Contribution of Protein G–Related a2-Macroglobulin–Binding ...

Contribution of Protein G-Related 2-Macroglobulin-

Binding Protein to Bacterial Virulence in a Mouse Skin

Model of Group A Streptococcal Infection

Antonia W. Toppel, Magnus Rasmussen, Manfred Rohde, Eva Medina, Gursharan

S. Chhatwal

Dow

nloaded from https://academ

ic.oup.com/jid/article/187/11/1694/884966 by guest on 19 July 2022

http://akrurl.com/.2zpmo

1694 • JID 2003:187 (1 June) • Toppel et al.

M A J O R A R T I C L E

Contribution of Protein G–Relateda2-Macroglobulin–Binding Protein to BacterialVirulence in a Mouse Skin Model of Group AStreptococcal Infection

Antonia W. Toppel,1 Magnus Rasmussen,2 Manfred Rohde,1 Eva Medina,1 and Gursharan S. Chhatwal1

1Department of Microbial Pathogenesis and Vaccine Research, Division of Microbiology, GBF-German Research Center for Biotechnology,Braunschweig, Germany; 2Department of Cell and Molecular Biology, Section for Molecular Pathogenesis, Lund University, Lund, Sweden

Protein G–related a2-macroglobulin–binding (GRAB) protein is a cell wall–attached determinant of group A

streptococcus (GAS) that interacts with the human protease inhibitor a2-macroglobulin (a2-M). Of 86 clinical

isolates tested, 23% could bind a2-M. However, all strains tested contained the grab gene. High levels of anti-

GRAB antibodies were found in the serum of convalescent GAS-infected patients, a finding that indicates that

this protein is expressed during the infection process. Among the a2-M–binding strains, 80% were skin isolates,

and 20% were throat isolates, findings that suggest that the skin environment is a preferential site for expression

of a2-M–binding activity. To test this possibility, we determined the role of GRAB in a mouse model of GAS

skin infection. The wild-type strain KTL3, which interacts with a2-M, showed high virulence. The isogenic

mutant of KTL3, MR4, devoid of surface-bound GRAB, was attenuated in virulence, compared with the wild-

type strain. Thus, mice infected with MR4 survived longer, developed smaller skin lesions, and exhibited lower

levels of bacterial dissemination than did those infected with KTL3. These results emphasize the role of GRAB

as a virulence factor of GAS.

Group A streptococci (GAS) are important human

pathogens able to cause a wide spectrum of clinical

manifestations, ranging from mild, selflimiting infec-

tions, such as pharyngitis and pyoderma, to more severe

invasive diseases, such as necrotizing fasciitis and strep-

tococcal toxic shock–like syndrome [1, 2]. Since the

1980s, a resurgence of severe, invasive GAS infections

has been observed, with a particularly high number of

necrotizing fasciitis cases [3–5]. Although the reasons

Received 17 September 2002; accepted 27 December 2002; electronically published15 May 2003.

All studies were approved by the local Animal Committee Board.

Reprints or correspondence: Dr. Gursharan Singh Chhatwal, Department of MicrobialPathogenesis and Vaccine Research, Division of Microbiology, GBF-German ResearchCenter for Biotechnology, Mascheroder Weg 1, D-38124 Braunschweig, Germany ([email protected]).

The Journal of Infectious Diseases 2003; 187:1694–703� 2003 by the Infectious Diseases Society of America. All rights reserved.0022-1899/2003/18711-0003$15.00

