GRAB (Protein G-related α2M-binding protein) is a surface protein of group A streptococci and exhibits high affinity for α2-macroglobulin (α2M), a broad-range protease inhibitor. It is the sole α2M-binding protein of group A streptococci that has been shown to promote bacterial virulence in a mouse model of skin infection. The binding site for α2M was predicted to be in the N-terminal A domain of GRAB. In the present study, the α2M-binding domain was first narrowed down to 34 amino acids (amino acids 34–67) using variable truncated N-terminal GRAB fusion proteins. The sequence of the identified domain was used to design overlapping synthetic peptides of different sizes, which were then immobilized on a membrane and assayed for their α2M-binding activity. The peptide screening revealed two binding motifs of ten amino acids length, located in the ΔA (N-terminal part of the A domain) region (amino acids 34–67) with the sequences PRIIPNGGTL (amino acids 41–50) and NAPEKLALRN (amino acids 56–65) respectively. These motifs were used for systematic mutational analysis by generating synthetic peptides containing individual amino acid substitutions at every position of the mapped binding regions. The results indicated a critical role for the arginine residue at position 42 in the first binding domain and at position 64 in the second binding region. Validation of arginine residues as the critical amino acids for α2M binding was achieved by site-directed mutagenesis and binding assays. Competitive inhibition assays with GRAB containing amino acid substitutions R42G (Arg42→Gly), R64G and R42G/R64G indicated differential contribution of the arginine residues at positions 42 and 64 to α2M-binding activity and, thus, their involvement in GRAB-induced virulence.
- α2-macroglobulin (α2M)
- Protein G-related α2-macroglobulin-binding protein (GRAB)
Group A streptococci (GAS; Streptococcus pyogenes) are important human pathogens that are commonly associated with mild and self-limiting diseases such as pharyngitis, pyoderma or asymptomatic colonization. Nevertheless, GAS have the potential to cause life-threatening invasive diseases like necrotizing fasciitis, bacteraemia, toxic shock syndrome and rheumatic heart disease [1,2]. In view of the resurgence of severe streptococcal invasive infections since the early 1980s [3–7], elucidation of the mechanisms that enable bacteria to invade the host and survive its defence mechanisms, including phagocytosis or complement-mediated cell lysis, is pivotal to a better understanding of GAS diseases [1,8–10].
The surface protein GRAB [Protein G-related α2M (α2-macroglobulin)-binding protein] is one of the GAS components promoting virulence  and was originally identified by the high degree of homology of its N-terminal part with the α2M-binding E domain of Protein G from group C and G streptococci, indicating a similar α2M-binding mechanism in both proteins [12–15]. GRAB is a protein with a molecular mass of 22.8 kDa, consisting of a cell-wall-anchoring region with a typical Gram-positive membrane anchor LPXTG motif [8,16], a varying number of 28-amino-acid-long repeats and a highly conserved, functional N-terminal domain responsible for the interaction with the broad-range protease inhibitor α2M . α2M is a 725 kDa homotetrameric plasma glycoprotein capable of inhibiting proteases of all classes and specificities through a unique mechanism involving steric entrapment and covalent binding [17–20]. The grab gene is present in almost all GAS isolates and contributes to bacterial virulence, as shown recently in a murine skin model of GAS infection [11,12]. Recruitment of a broad-range protease inhibitor such as α2M by the bacteria has been proposed as a mechanism leading to the protection of bacterial surface structures, such as the antiphagocytic M-protein, from proteolytic degradation [12,21,22]. Furthermore, depletion of the surrounding protease inhibitors probably leads to an increase in the amount of free proteases and enhances tissue destruction during the infection process [9,23,24]. Thus interaction with protease inhibitors, such as α2M, may enable bacteria to protect their surface structures and facilitate progressive dissemination in the tissue . The fact that neither GAS nor human pathogenic group G and C streptococci bind the electrophoretically fast form of α2M (f-α2M) that had already been complexed with proteases supports this hypothesis . Both pathogens only bind the electrophoretically slow form of α2M (s-α2M) with protease inhibitory activity [14,26–28]. Neither GRAB protein itself nor Protein G exhibits proteolytic activity, altering the conformational status of the bound plasma protein from s-α2M to f-α2M as demonstrated in previous studies . Thus surface-recruited α2M remains active and provides the bacterium with a mechanism to interact with foreign or its own proteases [12,14,15].
