Sword and shield: Linked group B streptococcal
?-hemolysin?cytolysin and carotenoid pigment
function to subvert host phagocyte defense
George Y. Liu*, Kelly S. Doran*, Toby Lawrence†, Nicole Turkson‡, Manuela Puliti§, Luciana Tissi§, and Victor Nizet*‡¶
Departments of *Pediatrics and†Pharmacology and‡Center for Marine Biotechnology and Biomedicine, The Scripps Institution of Oceanography, University
of California at San Diego, La Jolla, CA 92093; and§Microbiology Section, Department of Experimental Medicine and Biochemical Sciences, University of
Perugia, 06122 Perugia, Italy
Communicated by Donald R. Helinski, University of California at San Diego, La Jolla, CA, August 23, 2004 (received for review January 12, 2004)
Group B Streptococcus (GBS) is a major cause of pneumonia,
inside host phagocytic cells. The pore-forming GBS ?-hemolysin?
cytolysin (?H?C) encoded by cylE is an important virulence factor
as demonstrated in several in vivo models. Interestingly, cylE
deletion results not only in the loss of ?H?C activity, but also in the
loss of a carotenoid pigment of unknown function. In this study,
we sought to define the mechanism(s) by which cylE may contrib-
ute to GBS phagocyte resistance and increased virulence potential.
We found that cylE-deficient GBS was more readily cleared from a
mouse’s bloodstream, human whole blood, and isolated macro-
phage and neutrophil cultures. Survival was linked to the ability of
?H?C to induce cytolysis and apoptosis of the phagocytes. At a
lower bacterial inoculum, cylE also contributed to enhanced sur-
vival within phagocytes that was attributed to the ability of
carotenoid to shield GBS from oxidative damage. In oxidant killing
assays, cylE mutants were shown to be more susceptible to
hydrogen peroxide, hypochlorite, superoxide, and singlet oxygen.
Together, these data suggest a mechanism by which the linked
cylE-encoded phenotypes, ?H?C (sword) and carotenoid (shield),
act in partnership to thwart the immune phagocytic defenses.
recognized as a pathogen in adult populations, including the
elderly, pregnant women, and diabetics. One important viru-
lence factor of GBS is a surface-associated toxin known as the
?-hemolysin?cytolysin (?H?C). ?H?C is responsible for the
characteristic zone of clearing around GBS colonies grown on
blood agar plates and is capable of forming pores in a variety of
eukaryotic cell membranes. ?H?C is thought to contribute to
GBS pathogenicity by virtue of its cytolytic properties, its ability
to promote bacterial invasion of epithelial and endothelial
barriers, and its activation of host cytokines and other inflam-
matory mediators. GBS ?H?C mutants exhibit decreased viru-
lence in animal models of sepsis and meningitis (1, 2). GBS
?H?C-mediated cytotoxicity is blocked by the surfactant phos-
pholipid dipalmitoyl phosphatidylcholine (DPPC), perhaps ex-
plaining in part the increased susceptibility of premature,
surfactant-deficient neonates to severe GBS pneumonia (3).
Transposon mutagenesis studies mapped GBS ?H?C activity
to the cyl operon (4). A single ORF, cylE, is both necessary for
GBS hemolysin production and sufficient to confer ?-hemolysis
to Escherichia coli (5). The predicted 79-kDa protein product,
CylE, does not share homology with other proteins in the
GenBank databases. Interestingly, the cylE gene is also required
for GBS production of an orange carotenoid pigment, a unique
feature that is useful in distinguishing GBS from other ?-hemo-
lytic streptococci (6). GBS mutants in which the cylE gene is
removed are invariably both nonhemolytic and nonpigmented,
and both phenotypes are restored in single-gene complementa-
tion experiments where cylE is returned on a plasmid vector
(4, 5, 7).
roup B Streptococcus (GBS) is the leading cause of invasive
bacterial infections in human newborns and is increasingly
A clinical association with invasive human infections implies
that GBS can sometimes survive innate host defense mecha-
tissues. A principal role in innate immunity is played by host
neutrophils and macrophages, which can engulf and kill bacteria
by generation of reactive oxygen species and other antimicrobial
substances within the phagolysosome. Although streptococci are
commonly thought of as ‘‘extracellular pathogens,’’ GBS can
survive for prolonged periods within the phagolysosome of
macrophages (8, 9). And, although GBS lack the neutralizing
enzyme catalase, they can be ?10 times more resistant to killing
by H2O2than catalase-positive Staphylococcus aureus (10). The
mechanisms responsible for the enhanced survival of GBS are
Previous interpretations of the role played by the GBS ?H?C
in animal virulence have focused on cytolytic injury to host cells
or stimulation of inflammatory responses. Here we ask the
question of whether the decreased virulence of GBS cylE
GBS against host phagocytic clearance. We provide evidence
that cylE contributes to GBS survival in neutrophils and mac-
rophages and show that this effect is due to not only toxic
properties of ?H?C but also protection by the linked carotenoid
pigment against oxidative burst killing mechanisms.
Materials and Methods
Bacteria and Cell Lines. GBS used were WT strains NCTC10?84
(serotype V, hemolytic titer ? 64 units) and A909 (serotype Ia,
hemolytic titer ? 4 units) and the corresponding nonhemolytic,
nonpigmented isogenic allelic exchange mutants NCTC:cylE?cat
V-wt, Ia-wt, V?cylE, and Ia?cylE, respectively. Bacteria were
grown in Todd–Hewitt broth (THB) or on Todd–Hewitt agar
murine macrophages, and human neutrophils were maintained in
RPMI medium 1640 plus 10% FBS.