for the resurgence of severe GAS infections are not yet

clear, the emergence of GAS strains with increased vir-

ulence could be one of the factors [6, 7]. Severe soft-

tissue infections are characterized by intense tissue ne-

crosis that rapidly spreads from the original site of

infection [3]. For this purpose, GAS display a complex

array of virulence factors directed at facilitation of bac-

terial spread, by the degradation of host extracellular

matrix. Thus, in addition to being able to produce pro-

teases [8–14], DNase [15], and hyaluronidase [16], GAS

is able to interact directly with human components,

such as plasminogen, and is able to acquire plasminlike

enzymatic activity [17–21]. During invasive GAS in-

fection, these mechanisms might contribute to the deg-

radation of extracellular matrix and to the activation

of matrix metalloproteases [22, 23]. Concomitant with

the production and activation of proteases, GAS must

have developed mechanisms to acquire selfprotection

against proteolytic degradation. The recruitment of the

Dow



Contribution of GRAB to GAS Virulence • JID 2003:187 (1 June) • 1695

human protease inhibitor a2-macroglobulin (a2-M) to the sur-

face of GAS has been proposed as one such potential mech-

anism [24–26]. Human a2-M is a 725-kDa homotetrameric

protein capable of inhibiting a wide range of proteases [27–

30]. Interaction between GAS and a2-M seems to be mediated

mainly by the streptococcal surface protein G–related a2-mac-

roglobulin–binding (GRAB) protein [26], which was identified

by sequence homology to the a2-M–binding domain of protein

G from groups C and G streptococci [31–33]. The molecular

mass of the core protein is ∼22.8 kDa, and the core protein

contains a variable number of 28-aa repeated regions. The a2-

M–binding region is localized in the NH2-terminal part of

GRAB and is next to a repeat region. The grab gene seems to

be present in most Streptococcus pyogenes strains and is highly

conserved [26], a finding that suggests that the protein GRAB

plays a critical role in the physiology of S. pyogenes. A GAS

mutant strain devoid of surface-bound GRAB has been shown

to be less virulent than the wild-type strain, after intraperitoneal

infection of mice, a finding that indicates the potential of the

protein GRAB as a virulence determinant of GAS [26]. How-

ever, we believed that, because of the ability of GRAB to recruit

a protease inhibitor, the expression of GRAB might be advan-

tageous to GAS, in an environment with a high concentration

of proteases, such as in soft-tissue infection. Therefore, in this

study, we have evaluated the contribution of GRAB to GAS

virulence in a mouse skin infection model.

MATERIAL AND METHODS

Bacterial strains and growth conditions. The blood-isolated

S. pyogenes wild-type strain KTL3 (serotype M1) and the de-

rived GRAB-deficient mutant strain MR4 have been described

elsewhere [26]. MR4 was generated by the pFW13 suicide vec-

tor [34] and is devoid of surface-associated GRAB; instead, it

secretes a truncated form that lacks the cell wall–anchoring

region [26]. KTL3 was grown in Todd-Hewitt broth (GIBCO)

supplemented with 1% yeast extract (Difco), under static con-

ditions at 37�C, or onto blood agar plates (GIBCO). MR4 was

grown under similar conditions but was supplemented with

150 mg/mL kanamycin. Both strains exhibit similar growth

characteristics in broth. The clinical GAS isolates used for125I–a2-M–binding studies were obtained from different geo-

graphic areas, including Australia, India, and Germany. All

strains were isolated after 1993, from patients with skin and

throat infections. A high percentage of the strains were M-

nontypeable with a highly variable Vir type [35]. The remaining

strains exhibited many different M types, a finding that indi-

cates high heterogeneity among the different isolates. Many of

the clinical isolates used in this study have been characterized

by Goodfellow et al. [36].

a2-M–binding assay. The binding of a2-M to S. pyogenes

was assessed by use of 125I-labeled protein, according to pro-

cedures described elsewhere, and was standardized for groups

A, C, and G streptococci [24]. Radiolabeling of a2-M (Sigma)

was performed by use of carrier-free 125I (Amersham Pharma-

cia Biotech), by the chloramine T method, with a specific ac-

tivity of ∼1 mCi/mg of protein [24, 37]. In brief, bacteria were

cultured at 37�C in Todd-Hewitt broth supplemented with 1%

yeast extract and were harvested at different points of the

growth phase, were washed in PBS containing 0.05% Tween

20, and were incubated with 125I–a2-M for 45 min. Binding

assays were conducted in triplicate with ∼ cfu/25081.25 � 10

mL and 10 ng of 125I-labeled a2-M. After a washing step and

centrifugation, the activity retained in the pellet was measured

in a g-spectrometer and was expressed as percentage of the

added activity, as determined by precipitation with trichlo-

roacetic acid. KTL3 and MR4 strains were included as positive

and negative controls, respectively.

ELISA. After infection with GAS, antibody titers in serum

of convalescent patients were determined by ELISA. In brief,

96-well Nunc-ImmunoMaxiSorp assay plates (Nunc) were

coated with 50 mL/well His-tag–purified GRAB protein (strain

A 82, with grab sequence identical to that of strain SF370), at

a concentration of 5 mg/mL, in coating buffer (bicarbonate; pH

8.2). After incubation overnight at 4�C, plates were blocked

with 10% fetal calf serum (FCS) in PBS for 1 h at 37�C. Serum

diluted 1:50 in 10% FCS-PBS was added (100 mL/well), and

plates were incubated for 2 h at 37�C. After 4 washes with PBS

containing 0.05% Tween 20, peroxidase-conjugated mouse

anti–human IgG (PharMingen) was added, and plates were

further incubated, for 2 h at 37�C. After another 4 washes,

reactions were developed by ABTS, in 0.1 M citrate-phosphate

buffer (pH 4.35) containing 0.01% H2O2. Absorbance was de-

termined at 405 nm in a microtiter reader apparatus.

Polymerase chain reaction (PCR). PCR amplification of

the grab gene was performed by use of the following oligo-

nucleotides: primer 1, 5′-ATGGGAAAAGAAATAAAAGTGAA-

ATGC-3′ (position 61–87) of the gene, and primer 2, 5′-CTAA-

TTTTCTTTGCACTTTGAACTTAC-3′ (position 688–714) of

the gene of S. pyogenes reference strain SF370 (ATCC 700294;

GenBank accession no., GI:4589078). Chromosomal DNA from

the different clinical isolates was used as template, and the

manufacturer’s (QIAGEN) instructions were followed.

