|Home | About | Journals | Submit | Contact Us | Français|
Resistance to phagocytosis is a hallmark of virulent Streptococcus pyogenes (group A streptococcus). Surface-bound C5a peptidase reduces recruitment of phagocytes to the site of infection, and hyaluronic acid capsules and/or the M protein limit the uptake of streptococci. In this study the relative impact of M and M-like proteins and the C5a peptidase on the virulence of a serotype M49 strain was assessed. The capacities of isogenic strains with an insertion mutation in emm49; with a deletion mutation in scpA49 (C5a peptidase gene); and with a deletion that removes all three M-like genes, mrp49, emm49, and enn49, to colonize mice and resist phagocytosis were compared. Experiments confirmed results obtained in an earlier study, which showed that the M49 protein was not required for in vitro resistance to phagocytosis, and also showed that the M protein was not required for colonization of mice. Failure to produce all three M-like proteins, M49, Mrp, and Enn49, significantly reduced the ability of these streptococci to resist phagocytosis in vitro but did not significantly alter the persistence of streptococci on the oral mucosa. In vitro experiments indicate that M+ streptococci are phagocytized by polymorphonuclear leukocytes that have been activated with phorbol-12-myristate 13-acetate or recombinant human C5a. This observation may explain the finding that expression of M49 protein is not essential for short-term colonization of the mouse oral mucosa.
Although M protein and its antiphagocytic function have been the focus of decades of research, group A streptococci are now known to express several proteins on their surface. Moreover, as more serotypes are examined it has become apparent that some are more dependent on the hyaluronic acid capsule or on other unidentified gene products for resistance to phagocytosis (6, 29). The need to resist phagocytosis is evident, as streptococci have evolved overlapping mechanisms to avoid phagocytic defenses. The vir regulon of the serotype M49 streptococcus strain used in this study contains the emm49, mrp49, and enn49 genes, which encode the M49 protein and immunoglobulin G (IgG) and IgA binding proteins, respectively (7). Insertion mutations in any one of these genes partly reduced the capacity of this streptococcus to resist phagocytosis by whole blood or purified polymorphonuclear leukocytes (PMNs) (21). This finding prompted Podbielski et al. to suggest that resistance to phagocytosis depended on the cooperation of all three M-like proteins (21).
The early inflammatory response to streptococcus is complex and poorly understood. Activation of the alternative complement pathway in response to M+ streptococci is limited and little C3b is deposited on their surface (11). In contrast M− streptococci efficiently activate the alternative complement pathway and become circumferentially covered with C3b. The precise mechanism by which M protein limits activation of the alternative complement pathway and deposition of C3b is not known. In ischemia-induced models of inflammation C5a can be detected within minutes of the initial inflammatory insult. The inflammatory response is further amplified by subsequent accumulation of interleukin 8 and other cytokines (10). Both group A and group B streptococci express a C5a peptidase on their surface (8, 20). These enzymes are highly specific for C5a (30) and cleave the chemotaxin at its PMN binding site (5). The group A peptidase (SCPA) was shown to retard infiltration of granulocytes into the peritoneum (19) and subdermal sites of infection (13). Mutations in the peptidase gene (scpA) or immunization of mice with purified SCPA reduced the capacity of streptococci to colonize throats of mice following intranasal infection (12). Increased clearance of M+ SCPA− streptococci from the subdermal and oral mucosal sites of infection was unexpected. These streptococci are resistant to phagocytosis in vitro, i.e., when mixed with fresh human blood from a nonimmune donor. A possible explanation for these findings is that activated phagocytes are able to engulf M+ streptococci. In infectious foci containing M+ SCPA− bacteria, C5a may accumulate, recruit, and activate phagocytes more rapidly than in those containing M+ SCPA+ streptococci. Experiments reported here test this model and further examine the role of M-like proteins in resistance to phagocytosis in vitro.
