INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Group A streptococcus (Streptococcus pyogenes) is the cause of tonsillitis and impetigo and also causes severe diseases, including necrotizing fasciitis and the streptococcal toxic shock syndrome (2). In addition, group A streptococcal infections are sometimes complicated by one of the postinfectious sequelae, rheumatic fever and glomerulonephritis.

The M proteins of group A streptococci are major virulence factors that confer resistance to phagocytosis (14). The N-terminal sequences of the M proteins are highly variable, giving rise to the existing ~100 serotypes. Protection against phagocytosis is usually serotype specific. It should be noted that a large fraction of all group A streptococcal strains encode three proteins with structural features typical for M proteins (12, 18). These proteins, designated Mrp, Emm, and Enn, are encoded by adjacent genes on the chromosome and are regulated by a common positive regulatory gene element now designated mga (7, 20). At least two of the gene products (Mrp and Emm) from a single strain can contribute to resistance against phagocytosis, although Emm appears to be more important (22, 32). Strains with three genes encoding M-like proteins differ from those with a single gene in that they usually express opacity factor (OF), a lipoproteinase that is both surface bound and secreted (25, 26). Recent evidence indicates that OF contributes to group A streptococcal virulence (4).

Streptococcal mutants lacking M protein(s) are readily phagocytosed. However, the antiphagocytic property can be restored by complementation with the homologous M protein (19). Moreover, there is evidence that introduction of DNA encoding heterologous M proteins can provide phagocytosis resistance (5, 23, 27). For example, phagocytosis resistance was restored following integration of the gene encoding the M5 protein in the chromosome of an M protein-negative isolate derived from a serotype 24 strain (5). In contrast, the Emm4 protein (Arp4), derived from an OF⁺ type 4 strain, was unable to complement the phagocytosis resistance of an OF type 6 strain deleted of its M protein-encoding gene, leading to the conclusion that the expression of Emm4 is not sufficient to confer phagocytosis resistance (8). On the other hand, gene inactivation experiments suggest that Emm4 and similar proteins have antiphagocytic properties (22, 32). Taken together, these data suggested to us that the ability of M and Emm proteins to provide the antiphagocytic property might be limited by the genetic background of the strain in which it is expressed. To resolve this issue, we have analyzed the ability of different proteins to confer phagocytosis resistance in heterologous strains. We find that the ability of M proteins to provide protection against phagocytosis is indeed highly restricted by the genetic background of the strain. This finding suggests that phagocytosis resistance may require cooperation between M proteins and other strain-specific streptococcal component(s).

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains and plasmids. Strain M5 Manfredo was kindly provided by Michael Kehoe, University of Newcastle-upon-Tyne. The type 22 strain AL168 has been described previously (28). The type 6 strain JRS4 was from June R. Scott (Emory University, Atlanta, Ga.). The plasmid pLZ12Spec is an Escherichia coli-Streptococcus shuttle vector carrying a spectinomycin resistance marker (8), whereas derivatives of the temperature-sensitive suicide vector pJRS233 (21) used here carry resistance markers for erythromycin and kanamycin. Deletion of the emm5 gene in M5 Manfredo was achieved by homologous recombination by using a derivative of pJRS233 (10). This strain is referred to as M5. The mrp22 gene of strain AL168 was inactivated by insertional mutagenesis by using the conjugative transposon Tn916, whereas inactivation of the emm22 gene in the mrp22 negative strain was performed by homologous recombination by using the pJRS233 vector as described elsewhere (32).

Culture conditions. Streptococci were grown in Todd-Hewitt broth (TH), in 5% CO₂, at 37°C. Streptococci transformed with pLZ12Spec, or derivatives thereof, were selected on medium supplemented with spectinomycin at 100 mg/liter. For isolation of streptococci transformed with derivatives of pJRS233, erythromycin and kanamycin were used at concentrations of 1.0 and 200 mg/liter, respectively. E. coli LE392 was grown in Luria-Bertani medium, supplemented with spectinomycin at 20 mg/liter, if transformed with derivatives of pLZ12Spec, or with erythromycin at 500 mg/liter and kanamycin at 50 mg/liter, if transformed with derivatives of pJRS233.