Bactericidal assay. Resistance to phagocytosis was mea-

sured in whole human blood by a modification of the Lancefield

bactericidal assay for GAS [38]. Mid-log phase bacteria were

washed and were serially diluted in PBS. In sterile glass tubes,

bacterial suspension (100 mL) containing 102 cfu was added to

1 mL of fresh, heparinized human blood, and the resultant

solution was incubated on an orbital shaker for 1.5 h at 37�C.

The survival index was calculated as the colony-forming units

Dow




Figure 1. Capacity of Streptococcus pyogenes KTL3 obtained from dif-ferent growth phases to bind 125I–a2-macroglobulin (a2-M). Bacteria weretaken from culture at early, mid, and late logarithmic growth phases, aswell as at the early and late stationary phases. OD, optical density.

Figure 2. Detection of anti–protein G–related a2-macroglobulin–binding (GRAB) IgG antibodies in serum samples from convalescent group A streptococci(GAS)–infected patients ( ) and from uninfected control subjects ( ). Serum was obtained from patients with either pharyngitis (P 1–P 3) orn p 33 n p 8skin infections (S 1–S 30) caused by Streptococcus pyogenes. Anti-GRAB IgG was determined by conventional ELISA. Serum obtained from healthy personswas used as negative control (Ctr 1–Ctr 8). OD, optical density.

recovered after the 1.5-h incubation, divided by the initial in-

oculum added before incubation.

Resistances of both KTL3 and MR4 to phagocytosis were also

examined, by use of mouse neutrophils (polymorphonuclear leu-

kocytes [PMNLs]). For this purpose, BALB/c mice were injected

with carrageenan (1 mg/mouse) (Sigma) 48 h before PMNL

isolation. Carrageenan treatment increases the number of PMNLs

and reduces the number of macrophages, in the peritoneal cavity

[39]. Peritoneal lavage was then performed by use of 5 mL of

Dulbecco’s modified Eagle medium (DMEM)–HEPES (contain-

ing 10% heat-inactivated FCS)/mouse, and isolated cells were

adjusted to cells/mL. Mid-log phase bacteria were har-63 � 10

vested, were washed, and were diluted in tissue culture medium,

to and cfu/mL. Resting or phorbol-12-myristate-6 73 � 10 3 � 10

13-acetate (ICN Biomedicals)–activated PMNLs were combined

with equal volumes of bacterial suspension (250 mL:250 mL) and

were incubated for 1 h at 37�C on a rotary shaker. Bacteria

incubated in medium without PMNLs were used as controls.

PMNLs were then lysed with distilled H2O for 5 min, and serial

dilutions were plated on blood agar. The number of viable bac-

teria was determined and was compared with that of the controls.

Mouse strains. Inbred, pathogen-free 8-week-old BALB/c

and C3H/HeN mice were purchased from Harlan-Winkelmann.

Infection model. Before infection, fur was removed from

a –cm area on the backs of mice by use of an electric2 � 2

shaver. Mice were then injected with washed mid-log phase

harvested bacteria, at a volume of 100 mL of PBS containing

the appropriate inoculum dose. Bacteria were injected by use

of a 27-gauge needle, which raised the superficial bleb below

the skin. The number of injected microorganisms was deter-

mined by spectrophotometry (Novaspec II; Amersham Phar-

macia Biotech) and was verified by performance of colony

counts on blood agar–plated serial dilutions. The size, after

infection, of the skin lesions generated with either KTL3 or

MR4 was determined by use of a sliding caliper and a millimeter

scale, at intervals of 24 h.

Determination of bacterial loads in systemic organs. Mice

were infected subcutaneously with cfu of either KTL382.5 � 10

or MR4. Groups of 5 mice were killed at 24, 48, and 72 h after

inoculation, and livers and spleens were removed and homog-

enized in 5 mL of PBS. Organ homogenates were serially diluted

in PBS and were plated on blood agar plates. To exclude po-

tential revertants, samples were double-plated on blood agar

supplemented with kanamycin. Further confirmation was done

by replica plating of the mutant strain obtained from tissue of

infected mice.

Isolation and characterization of phagocytic cells from GAS-

infected mice. BALB/c mice were infected subcutaneously

with cfu of either KTL3 or MR4, and inflammatory85 � 10

cells attracted to the infected site were isolated by subdermal

Dow




Figure 3. Ability of Streptococcus pyogenes KTL3 and protein G–relateda2-macroglobulin–binding protein–deficient mutant MR4, obtained from cul-tures at different growth phases, to survive in whole human blood. OD,optical density.

Figure 4. A, Kaplan-Meier survival curves of C3H/HeN mice after subcutaneous challenge with cfu of Streptococcus pyogenes strain KTL3,82.5 � 10protein G–related a2-macroglobulin–binding protein–deficient MR4 mutant strain, or PBS alone (control). Animals were monitored, and deaths were recorded,daily, over 7 days. B and C, Bacterial loads in livers (B) and spleens (C) of C3H/HeN mice at 24, 48, and 72 h after inoculation with cfu of82.5 � 10either KTL3 or MR4. Each time represents a mean of 5 mice/group; bars indicate SEs.

lavage with 2 mL of HEPES-buffered DMEM containing 10%

FCS and 1% each of glutamine, penicillin, and streptomycin

and were analyzed by cell cytometry, to determine the per-

centage of macrophages to neutrophils. Cell cytometry was per-

formed by use of phycoerythrin-conjugated anti-mouse RB6

and fluorescein isothiocyanate–conjugated anti-mouse F480

antibodies (PharMingen).