In the present study, the GRAB–α2M interaction was analysed to map the minimal binding motif(s) and critical amino acid(s) of GRAB mediating the high-affinity interaction with α2M. Analysis of spot-synthesized synthetic peptides of GRAB and competitive inhibition experiments with recombinant GRAB derivatives identified two binding motifs, PR42IIPNGGTL and NAPEKLALR64N, in the ΔA (N-terminal part of the A domain) region (amino acids 34–67) of GRAB. Individual amino acid substitutions at every position in the motifs and competitive inhibition experiments using the mutated recombinant GRAB derivatives rGRAB42, rGRAB64 and rGRAB42/64 demonstrated that arginine residues are critical for the protein–protein interaction and, hence, have a pivotal role in the GRAB-induced virulence of GAS.
Bacterial strains, growth conditions and protein purification
S. pyogenes strains were grown in Todd–Hewitt broth (Invitrogen, Karlsruhe, Germany) supplemented with 1% yeast extract (Difco, Heidelberg, Germany) (referred to as THY) under static conditions at 37 °C or on blood agar plates (Becton Dickinson, Heidelberg, Germany). Epicurian Coli™ XL1-Blue cells as the host for recombinant pGEX-6P-1 (Amersham Biosciences) were grown in Luria–Bertani medium or on Luria–Bertani agar with ampicillin (100 μg/ml). The medium of Escherichia coli M15-[pREP4] containing recombinant pQE30 (Qiagen) was supplemented with 100 μg/ml ampicillin and 25 μg/ml kanamycin. Expressions of GST (glutathione S-transferase)- and His-tagged fusion proteins were induced with 1.5 mM isopropyl β-D-thiogalactoside (Sigma–Aldrich) after the culture reached an attenuance (D600) of 0.6 and growth continued for 5 h at 30 °C. Recombinant His-tagged GRAB (rGRAB) and its derivatives rGRAB42, rGRAB64 and rGRAB42/64 were purified by affinity chromatography under native conditions on Ni2+-nitrilotriacetic acid agarose according to the general guidelines of QIAexpressionist™ (Qiagen). GST fusion proteins rGST-ΔA, rGST-A, rGST-AR1 and rGST-AR2 were purified by affinity chromatography using glutathione–Sepharose 4B (Amersham Biosciences) according to the manufacturer's instructions.
Human plasma protease inhibitor α2M
The human plasma protease inhibitor α2M (M6159), purchased from Sigma–Aldrich, was added to 0.02 M Tris/0.13 M glycine (pH 8.0) containing 0.08 M trehalose and freeze-dried. According to the manufacturer, 1 mg of α2M inhibits at least 10 μg of trypsin with an activity of 10000 BAEE units/mg of protein (where BAEE stands for Nα-benzoyl-L-arginine ethyl ester hydrochloride). The trypsin-binding activity of α2M was validated by an assay on the basis of the ability of the protein to protect the entrapped trypsin (Sigma–Aldrich) from inactivation by soya-bean trypsin inhibitor (Sigma–Aldrich), using BAEE (Sigma–Aldrich) as the substrate . Briefly, purified α2M was incubated for 7 min at 30 °C with increasing molar ratios of trypsin (Sigma–Aldrich). The proteolytic activity of the remaining trypsin that was not entrapped by α2M was inhibited by adding double the molar amount of soya-bean trypsin inhibitor (Sigma–Aldrich). Activity of the remaining trypsin was determined by adding BAEE as substrate for trypsin at a final concentration of 2.5 mg/ml to the reaction and monitoring the ΔA253. Trypsin alone and trypsin/soya-bean trypsin inhibitor were used as controls. In addition, the percentages of slow- and fast-forms of α2M in the sample were determined by SDS/PAGE and subsequent densitometric analysis of the different bands. α2M preincubated with trypsin was used as a control (Figure 1). A comparison of the different bands revealed that the protein purchased from Sigma–Aldrich was mainly a slow form (s-α2M), non-protease-complexed α2M (82%). The amount of f-α2M (fast-form α2M) detected was less than 20%.
