Five Homologous Repeats of the Protein G-Related Protein MIG Cooperate in Binding to Goat Immunoglobulin G

Abstract

Protein MIG, from Streptococcus dysgalactiae, binds α₂-macroglobulin and immunoglobulin G (IgG). MIG-derived fusion proteins with one to five IgG-binding repeats differed up to 72,000-fold in avidity for goat IgG, indicating a considerable cooperativity of the repeats. Significant sequence variation in the IgG-binding repeats was recognized. Protein MIG interacted with goat IgG1 via both the Fc and Fab parts.

Streptococcus dysgalactiae SC1, originally isolated from a bovine mastitis specimen, expresses a 72-kDa protein (protein MIG) that reacts with immunoglobulins G (IgG) of different species and subclasses as a type III Fc receptor (12). The amino acid sequence of protein MIG shows relationship to protein G (8, 10).

In both molecules a domain in the N terminus binds to the proteinase inhibitor α₂-macroglobulin (α₂M) (10, 14). However, the binding regions do not reveal any homology and protein G binds the native form of α₂M while protein MIG reacts with α₂M-protease complexes (13, 14). The IgG-binding region of protein MIG consists of five repeated domains, compared to two to three repeated domains in protein G (5, 8, 17). The IgG-binding repeats of the two molecules are highly homologous (10). However, the first IgG-binding repeat of protein MIG differs from any of the protein G repeats by 12 to 13 amino acids, a difference that may alter the interaction with IgG. To evaluate the influence of these amino acid differences and of increasing numbers of repeats from protein MIG on its avidity to bind goat IgG, we constructed maltose-IgG-binding fusion proteins by using three different primers (Scandinavian Gene Synthesis) based on the mig gene (10). Primer 1 (5′-CTCGAATTCGTTCAACTAGAAGCACCTACA-3′) hybridized at the junction between the α₂M-binding domain and the first IgG-binding repeat of pAM1 (10), nucleotides 1118 to 1138, which added an EcoRI site 5′ to the start of the repeat. Primer 2 (5′-TCATTATTCAGTAACTGTAAAGGTTTTAGT-3′) hybridized to the extreme C terminus of each of the first four IgG-binding repeats, nucleotides 1339 to 1316, 1549 to 1526, 1759 to 1736, and 1969 to 1946, which added a tandem stop codon 3′ to the repeats. Primer 3 (5′-TCATTAAGGAACTTCAGTAACCATTTC-3′) hybridized at the junction between the fifth IgG-binding repeat and the cell wall-spanning region, nucleotides 2197 to 2177, which added a tandem stop codon 3′ to the repeat. The combination of primers 1 and 2 resulted in DNA fragments encoding one to four IgG-binding repeats, and the combination of primers 1 and 3 resulted in a DNA fragment encoding all five IgG-binding repeats. PCR products were ligated into pMAL-C2 (New England Biolabs), and after transformation to Escherichia coli, five different clones carrying pMAL1R to pMAL5R, which bear genes encoding maltose-binding proteins fused to various numbers of IgG-binding repeats, were isolated (Fig. 1A). Additionally, we constructed a clone with an N-terminal deletion of pMAL1R, designated pMAL1RCT, which expressed the last 28 amino acids (protein MAL1RCT) of the first IgG-binding repeat of protein MIG (Fig. 1B).

FIG. 1.

Schematic representation of the native MIG protein molecule and the fragments and peptides derived from the IgG-binding domain. (A) The numbers 1 through 5 indicate the five repeats of the IgG-binding domain, and the hatched area represents the cell wall-spanning and membrane-anchoring regions of the native molecule. The various fragments were expressed as fusions with the maltose-binding protein (MBP). The molecular masses of the proteins calculated from the amino acid sequences are given in parentheses. (B) Schematic representation of the peptides derived from the first IgG-binding repeat of protein MIG and the C1 domain of protein G. The numbers above the upper bar indicate the amino acid positions in the one repeat of protein MIG. The lower bar represents the truncated form of the repeat (1RCT). The arrows show the sequences of the 11-amino-acid-long peptides derived from the repeat and the corresponding region in the C1 domain of protein G. The identical amino acids in these two peptides are indicated by vertical bars.