Streptococcal strain CS101 is a spontaneous streptomycin-resistant derivative of a serum opacity-positive (OF+) serotype M49 strain. Strain MJ3-15 is strain CS101 with an internal in-frame deletion in the SCPA49 gene. CS101::pSF152 is an emm insertion mutant derived from strain CS101 and was provided by A. Podbielski (Institute of Medical Microbiology, University Hospital, Aachen, Germany). Streptococci were cultured in Todd-Hewitt broth (Oxoid, Basingstoke, United Kingdom) supplemented with 2% neopeptone (THB-neo) or 1% yeast extract (THY; Gibco, Pasley, United Kingdom) or on sheep blood agar. In some experiments streptococci were grown in culture medium containing streptomycin (200 μg/ml) or spectinomycin (60 μg/ml). Escherichia coli ER1821 (New England Biolabs, Inc., Beverly, Mass.) was used as the recipient for the thermosensitive suicide vector, plasmid pG+host5. pG+host5 was obtained from Appligene, Inc., Pleasanton, Calif. E. coli DH5α containing plasmid pMH109 and ER1821 containing plasmid pG+host5 were grown in Luria-Bertani broth containing chloramphenicol (10 μg/ml) and erythromycin (Erm; 300 μg/ml), respectively.
A 1.7-kb fragment of scpA49 containing the Mga binding site and promoter region was produced by PCR with primers scpA49For23 (5′ GGGGGG GGATCC TGTAACGGTGCAATAGAC 3′) and scpA49Rev1813 (5′ GGGGGG CCGCGG GGGTGCTGCAATATCTGGC 3′). Underlined nucleotides correspond to scpA49 sequences with coordinates nt 23 and 1813, respectively, and boldface nucleotides correspond to BamHI and SacII recognition sites, respectively. The amplified product was digested with BamHI and SacII and ligated into pG+host5. The resulting plasmid, pJCS17, containing 1.7 kb of scpA, was double cut with EcoRI and BamHI. The cat gene was removed from pMH109 following digestion with EcoRI and BamHI restriction enzymes and then cloned into plasmid pJCS17. One recombinant, plasmid pJCSC18, containing 1.7 kb scpA and 1.0 kb cat, was used in the next step. A 1.5-kb fragment of DNA containing the mga-mrp promoter sequence was produced by PCR with primers mgaFor1344 (5′ GGGGGG GTCGACGCTTTTGTTT TTCAGAGAC 3′) and mrpRev214 (5′ GGGGGG GAATTC ACTTTCTCAGTGAGTA GTG 3′). Underlined nucleotides correspond to mga and mrp sequences with coordinates nt 1344 and 214, respectively, and boldface nucleotides correspond to SalI and EcoRI recognition sites, respectively. The amplified mga-mrp fragment contains the mrp promoter. The PCR product was digested with SalI and EcoRI and ligated to pJCSC18. This ligation placed the cat gene under control of the mrp promoter. The resulting plasmid, pJCSCM6, containing mga-cat-scpA, was linearized with KpnI and electroporated into CS101 recipient cells. Transformants that resulted from a double crossover recombination were selected for chloramphenicol resistance (Cmr) and erythromycin sensitivity (Erms). Cmr Erms colonies were purified and confirmed to have the deletion by PCR.
One mutant, MJY1-3, was further analyzed. PCR primers corresponded to mga and scpA sequences outside the inserts carried by plasmid pJCSCM6 and corresponded to the cat sequence (13). The mgaFor1344, scpA49Rev1813, and mgaFor977 (5′ TCCTTAATAT GGTTCATACGG 3′) primers were specific for chromosomal sequences. The catRev753 primer (5′ GCGGTAAATAT ATTGAATTACC 3′) was specific for cat. Other primers used for the analysis of Δmrp-emm-enn mutants were emm49up187, SOR+ MRev, enn49up1598, and scpRev831 (13). The mgaFor40, mgaRev474 (18), orfxFor1201 (5′ AAAGGAGTAAAATTAACTG 3′), and orfxRev2414 (5′ TTTGACTACGATC TGTTC 3′) primers were also used to characterize deletion mutants. Taq DNA polymerase was obtained from Promega (Madison, Wis.).