Recombinant DNA techniques and PCR
Chromosomal DNA from the S. pyogenes strain A82 was prepared with Genomic-tip 100/G columns (Qiagen) according to the manufacturer's instructions and used as a template for PCR amplification of the grab gene. The grab sequence of GAS isolate A82 is identical with the grab sequence of S. pyogenes reference strain SF 370. PCR was performed with approx. 50 ng of chromosomal DNA as a template using Taq DNA polymerase (Qiagen) or Pfu DNA polymerase (Stratagene) under the buffer conditions recommended by the manufacturer. The PCR products were purified by gel extraction using the QIAquick PCR Purification kit (Qiagen). Plasmid DNA was isolated using QIAprep Spin Plasmid kit (Qiagen). T4 DNA ligase and the restriction enzymes were purchased from New England Biolabs. The transformation of recombinant plasmid DNA in E. coli was achieved by electroporation .
grab DNA fragments were amplified with oligonucleotides incorporating a BamHI restriction site at the 5′-end and a KpnI or EcoRI restriction site at the 3′-end. Oligonucleotides were designed from the grab sequence of the S. pyogenes strain SF 370 (A.T.C.C. 700294; GenBank® accession no. GI:4589078). PCR products were digested with BamHI–KpnI or BamHI–EcoRI respectively and ligated into similarly prepared expression vectors to allow the in-frame expression of GST- or His-tagged fusion proteins. The oligonucleotides, their features and the resulting grab DNA fragments generated by PCR are listed in Table 1.
Site-directed mutagenesis of grab
Site-directed mutagenesis of grab was performed by the method of Higuchi et al.  by incorporating single-base mismatches between the synthetic oligonucleotide and the template by PCR amplification. Each mutant grab DNA was generated in a two-step procedure. First, the overlapping DNA fragments designated as PCR1/PCR2 and PCR3/PCR4, containing the desired base-pair substitution at the selected position, were amplified by PCR (Table 1). Secondly, a fusion PCR was performed to generate the entire mutated DNA fragment from the overlapping fragments PCR1/PCR2 and PCR3/PCR4 using the oligonucleotides grab-BamHI and grab-KpnI (Table 1). The mutated DNA fragments were cloned in pGEM-T Easy under conditions recommended by the distributor (Promega) and, finally, after digestion with BamHI–KpnI, they were subcloned into the similarly prepared expression plasmid pQE30. The fusion PCR consisted of 2 cycles including 60 s of denaturation at 94 °C, 60 s of annealing at 70 °C and 120 s of extension at 72 °C, followed by 30 cycles including 60 s of denaturation at 94 °C, 60 s of annealing at 55 °C and 90 s of extension at 72 °C. The chromosomal DNA of the S. pyogenes strain A82 was used as a template. The oligonucleotides, the introduced point mutations and the resulting grab sequences are listed in Table 1. To generate the insert DNA of the double-mutant pQGR42/64 with single-point mutations at nucleotide positions 124 and 190, the plasmid pQGR42 was used as a template DNA by the method used for the generation of pQGR64. This strategy resulted finally in the expression of rGRAB42, rGRAB64 and rGRAB42/64. The integrity of the mutated grab sequences was confirmed by DNA sequence analysis using ABI PRISM dye terminator sequencing (PerkinElmer).
Antiserum against rGRAB was generated by Eurogentech (Seraing, Belgium). Rabbits were immunized intradermally with 50 μg of purified rGRAB encoded by pQGR in 1 ml of a 1:1 emulsion of antigen and complete Freund's adjuvant. The animals were boosted with 50 μg of rGRAB and incomplete adjuvant on days 14, 28 and 56 by the standard method. Polyclonal anti-rGRAB IgG of the derived serum fractions was purified using Protein A–Sepharose (Amersham Biosciences) as recommended by the distributor.
Electrophoresis and blot-overlay assay
Purified fusion proteins were subjected to SDS/PAGE (12% gel) by the method of Laemmli . The gels were then either stained with Coomassie Brilliant Blue or transferred on to a nylon membrane (Immobilon-P; Millipore) using a semi-dry blotting system (Bio-Rad) . Binding of purified human α2M (Sigma–Aldrich) to rGRAB and its derivatives was examined by the blot-overlay assay. Membranes were first blocked with 10 mM PBS containing 5% (w/v) skimmed milk for 1 h at room temperature (21 °C) and incubated for an additional hour with α2M (4 μg of α2M/ml of PBS). The binding of α2M to GRAB protein was detected with rabbit polyclonal IgG anti-human α2M (Sigma–Aldrich; 1:1000-diluted in PBS). Rabbit polyclonal IgG anti-rGRAB (1:500-diluted in PBS) was used as a control. Detection was performed by goat peroxidase-conjugated anti-rabbit IgG antibody (Sigma–Aldrich; 1:5000-diluted in PBS) followed by incubation with the substrate solution of 4-chloro-1-naphthol (1 mg/ml) and 0.1% H2O2 in 10 mM PBS.