Tissue collection and histology. Mice were injected sub-

cutaneously with cfu of either KTL3 or MR4, and skin85 � 10

lesions were removed after 48 h of inoculation, by wide mar-

ginal excision around the injection site. Tissue sections were

fixed in 10% phosphate-buffered formaldehyde for 24 h, were

washed extensively, and were stored in 70% ethanol. Tissue

dehydration and paraffin embedding were performed according

to standard protocols [40]. Samples were then sectioned to 7-

mm slices by a rotary microtome (Leica RM 2135), were stained

with buffered azure–eosin, and were examined by an Axioskop

microscope (Zeiss).

For immunofluorescence staining, fixed and immobilized

sections were rehydrated according to standard protocols [40]

and were placed in distilled H2O. After blocking for 1 h with

PBS containing 10% heat-inactivated FCS, samples were over-

laid with rabbit polyclonal anti–S. pyogenes serum [41] (1:100),

in PBS with 1% FCS, for 1 h at room temperature. Unbound

antibodies were removed by immersion of the samples in PBS.

In a second step, sections were incubated with a tetramethyl

rhodamine isothiocyanate–conjugated polyclonal goat anti–

rabbit IgG (Sigma; 1:200), in PBS with 1% FCS, for 1 h, were

extensively washed with PBS, and were mounted. Immunoflu-

orescence was then visualized by a fluorescence microscope

(Axioskop; Zeiss ).

Electron microscopy samples were fixed in a fixation solution

of 0.2% glutaraldehyde and 0.5% formaldehyde, in cacodylate

buffer (pH 6.9; 0.1 M cacodylate, 0.09 M sucrose, 0.01 M MgCl2,

and 0.01 M CaCl2), for 1 h on ice. After several washing steps

with cacodylate buffer containing 10 mM glycine, samples were

Dow




Figure 5. Size of skin lesions of C3H/HeN (A) or BALB/c (B) mice, after 48 h of subcutaneous infection with either Streptococcus pyogenes KTL3 (solid bars) or protein G–related a2-macroglobulin–binding (GRAB) protein–deficient mutant MR4 (open bars). Results represent means of 10 mice/group.*P ! .05. C, Photographs showing skin lesions developed on BALB/c mice at day 5 of subcutaneous infection with either KTL3 or MR4.

dehydrated according to the progressive-lowering-of-temper-

ature method, by use of a graded series of ethanol: 10% ethanol

on ice, 30% ethanol at �20�C, and 50%–100% ethanol at

�30�C. Samples were then infiltrated with Lowicryl K4M resin

(1 part ethanol:1 part K4M overnight, 1 part ethanol:2 parts

K4M for 24 h, and pure K4M resin for 48 h, with several

changes). Samples were polymerized by UV light (366 nm) for

2 days at �30�C and were further polymerized by UV light for

another 2 days, at room temperature. Samples were cut with

a diamond knife and were collected onto polyvinyl formal–

coated 300-mesh copper grids (Fluka). Grids were then incu-

bated with a 1:25 dilution of the anti-GRAB antibody (stock

solution, 1.8 mg/mL of IgG protein) for 12 h at 4�C. They were

then washed with PBS, were incubated with protein A–gold

complexes (15 nm in diameter) for 30 min at room temperature

(BritishBiocell), were washed again, with PBS containing 0.1%

Tween 20, were subsequently washed in distilled water, and

were air-dried. Counterstaining was performed, by use of 4%

aqueous uranyl acetate, for 5 min. Samples were then examined

in an Oberkochen transmission electron microscope (EM910;

Zeiss) at an acceleration voltage of 80 kV.

Statistical analysis. Statistical significance between sam-

ples was determined by Student’s t test.

RESULTS

a2-M binding among skin and throat GAS isolates. Eighty-

six strains of S. pyogenes isolated from patients with either

throat infection ( ) or skin infection ( ) were testedn p 40 n p 46

for their capacities to bind 125I-labeled a2-M. Strains binding

110% of the original added activity were considered to be dis-

playing binding activity. Twenty isolates (23%) bound a2-M,

whereas 66 isolates (77%) did not. Studies of the correlations

between the source of the isolates and their capacities to bind

a2-M showed that 80% of strains binding a2-M were skin iso-

lates, a finding that suggests a preferential expression of protein

GRAB during skin infection. However, it is possible that the

strains not binding a2-M had lost this function after subculture

under laboratory conditions. Thus, the clinical GAS isolates

were genotyped for the presence of the grab gene by PCR am-

plification with grab-specific oligonucleotide primers. All of the

strains tested contained the gene, a finding that indicates the

relevance of the protein GRAB for the biology of the bacterium.