PCR was used to confirm that mrp, emm, and enn genes were replaced by the cat gene in the chromosome of this strain. If the correct gene replacement had occurred primers would produce a PCR product of 2.6 kb. DNA from strain MJY1-3 produced a PCR product of this size. As expected wild-type CS101 streptococci did not yield a PCR product (data not shown). To confirm the boundaries of the Δmrp-emm-enn deletion, additional PCRs were done with the primers emmFor187 and emmRev1224 and ennFor191 and Rev831 (scpA promoter). As expected no PCR products resulted when DNA from strain MJY1-3 was amplified with these primers (data not shown). Moreover, amplification of MJY1-3 DNA between mga and scpA by using mgaFor977 and scpARev2322 primers produced the predicted 5-kb PCR product; whereas, amplification of DNA from the parent culture did not yield a PCR product because the distance between the mga and scpA genes is too great, approximately 11 kb (data not shown).
RNA was extracted from log-phase cultures of strains CS101 and MJY1-3 that were grown in THB-neo to an optical density at 560 nm (OD560) of 0.25 as previously described (3). Blots were hybridized at 42°C in a 50% formamide buffer. The 1.5-kb mga probe was produced by PCR with the mgaFor977 and mgaRev2014 primers. The 1.0-kb emm probe was generated by PCR with the emm49up187 and SOR+ MRev primers. The 1.2-kb enn probe was produced by PCR with the enn49up1598 and scpRev831 primers. The 1.7-kb scpA probe was generated by PCR with the scpFor23 and scpRev1813 primers. Probes were purified and labeled with [α-32P]dATP by the random primer method (NEN-Dupont, Boston, Mass.). The amount of total RNA added to each lane was equilibrated by first comparing the amount of rRNA in each preparation on ethidium bromide-stained gels.
Protein extracts were obtained from streptococci that were grown in 100 ml of THY at 37°C in a 5% CO2 atmosphere overnight. Cells were pelleted, washed twice with 5 ml of cold TE (50 mM Tris-Cl [pH 8.0], 1 mM EDTA) buffer containing 1 mM phenylmethylsulfonyl fluoride, and suspended in a solution containing 1 ml of TE-sucrose buffer (50 mM Tris-Cl [pH 8.0], 1 mM EDTA, 20% sucrose), 100 μl of lysozyme (100 mg/ml in TE-sucrose), and 50 μl of mutanolysin (5,000 U/ml in 0.1 M K2HPO4 [pH 6.2]). The mixture was rotated at 37°C for 2 h and then centrifuged for 5 min at top speed in an Eppendorf centrifuge. Electrophoresis and Western blot analysis were performed as described previously (13). The rabbit anti-serum against an M49 synthetic peptide was kindly provided by J. B. Dale (Department of Medicine, University of Tennessee, Memphis). The goat anti-rabbit antibody-alkaline phosphatase conjugate was obtained from Sigma (St. Louis, Mo.).
Streptococci were prepared for flow cytometry as previously described (16, 25). In brief, they were precultured in THY at 37°C and 10% CO2 for 18 h. A 1-ml aliquot was transferred to 10 ml of prewarmed THY and incubated to an OD600 of 0.38 to 0.42. Bacteria were then recovered by centrifugation, washed twice with phosphate-buffered saline (PBS), and resuspended in 1 ml of PBS. An intracellular nontoxic vital dye, biscarboxyethyl-carboxyfluorescein-pentaacetoxy-methylester (BCECF-AM; Boehringer, Mannheim, Germany), was added to a final concentration of 1 mmol/liter to the suspension of streptococci. After a 30-min incubation at 37°C, the now green fluorescence-labeled bacteria were sonicated to disrupt the streptococcal chains. The labeled bacteria were washed three times with PBS and then immediately used for the phagocytosis assay.
Prior to the last bacterial washing step, 1 ml of heparinized (10 IU/ml) whole blood from healthy human donors that contained approximately 2 × 106 PMNs was incubated at 37°C in an Eppendorf Thermomixer (Eppendorf, Hamburg, Germany) at 1,000 rpm for 15 min. After incubation, 107 BCECF-AM-labeled bacteria in 100 μl were removed at 0, 15, 30, and 60 min, immediately mixed with 2 ml of ice-cold lysis buffer (Becton-Dickinson, Heidelberg, Germany), and kept in ice until analysis.