Radiolabelling of α2M with 125I and competitive inhibition assay
Radiolabelling of the human protease inhibitor α2M (Sigma–Aldrich) was performed using carrier-free 125I (Amersham Biosciences) by the chloramine T method with a specific activity of approx. 1 mCi/mg of protein [34,35]. Binding of 125I-labelled α2M to S. pyogenes was assessed as described previously [11,15,27]. In competitive inhibition experiments, the binding of 12 ng of 125I-labelled α2M to 1.5×108 c.f.u./300 μl of the S. pyogenes strains A6 and KTL3 grown at 37 °C in THY to stationary phase was determined in the presence of increasing concentrations of unlabelled purified rGRAB and its derivatives. Recombinant GRAB, its derivatives and mutated GRAB were used at concentrations ranging from 6.6 ng/50 μl of PBS up to 100 μg/50 μl of PBS. The iodination of α2M neither influenced the conformational status of the protein as demonstrated by Müller and Rantamäki  nor the GRAB-binding ability as shown by the blot-overlay analysis.
Preparation of spot-synthesized peptides on membranes and analysis of α2M binding
The A region of GRAB (amino acids 34–91) was predicted to be the α2M binding domain. To map the binding site precisely, the amino acid sequence of the A domain was divided into 44 overlapping peptides, consisting of 15 amino acids each, with an offset of one amino acid per peptide spot . The initial sequence (spot 1) was V34DSPIEQPRIIPNGG48. The peptides were synthesized as an array of spots on an aminopegylated cellulose membrane (AIMS Scientific Products, Braunschweig, Germany) as described previously [37,38] and analysed for their α2M-binding activity in a protein overlay assay using native human α2M (Sigma–Aldrich) as a ligand (40 μg/ml α2M). The two α2M binding sites detected by that assay spanned parts of the 15-amino-acid regions P41RIIPNGGTLTNLLG55 and T51NLLGNAPEKLALRN65. These parts were subsequently divided again into synthetic, overlapping peptides of variable length and were chemically synthesized as above. To define the minimal binding motif, the region corresponding to amino acids 36–61 was divided into 68 overlapping synthetic peptides with an offset of one amino acid each. The membrane consisted of eight series of peptides with lengths varying from 8 up to 15 amino acids per spot (Table 2). The second region corresponded to amino acids 45–70 and was divided into 63 overlapping peptides, consisting of seven series of peptides with lengths varying from 9 up to 15 amino acids per spot (Table 2). In addition, peptide spot membranes were generated for systematic mutational analysis. The membranes contained decapeptides of the two proposed binding domains with individual amino acid substitutions at each position of the original peptide sequences PRIIPNGGTL and NAPEKLALRN. Detection of binding was performed by the blot-overlay assay as described previously . Peptide-specific signals were evaluated by densitometric analysis and the relative intensity of the spots was determined using the EASY Win32 analysis software (Herolab).