a2-M–binding activities during the different bacterial

growth phases. The binding capacity of KTL3 was deter-

mined during various growth phases by kinetic studies with125I-labeled a2-M. MR4 was used as negative control, to rule

out nonspecific binding to other bacterial determinants. Max-

Dow



Figure 6. A, Flow cytometric analysis of inflammatory neutrophils (PMNLs) (RB6) and macrophages (F480) present in skin lesions of BALB/c mice infected with either Streptococcus pyogenes KTL3 (upperpanels) or protein G–related a2-macroglobulin–binding (GRAB) protein–deficient mutant MR4 (lower panels), at 48 h of infection (solid line). Isotype-matched antibodies served as controls in all experiments(dotted line). B, C, and D, Histopathologic analyses of dermal sections taken from BALB/c mice at 48 h after inoculation with either KTL3 (B) or MR4 (C). Sections were stained with eosin–azure blue. Bacteriaare indicated with arrows. Insets in B and C show higher magnification of inflammatory PMNLs. Skin sections were also stained with tetramethyl rhodamine isothiocyanate–conjugated GAS-specific antibodiesand were visualized by immunofluorescence microscopy (D). Lower-magnification (left) and higher-magnification (right) photographs are shown. Five mice per group were used for histopathologic studies.

Dow




Figure 7. Transmission electron photographs showing surface localization of protein G–related a2-macroglobulin–binding (GRAB) protein in Streptococcuspyogenes KTL3 (A and B) and GRAB protein–deficient mutant strain MR4 (C and D), in skin sections of infected BALB/c mice, 48 h after inoculation. *,Bacterial capsule.

imal a2-M–binding activity of KTL3 was observed during the

early exponential growth phase, and activity slowly declined

until reaching a plateau (40% binding activity) during the sta-

tionary phase (figure 1).

Detection of anti-GRAB antibodies in serum of convalescent

patients. A few weeks after diagnosis, serum samples were

collected from convalescent patients with either pharyngitis or

skin infection caused by S. pyogenes. Serum samples were as-

sayed for the presence of anti-GRAB IgG antibodies. As shown

in figure 2, significantly higher levels of anti-GRAB antibodies

were found in the serum samples of all patients tested than in

the serum samples of healthy subjects from the control group,

which included persons from the same geographic area who

were seronegative for GAS antibodies by whole-cell ELISA.

Resistance to phagocytosis of MR4. The possibility that pro-

tein GRAB can contribute to the antiphagocytic capacity exhib-

ited by GAS was then examined. For this purpose, survival of

both KTL3 and MR4 were measured in fresh human blood from

3 separate donors, by a modified Lancefield bactericidal assay.

Both strains were equally resistant to phagocytotic killing in

whole blood. No differences in antiphagocytotic activity were

found between MR4 and KTL3, both of which had been obtained

from cultures at different growth phases (figure 3). Similar results

were obtained in killing assays by use of resting or phorbol-12-

myristate-13-acetate–activated mouse PMNLs (data not shown).

Survival times of mice after subcutaneous infection with

KTL3 or the MR4 strains of S. pyogenes. Groups of 10 C3H/

HeN mice were infected subcutaneously with cfu of82.5 � 10

Dow




either KTL3 or MR4, and survival was monitored over a 7-day

period. As shown in figure 4A, at day 7 after inoculation, 90%

of mice infected with MR4 survived infection, whereas only

40% of mice infected with KTL3 were still alive.

To determine whether increased survival was associated with

a lower rate of systemic bacterial growth, bacterial burden in

liver and spleen was determined at 24, 48, and 72 h after in-

oculation. Results in figure 4B and 4C show that bacterial dis-

semination from skin to systemic organs was comparable in

both MR4 and KTL3, at 24 h after infection. From this time

on, infection was stabilized in mice infected with MR4, and

bacterial clearance could be observed to some extent in the

spleens of these mice. In contrast, progressive bacterial growth

was observed in systemic organs of mice infected with KTL3.

All bacteria recovered, at different times after infection, from

the organs of mice infected with MR4 remained resistant to

kanamycin, as determined by double-plating in both the pres-

ence and the absence of the antibiotic and also by replica

plating.

S. pyogenes KTL3 generates larger lesions than does MR4,

after subcutaneous infection of mice. In a first set of ex-

periments, C3H/HeN mice were infected subcutaneously with

either KTL3 or MR4, and the development of skin lesions was

monitored on a daily basis. A white intensely inflamed area

was observed as soon as 24 h after inoculation. As the infection

progressed, the local inflammation expanded faster in mice

infected with KTL3 than it did in those infected with MR4,

giving rise to much larger lesions in mice in the former group

(figure 5A). These differences were also evident when half of

the inoculum dose was used (figure 5A). To rule out an influ-

ence of mouse background on the differences observed between

KTL3 and MR4, BALB/c mice, which are very resistant to GAS

infection [42], were infected subcutaneously with cfu85 � 10

of either KTL3 or MR4. Similar to what was observed in C3H/

HeN mice, in BALB/c mice, infection with MR4 generated

smaller and more superficial lesions than did infection with

KTL3 (figures 5B and 5C).