In order to assess the oxidative burst during the phagocytosis process, unlabeled bacteria were cultured and incubated in heparinized whole blood under identical conditions as described above. Dihydro-rhodamine 123 (DHR) was added to a final concentration of 10 mg/liter to the blood. This primarily nonfluorescent dye becomes fluorescent upon oxidation to rhodamine during the respiratory burst of activated PMNs (24, 28).
In some experiments phorbol-12-myristate-14-acetate (PMA; Sigma, Deisenhofen, Germany) and/or CD11b monoclonal antibody (MAb) (Becton Dickinson) was added to heparinized whole blood 5 to 10 min prior to the addition of the bacteria. The final concentrations were 20 ng/ml for PMA and 15% (vol/vol) for the CD11b MAb. Leukocytes were isolated from the lysis solution by centrifugation and were washed three times in PBS prior to analysis by flow cytometry. Flow cytometry was performed with a FACScan flow cytometer with Cellquest software (both from Becton Dickinson). The instrument settings were as follows: forward scatter (FSC) threshold set at 52, detector set at E00, 338, 528, and 560 for FSC, sideward scatter (SSC), fluorescence 1 (FL1), and FL2, respectively. Linear parameters were used for FSC and SSC, and logarithmic parameters were used for FL1 and FL2. PMNs were selectively analyzed by gating them according to their relative size (FSC) and granularity (SSC). The association of PMNs with the BCECF-AM-labeled streptococci and on the oxidative burst of the PMNs induced by unlabeled bacteria in the presence of DHR are expressed as increases of the green fluorescence of PMNs.
Phagocytosis of BCECF-AM-labeled streptococci by C5a-activated whole-blood PMNs was also quantified by flow cytometry. A 1-ml aliquot of heparinized (14 USP units/ml) whole blood from healthy human donors was incubated in a sterile 1.5-ml Eppendorf microcentrifuge tube with 50 μl of 10 μM recombinant human C5a (rhC5a; Sigma, St. Louis, Mo.) or without rhC5a for 45 min at 37°C on a Labquake rotator (8 rpm). After the incubation period, 100 μl of 107 BCECF-AM-labeled streptococci in PBS (pH 7.4) was added and the tubes were rotated at 8 rpm and at 37°C. Samples of 100 μl were removed at 0, 5, 15, 30, and 60 min, immediately mixed with 2 ml of ACK lysing buffer (Biofluids, Rockville, Md.), and incubated for 5 min at room temperature before being placed on ice. PMNs were isolated from the lysis solution by centrifugation (10 min, 4°C, 1,300 rpm; Beckman GS-6R centrifuge) and washed three times in ice-cold PBS (pH 7.4). The PMNs were resuspended in a final volume of 1 ml of PBS and analyzed by flow cytometry as described above.
Human blood bacteriocidal phagocytosis assays were performed as previously described (14). Briefly, log-phase cultures of streptococci were diluted in THY to 104 CFU/ml. One-tenth ml of diluted cultures and 0.9 ml of fresh human blood were mixed and rotated at 37°C for 3 h. Initial viable counts and counts after a 3-h rotation were determined by plating diluted samples onto blood agar (14).
Sixteen-hour cultures of challenge streptococci (1 × 108 to 9 × 108 CFU), grown in THB containing 20% normal rabbit serum and resuspended in 10 μl of PBS, were administered intranasally to 25-g female CD1 mice (Charles River Breeding Laboratories, Inc., Wilmington, Mass.) (2). Viable counts were determined by plating dilutions of cultures on blood agar plates. Throat swabs were taken daily from anesthetized mice for 7 to 10 days after inoculation and were streaked onto blood agar plates containing 200 μg of streptomycin/ml. After overnight incubation at 37°C, the number of β-hemolytic colonies on plates were counted. All challenge strains were marked by streptomycin resistance to distinguish them from β-hemolytic bacteria that might be present in the normal flora. The presence of one β-hemolytic colony was taken as a positive culture.