α2M binds to the ΔA domain of GRAB
The N-terminal part of the A domain, designated ΔA (Figure 2), was predicted to be the preferential binding site for α2M . Based on the previous results, five recombinant GRAB derivatives were constructed and produced as GST fusion proteins to narrow down the binding site for α2M in GRAB (Figure 3A). The ΔA fragment, comprising amino acids 34–67, represents the N-terminal part of GRAB and exhibits 71% identity to the E domain of Protein G . All recombinant proteins were separated by SDS/PAGE (Figure 3B) and functionally analysed for their α2M-binding activity under reducing conditions in blot-overlay assays (Figure 3C). Immunoblots with anti-GRAB IgG antibodies (Figure 3D) and anti-α2M IgG antibodies (Figure 3E) were used as controls. The binding of α2M to purified GST protein was tested negatively by Western-blot analysis and also by using 125I-labelled α2M as a ligand (results not shown). The results revealed that each of the rGRAB derivatives was capable of binding α2M. These results suggest that the α2M-binding motif is located in the ΔA domain of GRAB, whereas the C-terminal part of the A domain and the repeat regions did not play any role in the α2M-binding activity of GRAB. Recombinant GST that was used as a negative control interacted neither with anti-GRAB antibodies (Figure 3D) nor with α2M (Figure 3C). To compare the binding of native and trypsin-modified forms of α2M to the bacterial surface protein GRAB, additional binding assays were performed using α2M (Sigma–Aldrich) and trypsin-pretreated f-α2M (molar ratio α2M/trypsin: 1:2, 1:4 and 1:8) as ligands. The high molar trypsin ratio was used to avoid any residual s-α2M. Western-blot analysis did not reveal any binding of trypsin-pretreated f-α2M to rGRAB, whereas untreated α2M exhibited a strong binding ability (results not shown). α2M separated by SDS/PAGE and analysed by Western blotting using rGRAB as a ligand revealed no binding activity. The results indicate that the interaction between both proteins strongly depends on the conformational status of α2M and only native s-α2M is capable of interacting with the bacterial surface protein GRAB.
The function of the ΔA region of GRAB as an α2M-binding site was confirmed in competitive inhibition assays. The binding of radiolabelled 125I-α2M was assayed using two different S. pyogenes isolates KTL3 (M1 serotype) and A6 (M3 serotype) in the presence of increasing concentrations of rGST-ΔA and rGST-A. Both strains tested positive for α2M binding in previous experiments (results not shown), and GRAB was shown to be the only α2M-binding protein expressed by KTL3 . The results indicated that both rGST-ΔA and rGST-A were capable of inhibiting the binding of 125I-α2M to GAS by approx. 80%, thus confirming the role of ΔA as the binding region for α2M (Figure 3F). The inhibition of not more than 80% is probably due to fusion with GST. With His-tagged rGRAB, more than 97% inhibition could be achieved (results not shown).
α2M binding to spot-synthesized peptides deduced from the A region of GRAB
For a fine mapping of the binding site for α2M, the A region of GRAB was divided into 44 overlapping peptides, each consisting of 15 amino acids, with an offset of one amino acid each. These immobilized peptides were analysed for their ability to bind α2M. The blot-overlay assay revealed two reacting peptide series located between spots 7–10 and spots 17–18 (Figure 4). These two peptides represent amino acid sequences of ΔA at positions 40–57 and 50–65 of GRAB respectively.
Identification of two distinct binding motifs in the ΔA region of GRAB
Based on the results of the initial spot-synthesized peptides, two further sets of membrane-spotted peptides were analysed for α2M-binding activity to elucidate the minimal sequences required for binding (Table 2). For this purpose, two overlapping regions, one comprising amino acids 36–61 and the other comprising amino acids 45–70, were each divided into peptides with lengths varying from 8 up to 15 amino acids, with an offset of one amino acid per spot (Figure 5). This strategy resulted in 68 and 63 peptides respectively. The blot-overlay assay with α2M, identified 24 strongly reacting spots in the first region and 13 strongly reacting spots in the second region (Figures 5C and 5F). The results suggested an eight-amino-acid peptide, IIPNGGTL, as the minimal binding motif in the first region. Since the N-terminal extension by a proline and an arginine strongly enhanced the α2M-binding activity as shown by densitometric analysis, the decapeptide PRIIPNGGTL was proposed as the α2M-binding motif of the first region (Figure 5C). The nine-amino-acid peptide NAPEKLALR and the decapeptide NAPEKLALRN were identified as the minimal α2M-binding motifs of the second region (Figure 5F). C-terminal extension of the decapeptide with the GRAB-sequence-derived glutamic acid residue resulted in complete loss of α2M-binding activity. These results suggested that the two binding motifs contribute substantially to the GRAB–α2M interaction.