Binding of radiolabeled ligands to group A streptococci. For binding studies, overnight cultures of streptococci were harvested by centrifugation, washed twice in PBSA (0.15 M NaCl, 0.03 M phosphate, 0.02% sodium azide; pH 7.2), and were then resuspended in PBSA supplemented with 0.05% Tween 20 (PBSAT). The binding of radiolabeled proteins to bacteria, at different dilutions, was measured in a total volume of 250 µl of PBSAT. After incubation for 1 h at 20°C, the samples were centrifuged (4,000 × g, 10 min), the supernatant was discarded, and the remaining pellet was washed with 2 ml of PBSAT. After centrifugation and subsequent removal of the supernatant, the radioactivity associated with the pellet was measured in a gamma counter.

Bactericidal assay. Resistance to phagocytosis was analyzed in a bactericidal assay as described elsewhere (15, 21, 32). Briefly, an overnight culture of bacteria in TH, supplemented with 100 mg of spectinomycin per liter, was diluted 1:50 in TH and grown without agitation to an A₆₂₀ of 0.15 at 37°C. The bacteria were then diluted by 10⁴ in TH, and 100 µl of the suspension, containing ~50 CFU, was added to 1.0 ml of human blood supplemented with sodium heparin at 20 U/ml in acid-cleaned glass tubes (100 by 10 mm). The tubes were incubated with rotation at 37°C for 3 h. For some experiments, heparin was replaced with hirudin at 140 U/ml (Calbiochem). For experiments with hirudin, 2.2-ml polypropylene tubes were used instead of glass tubes, and 250 µl of fresh hirudinized blood was mixed with ~10 CFU of log-phase bacteria in TH. CFU in the samples were counted before and after incubation by using the pour plate method.

Emm protein and OF-encoding shuttle plasmids. Construction of pJRS264, an emm4 (arp4) gene containing derivative of pLZ12Spec has been described previously (8). A 2.1-kb EcoRI/SphI fragment of pKEJ1 harboring the emm5 gene (11) was inserted into pLZ12Spec, resulting in plasmid pLZemm5. For construction of pLZemm22, a 2.2-kb HindIII fragment of Sir2202 carrying the emm22(Sir22) gene (29), was ligated with HindIII-digested pLZ12Spec to generate pLZemm22. To construct pLZemm6, the emm6 gene was amplified by PCR by using chromosomal DNA from strain JRS4 as template and the synthetic oligonucleotides 5'-GGCTGGATCCTTAATAGCATTTAGGTC-3' and 5'-GCTAGGCATGCATAAAAGAGAGAACCG-3' as primers. Restriction sequences for BamHI and SphI were introduced by the primers. The amplified fragment was digested with BamHI and SphI and was subsequently ligated with BamHI/SphI-digested pLZ12Spec. To construct pLZsof22, the gene encoding opacity factor from type 22 strain AL168 was amplified by PCR by using the synthetic oligonucleotides 5'-GTCCCATGGCGGATCCTTAATTTTTATCTCACC-3' and 5'-GCAAGCTTGGCATGCTTAAA-GCCAAAGGCTTAGGG-3' as primers. Primer sequences were selected according to the published sequence for sof22 (25). The resulting fragment was digested with BamHI and SphI, for which restriction sequences were introduced by the primers and subsequently ligated with pLZ12Spec to generate pLZsof22.

Chromosomal replacement of emm5 by emm22 in strain M5 Manfredo. Integration of the emm22 gene into the chromosome of the M5 strain was achieved by homologous recombination. A 1,356-bp DNA fragment with the entire emm22 gene was amplified by PCR with the synthetic oligonucleotides 5'-CGAAGGATCCAAAAAAAGAGGAAGCCCCTTCC-3' and 5'-GGTCTGCATGCGATTGTTAGTTACTTAGCC-3' as primers and with chromosomal DNA from the AL168 strain as a template. The PCR fragment was digested with BamHI and SphI, recognition sequences for which had been introduced through the primers. The fragment was subsequently used to replace the Km2 cassette in a derivative of pJRS233 containing this cassette flanked by sequences surrounding the emm5 gene (10). In the resulting plasmid, the emm22 gene was flanked by two sequences (1,200 and 1,000 bp, respectively) that are derived from sequences upstream and downstream of the emm5 gene in the chromosome of M5 Manfredo. This plasmid was electroporated into M5. The transformed bacteria were first grown for 48 h at 30°C in the presence of erythromycin. Individual erythromycin-resistant bacteria were then used to inoculate 10 ml of TH and were incubated at 37°C, a temperature that does not allow pJRS233 or its derivatives to replicate in streptococci (21). Mutants in which Km2 in M5 had been replaced by emm22 were selected by screening streptococci able to grow at 37°C for the loss of kanamycin resistance. Integration of emm22 in the correct chromosomal location was verified by PCR.