In order to test whether M49, Mrp, and Enn cooperate to produce a fully phagocyte-resistant phenotype, a defined deletion mutant, Δmrp-emm-enn, was constructed by gene replacement by using the thermosensitive vector pG+host5 (see Materials and Methods for details). This was accomplished by replacing the sequence from within the 3′ end of mrp to the 5′ end of enn with the cat gene. The thermosensitive vector pG+host5 with the deletion cat construct was used to replace the wild-type sequence in the chromosome of strain CS101 by homologous recombination. The phenotype of this strain was fully characterized to insure that mga and scpA49 are normally expressed and that products from deleted genes are not produced. The M− strain CS101::pSF152 with a plasmid insertion in emm49 and SCPA− strain MJ3-15 with a site-directed nonpolar mutation in scpA49 were previously described (12, 21). To confirm that mutant streptococci no longer produce M protein, protein extracts were analyzed by Western blotting with anti-M49 serum. Protein extracts from the emm insertion mutants, strain CS101::pSF152, (Fig. (Fig.1,1, lane 2) and the Δmrp-emm-enn mutant MJY1-3 (Fig. (Fig.1,1, lane 1) did not react with specific M protein antibodies. Extracts from the parent and the SCPA− mutant, strain MJ3-15, showed the expected M49 protein band (Fig. (Fig.1,1, lanes 4 and 3, respectively).
The Δmrp-emm-enn mutant was confirmed to express mga and scpA transcripts by Northern hybridization. Total RNAs (10 μg) from wild-type CS101 and mutant MJY1-3 were blotted and hybridized to probes specific for emm, enn, mga, and scpA genes (Fig. (Fig.2).2). Blots showed that RNA from strain CS101 contained mRNA transcripts of the expected sizes, 1.2 and 1.0 kb, which hybridized to the emm and enn probes, respectively (Fig. (Fig.2,2, lanes 1). RNA from the mutant MJY1-3 culture lacked both emm and enn transcriptions (Fig. (Fig.2,2, lanes 2). Wild-type (Fig. (Fig.2,2, lanes 1) and mutant (Fig. (Fig.2,2, lanes 2) RNAs hybridized to mga and scpA probes, respectively. The mga transcript was 1.7 kb, the expected size (17, 23). The 1.5- and 2.9-kb minor bands that hybridized to mga and scpA probes are rRNA. It has been observed that specific mRNAs in purified streptococcal RNA can be trapped and coelectrophoresed with rRNA (3). The fact that the enn transcript was present at a lower concentration relative to that of emm mRNA was previously reported (7). In this strain of streptococcus scpA is known to be cotranscribed with two other 3′ open reading frames of unknown function (22), accounting for the 5.8-kb polycistronic mRNA that hybridized to the scpA probe. These results demonstrated that mutant MJY1-3 does not produce emm and enn transcripts and that mga and scpA are transcribed at levels similar to those for the parent culture.
If the M, Mrp, and Enn proteins collaborate to protect streptococci from phagocytic uptake, then strain MJY1-3 should be more sensitive to phagocytosis than a strain with a single mutation in either the mrp, emm, or enn gene. Both mutant and wild-type strains were tested for their capacity to resist phagocytosis by using whole-blood phagocytosis assays (Table (Table1)1) (14). The wild-type culture CS101 increased 67-fold during the 3-h rotation in fresh human blood. Mutant strains were all less able to resist phagocytosis, relative to the parent culture. The emm49 insertion mutant, strain CS101::pSF152, increased 24-fold. Strain MJY1-3 with the Δmrp-emm-enn deletion increased 9.2-fold, and to our surprise strain MJ3-15 only increased 21-fold during the 3-h rotation in human blood. These experiments confirm that resistance to phagocytosis is not solely dependent on the M protein and that the M49, Mrp, and Enn49 proteins contribute as a group to the anti-phagocytic phenotype of strain CS101. Strain CS101 and the above mutants have similar growth rates in THY and human plasma (data not shown).
To evaluate the relative importance of M and M-like proteins on colonization, 20 CD-1 outbred mice for each experimental condition were inoculated intranasally with 2.7 × 108 CFU of wild-type strain CS101 or 1.7 × 108 CFU of mutant MJY1-3. Throat swabs were taken from anesthetized mice and streaked onto blood agar plates containing streptomycin. Neither the parent or deletion strain persisted in the throat by 7 days postinoculation. Although the Δmrp-emm-enn mutant MJY1-3 appeared to be cleared from the nasopharynx somewhat more rapidly than the parent culture CS101, differences were not statistically significant (Fig. (Fig.3).3).