Histopathologic examination of skin lesions. BALB/c mice

were infected subcutaneously with cfu of either KTL385 � 10

or MR4, and skin sections were taken at 48 h after inoculation,

for histopathologic examination. Stained sections of skin iso-

lated from both groups of mice revealed a strong inflammatory

response, mainly composed of an intense infiltration of PMNLs,

a finding that was confirmed by cell cytometric analysis (figure

6A). High densities of streptococci can be observed at the in-

oculation sites of mice infected with either KTL3 (figure 6B)

or MR4 (figure 6C). However, in tissue sections from MR4-

infected mice, the infection seems to be contained at the in-

fection site by a thick surrounding layer of inflammatory cells

(figure 6C). In contrast, in KTL3-infected mice, the infection

was widely spread in epidermis, dermis, and subcutis (figure

6B). Invasion of the deepest layers of the skin was also more

pronounced in skin infected with the KTL3 strain (figure 6B).

The presence of high numbers of S. pyogenes in skin lesions

was further demonstrated by immunofluorescence staining

with GAS-specific serum antibodies (figure 6D). Additional

presence of protein GRAB in the surface of KTL3 was confirmed

by electron microscopy (figure 7A). The surface MR4 was in-

cluded as control (figure 7B).

DISCUSSION

The results presented here indicate that GRAB plays an im-

portant role in streptococcal virulence, because, in a mouse

model of skin infection, the presence of surface-associated pro-

tein GRAB confers a survival advantage to S. pyogenes. That

the grab gene is present in all clinical isolates tested [26] (this

study) and that anti-GRAB antibodies are present in the serum

of convalescent GAS-infected patients provide further evidence

for a role of the protein GRAB during the infection process.

S. pyogenes is generally an extracellular pathogen that colo-

nizes either the mucosal epithelium of the upper respiratory

tract or the epidermis of the skin. Local infections caused by

S. pyogenes, as well as particular skin infections, are character-

ized by both an extensive infiltration of PMNLs and serum

extravasations [43]. Although recruited PMNLs are important

for bacterial clearance [44], neutrophils, by the release of ox-

ygen-free radicals and proteases, also contribute to tissue dam-

age [45]. In addition, during soft-tissue infection, GAS not only

produce proteases and other products for degradation of the

extracellular matrix [8, 12, 14, 46], but they also activate host

metalloproteases [23]. Therefore, in these ecological niches, sur-

vival and persistence of S. pyogenes is ensured by a number of

strategies directed to circumvent the host immune system (e.g.,

antiphagocytic mechanisms) and probably also to acquire pro-

tection from a harsh environment [38, 47]. Regarding this last

point, the recruitment of the protease inhibitor a2-M to the

surface of GAS has been proposed as a potential mechanism

for bacterial protection against proteolytic degradation [26, 48].

a2-M is present in high concentrations in serum [29] and can

be transported to the site of GAS infection, after serum ex-

travasation, during the inflammatory reaction. That binding of

a2-M is predominant among GAS strains isolated from skin

infections underlines the importance of a2-M binding activity

of GAS during skin infection. Binding of a2-M to GAS is mainly

mediated by the protein GRAB [26]. In the present study, we

have shown that, in a mouse model of skin infection, a GRAB-

deficient mutant strain of S. pyogenes (MR4) is attenuated in

virulence, compared with the wild-type strain (KTL3). This

finding supports a previous report showing partial attenuation

in virulence of MR4 after intraperitoneal infection of mice [26].

The mechanism by which recruitment of a2-M confers survival

Dow




advantage to S. pyogenes is not yet clear. However, it can be

hypothesized that binding of a2-M to the bacterial surface

might contribute, on the one hand, to removal of a2-M from

the environment and, thus, to maintenance of the effective

activity of the proteases for more efficient tissue spreading and,

on the other hand, to protection of the bacterial surface pro-

teins from the action of these proteases. Support for these

notions is provided by the finding that GAS is unable to bind

a2-M when a2-M is already coupled with the proteases [24,

25]. However, because MR4 is deficient only in surface-bound

GRAB but is still able to release GRAB in the infection milieu

[26], the survival advantage of KTL3 that has been observed

in the mouse skin infection might be mainly due to the pro-

tection from protease degradation conferred by the presence

of GRAB on the bacterial surface. Thus, exposure of M pro-

tein and other surface proteins involved in antiphagocytic ac-

tivity to the activities of proteases might facilitate the uptake

of MR4 by the PMNLs present at the site of infection. Further

support for this view is provided by the preferential expression

of GRAB during logarithmic growth, expression that matches

the expression of the antiphagocytic M protein and C5a pep-

tidase [8].

In conclusion, our results indicate that protein GRAB, the

only a2-M–binding protein identified to date for S. pyogenes,

is important during the pathogenesis of skin infection. These

results highlight the potential of the protein GRAB as a vaccine

candidate against GAS infections.