Recombinant DNA techniques. Standard recombinant DNA techniques were used. Ligase and restriction enzymes were purchased from Promega. Transformation of E. coli was performed by the CaCl₂ method. Electroporation of streptococci was carried out as described previously (3).

Other methods and reagents. Human C4BP (complement factor 4b-binding protein) was a kind gift from Björn Dahlbäck, Malmö General Hospital, Lund University, Malmö, Sweden, and FHL-1 (complement factor H-like protein 1) was kindly provided by Peter Zipfel, Bernhard-Nocht Institute of Tropical Medicine, Hamburg, Germany. Polyclonal human serum immunoglobulin A (IgA) was from Cappel. Human fibrinogen and fibronectin was from Sigma. Proteins were labeled with ¹²⁵I by using the chloramine-T method (6). Measurement of lipoproteinase activity was made according to the method of Maxted et al. (16).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Construction of strains expressing homologous and heterologous M proteins. To study the ability of different M proteins to confer phagocytosis resistance to strains with different background, the well-defined M5 and M22 systems were used. The M5 Manfredo strain is OF and expresses a single M protein (17, 33) (Fig. 1), whereas the OF⁺ M22 strain AL168 contains an mga regulon encoding three M-like proteins (32) (Fig. 1), of which at least two (Mrp and Emm) are expressed on the streptococcal surface and confer phagocytosis resistance. The genes encoding M5 and Emm22 were cloned separately in the shuttle vector pLZ12Spec. The resulting plasmids, designated pLZemm5 and pLZemm22, respectively, were used to transform M5 (10), an M5-deficient derivative of M5 Manfredo, or AL168(mrp emm), a phagocytosis-sensitive derivative of AL168 lacking expression of the Mrp22 and Emm22 proteins (32). As a result, four strains were obtained, in which M5 and Emm22 were expressed in either a homologous or a heterologous background. The four strains were designated M5/pLZemm5, M5/pLZemm22, AL168(mrp emm)/pLZemm5, and AL168(mrp emm)/pLZemm22.

View larger version (18K): [in this window] [in a new window]   FIG. 1.   Schematic representation of the mga5 and mga22 regulons and the mutants used in this study. (A) Organization of the mga regulons of the type 5 strain M5 Manfredo and the type 22 strain AL168. Replacement of the emm5 gene with a gene encoding kanamycin resistance (km2) by homologous recombination gave rise to strain M5 (10). In the AL168(mrp emm) derivative, the mrp22 gene has been disrupted by insertion of the conjugative transposon Tn916, whereas a large part of the emm22 gene has been replaced by a kanamycin casette (32). The mga gene encodes the multigene activator of group A streptococcus, and the mrp, emm, and enn genes encode proteins in the M protein family, whereas the scpA gene encodes the streptococcal C5a peptidase. The regions in the proteins marked A, B, and C are distinct domains containing three to five different repeat units. (B) Organization of the mga regulon of the M5/emm22 derivative of the M5 Manfredo strain, in which the emm5 gene has been replaced by the emm22 gene. Protein ligands that are relevant for this study and their binding sites in M5 (Fg [fibrinogen] and FHL-1) and Emm22 (IgA and C4BP) are indicated.