In a second experiment the relative importance of the C5a peptidase and M49 protein for colonization was compared (Fig. (Fig.4).4). A small difference in the capacities of M+ SCPA+ streptococci (strain CS101) and M− SCPA+ streptococci (strain CS101::pSF152) to colonize the nasopharynx of mice was observed. Apparent differences, 4 and 5 days after infection, were not statistically significant as analyzed by the chi-square or Fisher exact tests. Significant differences in the persistence of wild-type and SCPA− mutant streptococci were observed on days 3 to 6 after inoculation (Fig. (Fig.4).4). By day 4, 55% (11 of 20) of mice infected with wild-type strain CS101 retained streptococci in their throats; whereas, only 25% (5 of 20) of mice inoculated with SCPA− mutant MJ3-15 produced positive throat cultures. These results suggest that expression of M49 and M-like proteins has a minor impact on the capacity of this M49 strain to initially colonize the mouse oral mucosa and confirmed our previous report which demonstrated that streptococci require SCPA for colonization (13).
The observations that M+ streptococci, i.e., those resistant to phagocytosis in vitro, are cleared from subdermal sites of infection (13) and the throat of intranasally infected mice are contrary to the long-held belief that M proteins primarily function to block phagocytic clearance from tissue. This contradiction led to experiments that tested the possibility that activated PMNs are able to phagocytize M+ streptococci.
In the first set of experiments PMNs were activated by exposure to PMA. Figure Figure55 shows representative results from four independent assays. PMNs were displayed according to their relative granularity and intensity of green fluorescence after coincubation with BCECF-AM-labeled streptococci. Immediately after addition of labeled bacteria only a few PMNs fluoresced green (time 0) (Fig. (Fig.5,5, I). The proportion of fluorescent PMNs depended on the culture of S. pyogenes used and the presence or absence of PMA. In the absence of PMA, the wild-type strain CS101 and SCPA− strain MJ3-15 were phagocytized by only 10 to 20% of the granulocytes. In contrast the Δmrp-emm-enn deletion mutant MJY1-3 was phagocytized by greater than 90% of the PMNs after 30 min (Fig. (Fig.5,5, II).
Upon addition of PMA to the heparinized blood, 5 min prior to the addition of the BCECF-AM-labeled bacteria, the proportion of PMNs to become associated with M+ streptococci, strains CS101 and MJ3-15, increased from 10 to 20% to approximately 90% (Fig. (Fig.5,5, III). These M+ strains showed nearly identical behaviors. Decreasing the concentration of PMA proportionately reduced the PMNs associated with streptococci (data not shown). Even though sensitive to phagocytosis without PMA activation, association of the Δmrp-emm-enn deletion strain MJY1-3 with PMNs was accelerated by exposure of blood to PMA (Fig. (Fig.5,5, III).
Experiments designed to assess the oxidative burst of PMNs exposed to streptococci were performed with DHR. DHR is freely permeable, localizes in the mitochondria, and after oxidation by H2O2 and O2− to rhodamine 123, emits a bright green fluorescent signal upon excitation by blue light (488 nm) (28). When streptococci were added to PMNs which had been loaded with DHR the oxidative burst paralleled the results of the phagocytosis assays (data not shown). There were no differences in the kinetics of induction of the oxidative burst by PMNs associated with M+ SCPA+ CS101 or M+ SCPA− MJ3-15 (data not shown).
PMA is expected to upregulate Mac-1 expression on PMNs (4). Inhibition of that activation pathway should reduce phagocytosis of streptococci. Addition of CD11b MAbs to heparinized blood 5 min after the addition of PMA and 5 min prior to adding streptococci reduced the association of PMNs with CS101 and MJ3-15 streptococci (Fig. (Fig.6B).6B). This reduction was, however, less striking than that previously reported for other strains of S. pyogenes (25). Nevertheless, these results demonstrate that PMA leads to Mac-1-dependent phagocytosis of S. pyogenes (26). Strain CS101 and the SCPA− strain MJ3-15 behaved similarly in this assay. The C5a peptidase would not be expected to impact on the phagocytic potential of PMA-activated PMNs.