Sequencing of the clones over the vector insert junctions revealed a base substitution in pMAL4R, which caused an amino acid change (Glu→Gly). The expression of this construct was very weak, and the molecule was therefore not further studied. Unlike the other proteins, protein MAL1RCT showed very weak binding when it was applied to IgG-Sepharose. In Western blots, all purified fusion proteins except protein MAL1RCT revealed IgG-binding activity (data not shown). Factor Xa cleavage of MAL1R resulted in one free repeat, which no longer reacted with IgG in Western blot assays (data not shown). To investigate the influence of an increasing number of IgG-binding repeats on the interaction with goat IgG subclasses, the fusion proteins (MAL1R to MAL5R) were absorbed in wells of microtiter plates and goat IgG1 and IgG2, purified as recently described (15), were allowed to compete for the binding of the goat IgG1-alkaline phosphatase conjugate (alkP-IgG1; Bio-Rad). As shown in Fig. 2, the avidities of MIG constructs for goat IgG1 increased as the number of IgG-binding repeats increased. Similar curves were obtained with goat IgG2 (data not shown). Data were analyzed with a curve-fitting program (9, 11), and the IgG concentrations giving 50% inhibition (IC₅₀) of binding by the constructs were calculated (Table 1). These values did not reveal significant differences between goat IgG1 and IgG2, indicating similar reactivities of the fusion proteins to both subclasses. It is therefore unlikely that any of the IgG-binding repeats could interact exclusively with only one of the goat IgG subclasses.

FIG. 2.

Inhibition by goat IgG1 of the binding of alkP-goat IgG1 to various MIG-derived fusion proteins immobilized on microtiter plates.

TABLE 1.

Goat IgG1 and IgG2 IC₅₀ of the binding between alkaline phosphatase-conjugated goat IgG1 and various fusion proteins immobilized on the surfaces of microtiter wells

Fusion protein on surface of well	Mean IC50 ± SD (nM)a
	IgG1	IgG2
MAL1R	36.8 ± 6.8	55.6 ± 26.7
MAL2R	16.7 ± 5.0	18.9 ± 10.2
MAL3R	4.6 ± 0.8	3.0 ± 0.7
MAL5R	1.0 ± 0.2	1.5 ± 0.4

^aValues were calculated from the curves shown in Fig. 2. 

Earlier experiments with protein G variants containing two or three repeats have revealed a 10-fold higher affinity for the three-repeat molecule (2). As protein MIG contains five IgG-binding repeats, it was of interest to evaluate how the increasing number of repeats would contribute to the avidity for IgG binding. However, the binding regions of fusion proteins immobilized on solid surfaces may become partly blocked or sterically hindered. We therefore also investigated their interaction with goat IgG1 in solution by allowing them to compete with the binding of alkP-IgG1 to MAL5R-coated microtiter plates. Complete inhibition was observed with all fusion proteins (Fig. 3). The IC₅₀ differed greatly and were 400-, 3,400-, and 72,000-fold higher for MAL3R, MAL2R, and MAL1R, respectively, than for MAL5R. The IC₅₀ recalculated with regard to the number of IgG-binding repeats involved in each assay still differed greatly. We recognized a linear relationship between the log₁₀ units of the IC₅₀ and the number of repeats in the fusion protein (data not shown). Thus, the five IgG-binding repeats of protein MIG obviously cooperate to a considerable extent in the binding of IgG. In light of the results of experiments with passively immobilized protein MIG fragments, it is becoming clear that the full IgG-binding activity of protein MIG is available only when the proteins interact in solution.

FIG. 3.