Binding properties of streptococcal strains expressing homologous and heterologous M proteins. To analyze whether the surface expression of the M5 and Emm22 proteins was similar in the different bacterial backgrounds, we determined the ability of the complemented streptococci to bind radiolabeled ligands known to specifically interact with the M5 and Emm22 proteins, respectively (Fig. 2). Surface expression of the M5 protein was monitored by using the capacity of the streptococci to bind FHL-1 and fibrinogen, both of which bind to M5, but not to Emm22 (10, 13). Similarly, C4BP and IgA, both of which interact with Emm22 but not with M5 were used to determine the expression level of the Emm22 protein (11, 31). The two strains expressing M5 in homologous and heterologous backgrounds showed similar FHL-1 and fibrinogen-binding capacity. The slight differences noted are unlikely to reflect different M5 protein expression levels, particularly since for fibrinogen the binding capacity of M5/pLZemm5 and AL168(mrp emm)/pLZemm5 was somewhat reduced as compared to M5 Manfredo, whereas for FHL-1 the binding capacity of M5/pLZemm5 was slightly improved over the other two strains (Fig. 2A and B). The two strains expressing Emm22 in homologous and heterologous background showed C4BP- and IgA-binding levels that were practically identical (Fig. 2C and D). No difference in the binding patterns was noted between complemented isolates that had been cultured with or without antibiotic selection (data not shown), suggesting that the plasmids were not highly unstable. Finally, the complemented strains all possessed the hair-like surface structures that are typical for M proteins (30), as determined by transmission electron microscopy (data not shown).

View larger version (30K): [in this window] [in a new window]   FIG. 2.   Binding of purified ligands to streptococci expressing the M5 and Emm22 proteins. The binding of 125I-labeled ligands to the indicated group A streptococcal strains was measured as a function of bacterial number. Fibrinogen (A) and FHL-1 (B) are known to interact specifically with the M5 protein, whereas C4BP (C) and IgA (D) bind to the Emm22 protein. The data presented are based on three independent experiments with triplicate samples. The variation was <5% of the binding, and therefore error bars have not been introduced.

Ability of M proteins to confer antiphagocytic property in a homologous or heterologous background. The ability of the four complemented strains to resist phagocytosis was compared with the parental M protein-expressing strains, and with the corresponding M negative strains, in the bactericidal assay in whole blood by using heparin as the anticoagulant (15). Ten independent experiments with blood from six different blood donors were carried out (Fig. 3 and Table 1). There was a donor-dependent interexperimental variation, but the outcome of each individual experiment showed a similar pattern, namely, that M5 and Emm22 provided phagocytosis resistance in the homologous but not in the heterologous background. Similar results were obtained with hirudin, a specific inhibitor of thrombin, as the anticoagulant (data not shown). Together, these data show that the ability of M proteins to provide protection against phagocytosis is restricted by the genetic background of the strain.

View larger version (20K): [in this window] [in a new window]   FIG. 3.   Growth of group A streptococcal strains in human blood. The multiplication factor indicated on the ordinate represents the factor of the increase in number of CFU during a 3-h incubation in rotating tubes. (A) Experiments with strains expressing the M5 protein or no M protein (strain M5). (B) Experiments with strains expressing the Emm22 protein or no M protein (AL168mrpemm). The data are also shown in Table 1.

View larger version (14K): [in this window] [in a new window]   FIG. 4.   Binding of host ligands to an M-negative type 5 expressing Emm22 and OF22. The emm5 gene of M5 Manfredo was replaced by emm22, resulting in strain M5/emm22. (A) This strain expresses Emm22 on its surface, as shown by its ability to bind C4BP and IgA. The gene encoding the opacity factor (sof) was cloned from the type M22 strain AL168 and inserted in pLZ12Spec. The resulting plasmid was then introduced into M5/emm22, producing strain M5/emm22(pLZsof22). (B) The ability of M5/emm22(pLZsof22) to bind fibronectin, a ligand for OF, was compared with the fibronectin-binding capacity of the M5 Manfredo, M5, and M5/emm22 strains.

The inability of M proteins to confer the antiphagocytic property to heterologous strains is not limited to the M5 and M22 systems. To analyze whether strain-specific protection against phagocytosis is limited to the M5-M22 system, M5 and AL168(mrp emm) were complemented with the emm6 gene cloned in pLZ12Spec. Binding analysis with fibrinogen and FHL-1 showed that the resulting heterologous strains expressed the M6 protein at levels comparable to those of the wild-type strain (data not shown). The complemented strains were analyzed in the bactericidal assay. Expression of the M6 protein in M5 supported growth in blood, whereas its expression in AL168(mrp emm) did not (Table 2). It has been shown previously that expression of Emm4 (Arp4) failed to protect M6-deficient streptococci against phagocytosis (8). In agreement with this finding, M5 streptococci complemented with the emm4 gene on the pLZ plasmid (pJRS264) were readily phagocytosed (Table 2). In contrast, when pJRS264 was used to complement AL168(mrp emm), the resulting isolate became resistant against phagocytosis (Table 2), showing that the Emm4 protein does indeed have the functional characteristics of an M protein.