Initial interactions of PMA and C5a with PMNs that lead to upregulation of Mac-1 and induction of an oxidative burst differ (4). Therefore, experiments were performed to determine whether PMNs that were activated by prior exposure to C5a also develop the capacity to associate with M+ streptococci. Whole blood was preincubated with 0.5 μM rhC5a. The oxidative burst was determined to be maximal after 45 min of incubation by using the assay described above. The kinetics of association of BCECF-AM-labeled M+ SCPA+ streptococci (strain CS101) with PMNs activated by rhC5a was examined. The mean fluorescence of PMNs that were exposed to C5a increased more rapidly and to a greater intensity than that of those not preincubated with C5a (Fig. (Fig.7A).7A). Data representative of five experiments, which were performed on different days, are shown in Fig. Fig.7.7. Differences in fluorescence of PMN associated with streptococci were greatest after a 15-min incubation. A total of 92% of C5a-activated PMNs were associated with M+ SCPA+ streptococci; whereas, only 52% of PMNs were associated with streptococci by this time, if blood was not pre-incubated with C5a. Data presented in Fig. Fig.7B7B is taken from the 15-min time point of one experiment. C5a- and PMA-activated PMNs were equally associated with M+ CS101 streptococci. Strain MJY1-3, the Δmrp-emm-enn deletion mutant, was phagocytized without prior activation with PMA or C5a. These differences were small but statistically significant and reproducible when experiments were repeated. This experiment is complicated by the fact that the alternative complement pathway is rapidly activated with production of C5a and subsequent activation of PMNs by introducing streptococci to fresh blood (11).
The surface of the group A streptococcus is overlaid with a mosaic of polysaccharides, α-helical M-like proteins, and enzymes. Although little is known about the sequence of events that lead to clearance of group A streptococci from tissue, it is clear that the organism has evolved a complex system to avoid the consequences of the early inflammatory response. The capacity of this species to resist phagocytosis is a hallmark of its virulence mechanisms. This characteristic has been attributed to the M protein and more recently to the hyaluronic acid capsule (6, 29). Resistance to the phagocytic response is blocked at two stages. Surface-bound C5a peptidase destroys the early chemotactic signal that attracts PMNs and mononuclear phagocytes to the site of infection (13, 30). M protein limits the deposition of C3b opsonin onto the bacterial surface, thereby blocking recognition of streptococci by phagocyte receptors (11).
More recent studies have revealed that the array of M-like proteins, their interaction with plasma proteins, and their impact on resistance to phagocytosis in vitro varies dramatically among the different serotypes (1, 9, 21). The M49 strain used in this study expresses at least three M-like proteins on its surface, the M49, Mrp49, and Enn49 proteins (7). Knockout mutations in emm49 and mrp genes have minimal impact on the ability of the organism to resist phagocytosis (21). Phagocytosis assays were performed as originally described by Lancefield (14). In this system resistance to phagocytosis is a complex, dynamic process that involves complement activation, opsonization, and multiplication of streptococci over the time period of the assay. Experiments reported here confirmed that the insertion mutation in emm49 had a small effect on the capacity of the organism to resist phagocytosis. This finding raised the question of whether all three proteins contribute to the resistance phenotype in an additive manner. A deletion mutant was constructed, which lacks all three M-like proteins on its surface, to test this possibility. The M49− Mrp49− Enn49− deletion mutant was more sensitive to phagocytosis than the wild type or M− culture, supporting the suggestion that all three proteins are required for full resistance to phagocytosis (19). This deletion mutant did not, however, decrease in number beyond that originally inoculated into the rotated blood. Streptococci with a deletion that removes the entire gene cluster mga-scpA49 were completely eliminated in rotated blood (data not shown). This suggests that some other factor controlled by the positive activator, Mga, may also contribute to the resistance phenotype. Although both the M49− insertion and the M49− Mrp49− Enn49− deletion mutants appeared to be cleared from the oral mucosa somewhat more quickly than their M+ parent culture following intranasal inoculation of mice, the differences were not statistically significant (Fig. (Fig.3).3). In contrast to those mutants the SCPA− strain was cleared significantly more rapidly from the throat than was the parent culture. Thus, SCPA activity appears to be more important than M-like proteins for this strain of streptococcus to persist on the oral mucosa.