Systematic mutational analysis by spot-synthesized peptides
From the results of the previous spot membrane analysis, it was predicted that two binding sites of GRAB are involved in the GRAB–α2M complex formation. In an attempt to locate the amino acids critical for this interaction, every single amino acid of the peptides PRIIPNGGTL and NAPEKLALRN was individually substituted. Membrane-spotted peptides containing the amino acid substitutions were analysed for their ability to bind soluble α2M. Results of the blot-overlay assays revealed that substitutions with glutamic and aspartic acid residues resulted in the complete loss of binding activity in both the binding motifs. In contrast, substitution of each amino acid of the second binding region with a lysine residue resulted in enhanced α2M-binding activity. However, the analysis indicated that the arginine residue at position 42 located in the first motif and the arginine residue at position 64 located in the second motif are crucial for the α2M-binding activity of GRAB (Figures 6B and 6D). Although substitution of the lysine residue at position 60 resulted in a decrease in the α2M binding, this lysine residue was not considered to be crucial for the interaction, since a previous membrane-spotted peptide analysis (Figures 4 and 5F) demonstrated that peptides containing the lysine residue but not the arginine residue (spot 16; Figure 4) are not capable of interacting with α2M.
Effect of site-directed mutagenesis of Arg42 and Arg64 of rGRAB on α2M-binding activity
To confirm the critical role of the arginine residues at amino acid positions 42 and 64 of GRAB in α2M binding, site-directed mutagenesis was performed. Recombinant GRAB encoded by pQGR was selected for individual amino acid substitutions resulting in rGRAB42, rGRAB64 and rGRAB42/64 (Figure 7A). Binding of α2M was first examined using SDS/PAGE (Figure 7B). Results showed that substitution of the arginine residue at position 42 did not affect the binding of α2M to rGRAB42. In contrast, substitution of the arginine residue at position 64 of GRAB resulted in the complete loss of α2M-binding activity under reducing conditions. Binding of α2M to rGRAB42/64 with amino acid substitutions of both arginine residues was also completely abolished (Figure 7D). Immunoblot analysis using anti-GRAB IgG antibodies was used as a control (Figure 7E). Some other bands in Figures 7(C) and 7(D) are probably GRAB aggregates capable of interacting with α2M.
The effect of amino acid substitutions on α2M-binding activity was also investigated by competitive inhibition assays (Figure 8). The recombinant proteins rGRAB, rGRAB42, rGRAB64 and rGRAB42/64 were used for competitive inhibition of α2M binding to the S. pyogenes strain KTL3. Proteins were added separately in increasing amounts from 6.6 ng up to 100 μg per reaction. The competitive inhibition with rGRAB42 confirmed the result of the blot-overlay assay that the substitution of R42G did not change the binding of α2M to GRAB42. Binding of 125I-α2M to GAS was inhibited by rGRAB42 and, therefore, the inhibition was similar to that of the wild-type protein rGRAB. Interestingly, rGRAB64, which did not show any α2M-binding activity under reducing conditions, could inhibit binding of α2M to the bacteria. However, 0.4 nmol (6.25 μg) of rGRAB64 was needed to block nearly 50% of the α2M-binding activity of 1.5×108 c.f.u. of the S. pyogenes strain KTL3, whereas only 0.006 nmol (00976 μg) of wild-type GRAB was sufficient. Substitution of both arginine residues at positions 42 and 64 abolished α2M-binding activity under these conditions, as indicated by the lack of inhibiting activity. Therefore these results suggest that Arg64 is the most crucial amino acid residue in this interaction, followed by Arg42, and both together are crucial for the GRAB–α2M complex formation.
α2M is a plasma protein with a broad spectrum of biological properties. In addition to its well-characterized function as protease inhibitor, α2M can interact with cytokines such as IL-1β (interleukin-1β), IL-2, IL-6 and IL-8 as well as with growth factors like PDGF-β (platelet-derived growth factor-β), TGF-β (transforming growth factor-β) or NGF-β (nerve growth factor-β), depending on its conformational status (s-α2M or f-α2M) [39–44]. The ability of α2M to interact with all four classes of proteases, irrespective of their origin (exo- or endo-genic), with less efficiency compared with the class-specific inhibitors and the high concentration of α2M in human plasma makes α2M part of either a protective mechanism decreasing the concentration of soluble plasma proteases to physiological levels or a mechanism against the invading pathogens by presenting entrapped foreign proteases to antigen-presenting cells and therefore stimulating the host immune response [45,46]. The nature of these interactions and its impact on the infection by pathogens remain unclear and open to question. Especially, the ability of α2M to interact with a wide spectrum of proteases has been, in turn, used by invading bacteria such as S. pyogenes to protect their own surface structures from degradation during the infection process [11,12].