The opacity factor is unable to restore the antiphagocytic property of Emm22 in a heterologous background. The restriction in the ability of M proteins to provide protection against phagocytosis in different genetic background suggests that other strain-specific components affect the ability of M proteins to confer phagocytosis resistance. One such protein is the OF. To analyze whether coexpression of OF allows Emm22 to confer phagocytosis resistance in a heterologous background, we introduced the plasmid pLZsof22 encoding the OF22 protein into M5/emm22. The resulting strain, M5/emm22(pLZsof22), had lipoproteinase activity (data not shown) and bound fibronectin (Fig. 4B), a characteristic property of OF (25). However, expression of OF22 did not rescue M5/emm22(pLZsof22) streptococci in the bactericidal assay (Table 2). Other factors that may affect the strain-specific restriction must therefore be analyzed.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Our observations clearly suggest that M proteins are restricted in their ability to provide group A streptococci with resistance against phagocytosis. This conclusion is based on the finding that the M5 and M6 proteins, originating from isolates of the OF lineage, conferred phagocytosis resistance in the background of an OF strain of type 5 but failed to do so in the OF⁺ strain of type 22. Similarly, the Emm4 and Emm22 proteins, both originating from OF⁺ strains, only conferred phagocytosis resistance to the OF⁺ strain.

Several explanations for the surprising phenomenon described here can be imagined. A trivial reason would be that M protein encoding plasmids are unstable in some strains. However, several lines of evidence indicated that the plasmids used here were not lost during the course of the experiments. More importantly, such an explanation does not account for the failure of M5 bacteria carrying the emm22 gene inserted into the chromosome to grow in blood. A second alternative could be that certain M proteins are not properly surface expressed in a heterologous background. Our data do not support this explanation, since the ability of the complemented strains to bind different ligands were similar to that of the parent strains. However, it seems possible that there exist lineage specific systems necessary for correct folding, processing, or cell surface distribution of M proteins and that these systems influence the ability of M proteins to confer phagocytosis resistance. For example, it has been demonstrated that the streptococcal cysteine proteinase cleaves proteins belonging to the M protein family, thereby providing a mechanism for modulating the surface expression of these proteins (1, 24). It remains to be investigated whether such protease activity can show the substrate specificity required to explain the present findings. The binding data with C4BP and FHL-1, which bind to the N-terminal surface-exposed portion of Emm22 and M5, respectively (10, 11), do not support this explanation since the complemented strains had the expected binding properties, showing that their M proteins are not processed in any extensive way. The failure of emm5 to provide the AL168(mrp emm) strain with the capacity to survive in blood could partially be explained by the requirement for expression of both the mrp22 and emm22 gene products. However, this explanation is not sufficient, since emm22 by itself provides the mutant strain with phagocytosis resistance (Fig. 3) (32).

Until now, there has been no evidence that M proteins would require additional factors to exert their antiphagocytic effect. However, this may well be the explanation behind the present findings. The opacity factor is one such factor, but the finding that simultaneous expression of Emm22 and OF22 did not provide the OF M5 strain with the antiphagocytic property indicates that OF is not a likely candidate. Moreover, a requirement for OF production would not explain the failure of the M5 and M6 proteins to provide protection in the OF⁺ background. Other factors should therefore be evaluated. Among known surface proteins, the T proteins are perhaps the most interesting candidates, since they appear to be variable and since expression of a specific T protein is usually associated with expression of certain M serotypes (9).

Regardless of the molecular explanation, the present findings show that there are functional barriers between group A streptococci of different serotypes. In addition, the data imply that great care must be taken when interpreting functional data derived from experiments involving the introduction of M and Emm protein-encoding genes into different strains of group A streptococci.