References

1. Cunningham MW. Pathogenesis of group A streptococcal infections.Clin Microbiol Rev 2000; 13:470–511.

2. Sharkawy A, Low DE, Saginur R, et al. Severe group A streptococcalsoft-tissue infections in Ontario: 1992–1996. Clin Infect Dis 2002; 34:454–60.

3. Bisno AL, Stevens DL. Streptococcal infections of skin and soft tissues.N Engl J Med 1996; 334:240–5.

4. Chelsom J, Halstensen A, Haga T, Hoiby EA. Necrotizing fasciitis dueto group A streptococci in western Norway: incidence and clinicalfeatures. Lancet 1994; 344:1111–5.

5. Kaul R, McGeer A, Low DE, Green K, Schwartz B, Simor AE. Pop-ulation-based surveillance for group A streptococcal necrotizing fas-ciitis: clinical features, prognostic indicators, and microbiologic analysisof seventy-seven cases. Am J Med 1997; 103:18–24.

6. Johnson DR, Wotton JT, Shet A, Kaplan EL. A comparison of groupA streptococci from invasive and uncomplicated infections: are virulentclones responsible for serious streptococcal infections? J Infect Dis2002; 185:1586–95.

7. Musser JM, Kapur V, Szeto J, Pan X, Swanson DS, Martin DS. Geneticdiversity and relationships among Streptococcus pyogenes strains ex-pressing serotype M1 protein: recent intercontinental spread of a sub-clone causing episodes of invasive disease. Infect Immun 1995; 63:994–1003.

8. Rasmussen M, Bjorck L. Proteolysis and its regulation at the surfaceof Streptococcus pyogenes. Mol Microbiol 2002; 43:537–44.

9. von Pawel-Rammingen U, Johansson BP, Bjorck L. IdeS, a novel strep-tococcal cysteine proteinase with unique specificity for immunoglob-ulin G. EMBO J 2002; 21:1607–15.

10. Svensson MD, Scaramuzzino DA, Sjobring U, Olsen A, Frank C, BessenDE. Role for a secreted cysteine proteinase in the establishment of hosttissue tropism by group A streptococci. Mol Microbiol 2000; 38:242–53.

11. Ashbaugh CD, Wessels MR. Absence of a cysteine protease effect onbacterial virulence in two murine models of human invasive group Astreptococcal infection. Infect Immun 2001; 69:6683–6.

12. Ashbaugh CD, Warren HB, Carey VJ, Wessels MR. Molecular analysisof the role of the group A streptococcal cysteine protease, hyaluronicacid capsule, and M protein in a murine model of human invasivesoft-tissue infection. J Clin Invest 1998; 102:550–60.

13. Raeder R, Harokopakis E, Hollingshead S, Boyle MDP. Absence ofSpeB production in virulent large capsular forms of group A strep-tococcal strain 64. Infect Immun 2000; 68:744–51.

14. Lukomski S, Montgomery CA, Rurangirwa J, et al. Extracellular cys-teine protease produced by Streptococcus pyogenes participates in thepathogenesis of invasive skin infection and dissemination in mice.Infect Immun 1999; 67:1779–88.

15. Wannamaker LW. The differentiation of three distinct desoxyribonu-cleases of group A streptococci. J Exp Med 1958; 107:797–812.

16. Hynes WL, Hancock L, Ferretti JJ. Analysis of a second bacteriophagehyaluronidase gene from Streptococcus pyogenes: evidence for a thirdhyaluronidase involved in extracellular enzymatic activity. Infect Im-mun 1995; 63:3015–20.

17. Tewodros W, Norgren M, Kronvall G. Streptokinase activity amonggroup A streptococci in relation to streptokinase genotype, plasmin-ogen binding, and disease manifestations. Microb Pathog 1995; 18:53–65.

18. Wang H, Lottenberg R, Boyle MDP. Analysis of plasmin(ogen) acqui-sition by clinical isolates of group A streptococci incubated in humanplasma. J Infect Dis 1994; 169:143–9.

19. Wang H, Lottenberg R, Boyle MDP. Analysis of the interaction of groupA streptococci with fibrinogen, streptokinase and plasminogen. MicrobPathog 1995; 18:153–66.

20. Wang H, Lottenberg R, Boyle MDP. A role for fibrinogen in the strep-tokinase-dependent acquisition of plasmin(ogen) by group A strep-tococci. J Infect Dis 1995; 171:85–92.

21. Christner R, Li Z, Raeder R, Podbielski A, Boyle MDP. Identificationof key gene products required for acquisition of plasmin-like enzymaticactivity by group A streptococci. J Infect Dis 1997; 175:1115–20.

22. Li Z, Ploplis VA, French EL, Boyle MDP. Interaction between groupA streptococci and the plasmin(ogen) system promotes virulence in amouse skin infection model. J Infect Dis 1999; 179:907–14.

23. Burns EH Jr, Marciel AM, Musser JM. Activation of a 66-kilodaltonhuman endothelial cell matrix metalloprotease by Streptococcus pyo-genes extracellular cysteine protease. Infect Immun 1996; 64:4744–50.

