be used for efficient antibody production against short peptides encoded by synthetic oligodeoxynucleotides.     

An anti-peptide antibody that recognizes the dopamine D2 receptor from bovine striatum.

The bovine dopamine D2 receptor was purified by wheat-germ-agglutinin-Sepharose chromatography and affinity chromatography, using the D2-receptor-specific agonist N-0434. Purification yields a preparation with a major protein band of 95 kDa. In order to ascertain the identity of this protein, polyclonal antibodies against the dopamine D2 receptor have been raised using synthetic peptides based on the predicted amino acid sequence of the cloned D2 receptor. For the initial screening of these antibodies, three fusion proteins consisting of beta-galactosidase and receptor fragments were constructed. One antiserum reacted strongly with the corresponding D2 receptor fusion protein, both on Western blots and in immunoprecipitation experiments. In each case, recognition was inhibited by competition with free peptide. On Western blots of partially purified receptor preparations from bovine striatum, the antiserum specifically recognized a 95-kDa glycoprotein. From similar preparations, the antiserum precipitated a substantial proportion of active D2 receptor, as determined by a decrease in [3H]spiperone binding in the supernatant. Active receptor could be released from the immunoprecipitate by addition of free peptide. Immunocytochemical analysis of cells transiently transfected with DNA coding for the D2 receptor showed specific staining of transfected cells. The antibody raised against a sequence in the third intracellular loop is able to shift the affinity of the receptor for dopamine from high to low, indicating that the antiserum may be interfering with receptor-GTP-binding-protein interactions.     

The primary structure of the Cytisus scoparius seed lectin and a carbohydrate-binding peptide.

The complete amino acid sequence of 2-acetamido-2-deoxy-D-galactose-binding Cytisus scoparius seed lectin II (CSII) was determined using a protein sequencer. After digestion of CSII with endoproteinase Lys-C or Asp-N, the resulting peptides were purified by reversed-phase high performance liquid chromatography (HPLC) and then subjected to sequence analysis. Comparison of the complete amino acid sequence of CSII with the sequences of other leguminous seed lectins revealed regions of extensive homology. The amino acid residues of concanavalin A (Con A) involved in the metal binding site are highly conserved among those of CSII. A carbohydrate-binding peptide of CSII was obtained from the endoproteinase Asp-N digest of CSII by affinity chromatography on a column of GalNAc-Gel. This peptide was retained on the GalNAc-Gel column and was presumed to have affinity for the column. The amino acid sequence of the retarded peptide was determined using a protein sequencer. The retarded peptide was found to correspond to the putative metal-binding region of Con A. These results strongly suggest that this peptide represents the carbohydrate-binding and metal ion-binding sites of CSII.     

Purification and characterization of an algal (Dunaliella tertiolecta) protein cross-reacting with an anti-Ga-common antibody

A 26-kDa and a 36-kDa protein that cross-reacted with anti-G_a-common and anti-G_β antibodies, respectively, were detected in Dunaliella cells. The 26-kDa protein was solubilized from a crude membrane fraction with deoxycholate and purified to homogeneity by DE52 and hydroxylapatite chromatography and DEAE-5PW high performance liquid chromatography (HPLC). The hydroxylapatite-purified preparation had GTPγS binding and GTPase activities, but the homogeneous 26-kDa protein had none. The sequence of the 28 N-terminal amino acids of the 26-kDa protein had no homology to any GTP binding protein thus far reported.     

Purification and Characterization of Canine Serum Ferritin-binding Proteins

Ferritin-binding protein (FBP) is known to interact with circulating ferritins in mammals. Canine FBPs were purified from canine serum by affinity chromatography and were identified as IgM, IgG, and IgA by immunoblotting with alkaline phosphatase-labeled antibodies to canine IgM, IgG, and IgA heavy chains. Following further purification by application to a Sephacryl S-300 column, canine FBPs were separated into 81.3- and 27.7-kDa bands by sodium dodecyl sulfate-polyacryamide gel electrophoresis, and the 81.3-kDa band reacted with the anti-canine IgM heavy chain antibody. Purified canine FBP bound to canine liver ferritin, but not to canine albumin and transferrin. FBP showed greater binding to the expressed bovine ferritin H-chain homopolymer than to the expressed bovine ferritin L-chain homopolymer. The binding of FBP with canine liver ferritin was dose-dependently inhibited by anti-rat liver ferritin antibody, and the anti-ferritin antibody dissociated the bound FBP in a dose-dependent manner, even after binding FBP with liver ferritin. The canine ferritin H subunit peptide fragment with amino acid residues 148–155 (NH₂-GDHVTNLR-COOH) in its C-terminal region was recognized by FBP. These results indicate that canine serum FBPs are autoantibodies to ferritin (IgM, IgG, and IgA) and that anti-ferritin autoantibody (IgM) recognizes the C-terminal region of ferritin H subunit.     

A physicochemical study of protein G, a molecule with unique immunoglobulin G-binding properties

Protein G, an IgG-binding molecule, was prepared from the cell walls of a group G streptococcal strain, G-148. The protein could be extracted from the cells by papain digestion and purified by the sequential use of ion-exchange chromatography on DEAE-Sephadex, affinity chromatography on Sepharose-coupled human IgG, and gel chromatography on Sephadex G-200. Two protein bands with similar molecular weight, 34,000 and 36,000, were obtained when analyzing the pure protein G on sodium dodecyl sulfate-polyacrylamide gel electrophoresis. The yield using this purification scheme was 27% of the protein G solubilized from the cells or 70 micrograms/ml packed bacteria. The Stokes radius and frictional ratio of protein G were determined to 3.53 nm and 1.64, respectively, suggesting an elongated fibrous molecule. The protein did not contain any intrachain disulfide bonds. The amino acid composition of protein G was determined and was found to be different from that of protein A, the well known staphylococcal IgG-binding protein. The equilibrium constants of the reactions between protein G and human, rabbit, mouse, and goat polyclonal IgG, determined by Scatchard plots, ranged between 1 X 10(10) and 7 X 10(10), for rat polyclonal IgG 1.4 X 10(9), and human monoclonal IgG1, IgG2, IgG3, and IgG4 between 2 X 10(9) and 6 X 10(9). These affinity constants were always greater than the corresponding values for protein A. The binding between protein G and various polyclonal and monoclonal IgG was pH dependent between 2.8 and 10, strongest at pH 4 and 5, and weakest at pH 10.     

A small bispecific protein selected for orthogonal affinity purification

A novel protein domain with dual affinity has been created by randomization and selection. The small alkali-stabilized albumin-binding domain (ABD*), used as scaffold to construct the library, has affinity to human serum albumin (HSA) and is constituted of 46 amino acids of which 11 were randomized. To achieve a dual binder, the binding site of the inherent HSA affinity was untouched and the randomization was made on the opposite side of the molecule. Despite its small size and randomization of almost a quarter of its amino acids, a bifunctional molecule, ABDz1, with ability to bind to both HSA and the Z2 domain/protein A was successfully selected using phage display. Moreover, the newly selected variant showed improved affinity for HSA compared to the parental molecule. This novel protein domain has been characterized regarding secondary structure and affinity to the two different ligands. The possibility for affinity purification on two different matrices has been investigated using the two ligands, the HSA matrix and the protein A-based, MabSelect SuRe matrix, and the new protein domain was purified to homogeneity. Furthermore, gene fusions between the new domain and three different target proteins with different characteristics were made. To take advantage of both affinities, a purification strategy referred to as orthogonal affinity purification using two different matrices was created. Successful purification of all three versions was efficiently carried out using this strategy.