Competition between MAL5R applied to the surfaces of microtiter plates and various protein MIG constructs in solution for binding of labeled goat IgG1. Samples from dilution series of MAL1R (•), MAL2R (□), MAL3R (▵), and MAL5R (○), mixed with an equal volume of 10⁵-diluted alkP-goat IgG1, were incubated in wells of enzyme-linked immunosorbent assay plates coated with MAL5R.

An inhibitory effect on the binding of IgG1 to MAL5R-coated wells was also observed with the truncated form of the first IgG-binding repeat (MAL1RCT) (Fig. 4), but the shape of the dose-response curve differed from those of the curves obtained with the other constructs. The MIG-derived amino acid sequence of the truncated form differs in five residues from the corresponding part of the C1 domain in protein G. This region of the protein G domain contains most of the amino acid residues responsible for binding to the Fc fragments of immunoglobulins (3, 7). Frick et al. (6) demonstrated that not only a 28-amino-acid synthetic peptide of the C1 domain (corresponding to our MAL1RCT) but also an even smaller peptide (PG8) of only 11 amino acids could efficiently inhibit the interaction between protein G and human Fc fragments immobilized on polyacrylamide beads. Based on the three-dimensional structure of a single protein G repeat, the 11 residues are localized in the C terminus of the α-helix, the N terminus of the third β-strand, and the loop region connecting these two structural elements. Recent crystal structure data confirm that these 11 amino acids are involved in direct contact with IgG Fc (16). In our experiments, a corresponding 11-amino-acid synthetic peptide from the first repeat of protein MIG could, like the PG8 peptide from protein G, inhibit the binding of a protein G fragment (fragment CDC) to human Fc but only at marginally higher concentrations (Fig. 5). Nevertheless, the activity of the MIG-derived peptide suggests that despite the amino acid differences (4 of 11 amino acids [Fig. 1]), the secondary structure as well as the mode of action of the first MIG repeat may be similar to those aspects of the C1 domain of protein G.

FIG. 4.

Inhibition of binding of labeled goat IgG1 to surface-bound MAL5R by MAL1RCT. The 100% binding value represents the interaction between the labeled IgG1 and MAL5R in the absence of inhibitor.

FIG. 5.

Inhibition of the binding of ¹²⁵I-labeled protein G to immobilized human Fc by synthetic peptides derived from protein MIG (○) or protein G (•).

An earlier study by Eliasson (4) indicated that protein G binds bovine IgG exclusively via the F(ab′)₂ part. In contrast, human Fc fragments strongly interact with protein G but human Fab fragments reveal only 10% of the activity of Fc fragments (1). However, binding of goat IgG to MAL5R may be only partly inhibited (to about 55%) by human IgG Fc (Fig. 6) whereas binding of labeled rabbit IgG to the same MIG protein construct was completely inhibited by the human Fc fragment. Moreover, the interaction between ¹²⁵I-labeled goat IgG Fc fragments and the native MIG protein on streptococcal cells was completely blocked by both human Fc and goat IgG (data not shown). Thus, there is a clear interaction between protein MIG and goat IgG via both the Fc and the Fab part of the molecule.

FIG. 6.

Inhibition of binding of labeled goat IgG1 or labeled rabbit IgG to MAL5R-coated surfaces by human Fc fragment. The 100% binding value represents the interaction between labeled IgG and MAL5R in the absence of inhibitors.

Although the primary structure (10) suggests that the valency of protein MIG for IgG is 5, the functional valency has so far not been determined. Preliminary biosensor data from our laboratory indicate that a single MIG protein molecule may interact with at least three IgG molecules. It is tempting to speculate that the relatively large IgG-binding domain of protein MIG with five binding repeats, connected by extension sequences of 15 amino acids, may mediate the strong interaction with goat IgG by simultaneously binding to domains in both the Fc and the Fab part. If such interactions occur in vivo, they may lead to conformational changes of the IgG molecule, interfere with complement activation and phagocytosis, and thus act like a virulence factor.