The recruitment of non-protease-complexed s-α2M to the GAS surface is mainly mediated by the low-molecular-mass cell-wall-attached protein GRAB. The advantages of this interaction to the bacterial invader are manifold. Binding a broad-range protease inhibitor such as α2M provides the bacterium with an important tool to minimize protease degradation of, for example, antiphagocytic surface proteins or to recruit inflammatory mediators to the bacterial surface to enhance local inflammation for a more efficient tissue spreading. Moreover, this mechanism may reduce the concentration of α2M in the surrounding environment maintaining protease activity and therefore tissue degradation. The presence of GRAB and its recruitment of α2M on to the bacterial surface facilitates bacterial survival and influences the infection process . The fact that GRAB protein contributes to bacterial virulence led to our investigation of the nature of this interaction to predict the structure and stability of the complex or to identify the proteins engaging similar strategies for α2M assembly. The present study provides evidence that the recruitment of α2M via the surface protein GRAB is mediated by two distinct but adjacent motifs (PR42IIPNGGTL and NAPEKLALR64N), which are both located in the ΔA region (amino acids 34–67). Each domain contains a single basic charged arginine residue with distinct roles in the binding facility. Individual substitution of the particular residues influenced the binding in a divergent manner. Substitution of Arg42 had no detectable influence on α2M-binding activity. In contrast, substitution of Arg64 almost resulted in loss of α2M-binding activity of GRAB. Although it was not possible to detect any interaction under reducing conditions, the protein displayed a reduced affinity for α2M under physiological conditions. The amino acid substitutions of both residues (Arg42 and Arg64) emphasized the effect of substitution of Arg64. The results obtained strongly indicate that Arg64 is the amino acid of GRAB most critical for the interaction with α2M. As amino acid substitutions of Arg42 and Arg64 almost completely abolished α2M-binding affinity under native conditions, the contribution of the arginine residue at position 42 to complex formation became visible. It is possible that both arginine residues play a role in tertiary-structure formation, assembling the two adjacent motifs in an appropriate position to bind α2M with high affinity or just enhance the binding because of their basic charged residues. It is probable that the high affinity of the binding site NAPEKLALR64N for α2M is masking the contribution of Arg42. As a result, binding of α2M to rGRAB42 is similar to wild-type rGRAB. It is probable, but not confirmed, that the flanking amino acids TLTNLL contribute to the high affinity of the GRAB–α2M interaction as suggested by spot-membrane analysis. A synthetic peptide IIPNGGTLTNLLGNA (Figure 4, spot 10) that did not contain the critical arginine residues was not capable of competitively inhibiting the binding of α2M to GAS (results not shown).
Interpretation of these results indicate that the α2M-binding activity of GRAB depends on the interplay of the protease inhibitor α2M with two adjacent motifs in GRAB. One of these two motifs contains the Arg64 and might be directly contributing to the interaction and the other motif, containing Arg42, may contribute to high-affinity binding after correct folding of the complex structures or its basic charge. In view of the high sequence similarity between GRAB protein and Protein G, it is tempting to speculate that both proteins have a common ancestor and, accordingly, a closely related mechanism for α2M recruitment. The results of the present study indicate the involvement of charged motifs in the ΔA region of GRAB in α2M binding and, therefore, in virulence of GAS. These motifs might represent targets for developing new intervention strategies.
We thank S. Daenicke for expert assistance with SPOT-peptide synthesis, N. Janze for technical support and Dr D. Zähner (all from the German Research Center for Biotechnology) for helpful and constructive discussions. We are grateful to M. Rasmussen (Department of Cell and Molecular Biology, Lund University, Lund, Sweden) for providing the GAS strain KTL3.
Abbreviations: ΔA, N-terminal part of the A domain; α2M, α2-macroglobulin; f-α2M, electrophoretically fast form of α2M; s-α2M, electrophoretically slow form of α2M; BAEE, Nα-benzoyl-L-arginine ethyl ester hydrochloride; GAS, group A streptococci; GST, glutathione S-transferase; (r)GRAB, (recombinant) Protein G-related α2M-binding protein
- The Biochemical Society, London