24. Chhatwal GS, Muller HP, Blobel H. Characterization of binding of hu-man a2-macroglobulin to group G streptococci. Infect Immun 1983;41:959–64.

25. Chhatwal GS, Albohn G, Blobel H. Novel complex formed between anonproteolytic cell wall protein of group A streptococci and a2-mac-roglobulin. J Bacteriol 1987; 169:3691–5.

26. Rasmussen M, Muller HP, Bjorck L. Protein GRAB of Streptococcuspyogenes regulates proteolysis at the bacterial surface by binding a2-macroglobulin. J Biol Chem 1999; 274:15336–44.

27. Sottrup-Jensen L, Stepanik TM, Kristensen T, et al. Primary structureof human a2-macroglobulin. J Biol Chem 1984; 259:8318–27.

28. Gonias SL, Roche PA, Pizzo SV. Purification and characterization ofhuman a2-macroglobulin conformational variants by non-ideal high per-formance size-exclusion chromatography. Biochem J 1986; 235:559–67.

29. Coan MH, Roberts RC. A redetermination of the concentration of a2-macroglobulin in human plasma. Biol Chem Hoppe Seyler 1989; 370:673–6.

30. Moncino MD, Roche PA, Pizzo SV. Characterization of human a2-macroglobulin monomers obtained by reduction with dithiothreitol.Biochemistry 1991; 30:1545–51.

31. Sjobring U, Trojnar J, Grubb A, Akerstrom B, Bjorck L. Ig-binding

Dow




bacterial proteins also bind proteinase inhibitors. J Immunol 1989;143:2948–54.

32. Sjobring U, Bjorck L, Kastern W. Streptococcal protein G gene structureand protein binding properties. J Biol Chem 1991; 266:399–405.

33. Muller HP, Rantamaki LK. Binding of native a2-macroglobulin to hu-man group G streptococci. Infect Immun 1995; 63:2833–9.

34. Podbielski A, Spellerberg B, Woischnik M, Pohl B, Luttiken R. Novelseries of plasmid vectors for gene inactivation and expression analysisin group A streptococci (GAS). Gene 1996; 177: 137–47.

35. Gardiner DL, Goodfellow AM, Martin DR, Sriprakash KD. Group Astreptococcal Vir types are M-protein gene (emm) sequence type spe-cific. J Clin Microbiol 1998; 36:902–7.

36. Goodfellow AM, Hibble M, Talay SR, et al. Distribution and antige-nicity of fibronectin binding proteins (SfbI and SfbII) of Streptococcuspyogenes clinical isolates from the northern territory, Australia. J ClinMicrobiol 2000; 38:389–92.

37. Hunter WH, Greenwood FC. Preparation of iodine-131 labeled humangrowth hormone of high specific activity. Nature 1962; 194:495–6.

38. Lancefield RC. Current knowledge of type-specific M antigens of groupA streptococci. J Immunol 1962; 89:307–13.

39. Ishizaka S, Kuriyama S, Tsujii T. In vivo depletion of macrophages bydesulfated iota-carrageenan in mice. J Immunol Methods 1989; 124:17–24.

40. Machling RC. An inexpensive and reliable method of negative-pressureparaffin wax embedding. J Med Lab Technol 1969; 26:271.

41. Molinari G, Talay SR, Valentin-Weigand P, Rohde M, Chhatwal GS.The fibronectin-binding protein of Streptococcus pyogenes, SfbI, is in-volved in the internalization of group A streptococci by epithelial cells.Infect Immun 1997; 65:1357–63.

42. Medina E, Goldmann O, Rohde M, Lengeling A, Chhatwal GS. Geneticcontrol of susceptibility to group A streptococcal infection in mice. JInfect Dis 2001; 184:846–52.

43. Scaramuzzino DA, McNiff JM, Bessen DE. Humanized in vivo modelfor streptococcal impetigo. Infect Immun 2000; 68:2880–7.

44. Smith JA. Neutrophils, host defence, and inflammation: a double-edgedsword. J Leukoc Biol 1994; 56:672–86.

45. Weiss SJ. Tissue destruction by neutrophils. N Engl J Med 1989; 320:365–76.

46. Saouda M, Wu W, Conran P, Boyle MDP. Streptococcal pyrogenicexotoxin B enhances tissue damage initiated by other Streptococcuspyogenes products. J Infect Dis 2001; 184:723–31.

47. Podbielski A, Schnitzler N, Beyhs P, Boyle MD. M-related protein(Mrp) contributes to group A streptococcal resistance to phagocytosisby human granulocytes. Mol Microbiol 1996; 19:429–41.

48. Song XM, Perez-Casal J, Bolton A, Potter AA. Surface-expressed Migprotein protects Streptococcus dysgalactiae against phagocytosis by bo-vine neutrophils. Infect Immun 2001; 69:6030–7.

Dow



Contribution of Protein G–Related a2-Macroglobulin–Binding ...

Documents