Our conclusions are consistent with those of Husmann et al. who investigated the impact of M-like proteins of a S. pyogenes strain that is unusually virulent for mice (9). Deletion of genes that encode M-like proteins had little influence on the potential of this strain to colonize the oral mucosa of C57BL/10SnJ mice. In contrast to our findings, a mutation in scpA did not reduce the potential of this mouse pathogen to colonize the oral mucosa. However, a larger dose of this mutant, relative to wild-type streptococcus, was required to cause pneumonia following intratracheal inoculation, suggesting that SCPA contributed to the potential of this strain to induce pneumonia in mice. These differences are not easily explained. Virulence in mice appears to be highly dependent on expression of a large hyaluronic acid capsule. In fact, the vir gene cluster of the M50 culture is remarkably down regulated and M and M-like proteins are barely expressed (31).
It was surprising that the M49+ SCPA49− mutant was somewhat less resistant to phagocytosis than wild-type streptococci (Table (Table1).1). This observation may be explained by the fact that M+ streptococci can be phagocytized by C5a-activated PMNs. The SCPA− phenotype could permit more rapid accumulation of C5a and concomitant activation of PMNs. The group B streptococcal C5a peptidase was also reported to effect resistance to phagocytosis (27). However, because a chemical gradient of C5a is not maintained in rotated blood, the impact of C5a on the characteristic response of PMNs in rotated blood is likely to be small. This explanation was supported by the finding that the mean fluorescence of PMNs increased more rapidly and to a greater extent when blood was mixed with BCECF-AM-labeled SCPA− streptococci than SCPA+ isogenic streptococci under the same conditions (data not shown). We previously showed that SCPA increases the ability of streptococci to colonize the oral mucosa of mice and that immunization of mice with purified protein or inactivation of SCPA by mutation enhanced clearance of streptococci from the oral mucosa (12).
Schnitzler et al. developed a flow cytometer method to quantitate phagocytosis of streptococci in whole blood (25) and performed preliminary experiments which indicated that M+ streptococci are phagocytized by PMNs in whole blood, when they were activated by PMA (26). Flow cytometry, in itself, does not distinguish between intracellular streptococci and those bound to the phagocyte surface. These investigators, however, used interference contrast and fluorescence microscopy to demonstrate that PMNs associated with M+ streptococci were internalized. We confirmed their results and showed that PMA- and C5a-activated PMNs also developed the capacity to associate M49+ streptococci. Although the wild-type M+ culture readily associated with rhC5a-activated PMNs, evidence that the streptococci were killed by activated PMNs was not observed (data not shown). Because the M protein blocks opsonization of streptococci by C3b we hypothesize that they are internalized by PMNs via receptors other than CR3 or CR1. Therefore, they may enter a vacuole or compartments of the PMN that cannot fuse with lysosomal granules. Animal experiments reported by Lukomski et al. suggested that M3 streptococci kill PMNs recruited to the peritoneum following infection of mice (15). Experiments are planned to investigate the possibility that M+ streptococci not only survive phagocytosis by PMNs but kill these phagocytes once ingested.
S. pyogenes colonization of the oral mucosa with subsequent disease is a complex process that is clearly dependent on a variety of extracellular and surface-bound macromolecules. Moreover, it has become evident that this species has evolved different but overlapping mechanisms to adhere to epithelial tissue, resist the inflammatory phagocytic response, and invade deeper tissue. It is no longer rational to point toward the M protein as the sole requirement for virulence, nor is it reasonable to assume that different serotypes or even strains within a serotype successfully infect mice or humans by using the same set of virulence factors.
This study was funded by a grant from the Public Health Service (AI10016).
We thank Tim Leonard for graphic presentations and Melodie Bahan for transcribing the manuscript.