Murine Model of Sepsis. Six- to 8-week-old CD-1 mice (Charles
River Laboratories) were injected in the tail vein with 108early
bacteremia was assessed by blood collection and enumeration of
GBS colony-forming units (cfu) on THA.
Human Whole Blood Killing Assay. GBS was grown to early logarith-
mic phase, washed, and resuspended in PBS. Inocula of 104cfu in
100 ?l were mixed with 300 ?l of freshly drawn human blood in
Abbreviations: ?H?C, ?-hemolysin?cytolysin; cfu, colony-forming units; DPI, diphenylene
iodonium; DPPC, dipalmitoyl phosphatidylcholine; GBS, group B Streptococcus; moi, mul-
tiplicity of infection; THA, Todd–Hewitt agar; THB, Todd–Hewitt broth; TUNEL, terminal
deoxynucleotidyltransferase-mediated dUTP nick end labeling.
¶To whom correspondence should be addressed. E-mail: firstname.lastname@example.org.
© 2004 by The National Academy of Sciences of the USA
www.pnas.org?cgi?doi?10.1073?pnas.0406143101 PNAS ?
October 5, 2004 ?
vol. 101 ?
no. 40 ?
dilutions were plated on THA for enumeration of cfu.
Neutrophil Purification. Human neutrophils were purified from
healthy human volunteers by using a Histopaque gradient
(Sigma) according to the manufacturer’s directions. Neutrophil
purity was ?95%.
Macrophage and Neutrophil Killing Assays. Three days after i.p.
injection of 3 ml of sterile thioglycolate broth, C57BL?6 mouse
(Charles River Laboratories) macrophages were isolated by peri-
toneal lavage with PBS. Specific inocula of logarithmic-phase GBS
were added and brought into proximity with human neutrophils or
murine macrophages by centrifugation at 600 ? g for 10 min. After
3–4 h at 37°C in 5% CO2, Triton X-100 (0.02% final) was added to
lyse the phagocytes. Lysates were plated on THA at 37°C for
enumeration of cfu. In ?H?C-blocking experiments, the inhibitor
DPPC was suspended in PBS by sonication and added to GBS for
5 min before incubation with macrophages.
Macrophage Cytotoxicity and Apoptosis Assays. To determine cy-
totoxicity, macrophages were washed free of GBS by using PBS,
incubated in 0.4% trypan blue for 30 min at 37°C, fixed in 4%
paraformaldehyde, and counterstained in 0.1% eosin. For DNA
fragmentation studies, logarithmic-phase GBS was added to 4 ?
cells were washed twice and resuspended in fresh RPMI medium
1640 plus 10% FBS plus 10 ?g?ml gentamicin plus 5 ?g?ml
penicillin. At 24 h, the phagocytes were spun down, lysed in a
solution containing 50 mM Tris (pH 7.2), 10 mM EDTA, 0.01%
SDS, 0.2 mg?ml RNase, and 0.5 mg?ml proteinase K (4 h at
55°C), and DNA was harvested by ethanol precipitation for
laddering analysis on a 2% agarose gel. For terminal deoxynucle-
otidyltransferase-mediated dUTP nick end labeling (TUNEL)
assay, macrophages and GBS (or GBS ?H?C extract) were
coincubated in an eight-well chamber slide as above for 24 h,
spun down, fixed in 4% paraformaldehyde, and assayed for
TUNEL according to manufacturer’s instructions (Roche Di-
agnostics). For in vivo studies, 8-wk-old C57BL or CD1 mice
were injected i.p. with 1 ? 108cfu of GBS. At 24 h, spleens were
harvested, fixed in paraformaldehyde, paraffin sectioned, and
stained for TUNEL.
Preparation of GBS Hemolysin or Carotenoid Extracts. V-wt or
V?cylE cells were grown in THB to OD600? 0.8, pelleted, and
resuspended in 100 ml of PBS plus 2% starch plus 1% glucose.
After 1–2 h at 37°C, the bacteria were pelleted, and the starch-
bound hemolysin in the filtered supernatant was precipitated
with an equal volume of ice-cold methanol at ?20°C, then
resuspended in 1–2 ml of PBS. Hemolytic titer was quantitated
as described in ref. 3. The carotenoid extract was prepared as
above except the extractant solution used was THB plus 60 mM
sodium phosphate (pH 7.0) plus 0.2% starch plus 1% glucose.
Expression of S. aureus Carotenoid in GBS. Primers CRTfwd (5?-
3?) were used to amplify the crtM and crtN genes (11) from a
pigmented S. aureus clinical isolate. The fragment was direc-
tionally subcloned into expression vector pDCerm (12) and used
to transform GBS mutant Ia?cylE; as a control, Ia?cylE was
transformed with pDCerm vector alone.
GBS Intracellular Survival Assays. Neutrophil assays were per-
formed by using a protocol published in ref. 13. For macrophage
assays, early stationary phase GBS was spun onto a cell mono-
layer and cultured at 37°C under a 5% CO2?95% air atmosphere.
After 30 min, phagocytes were washed three times with RPMI
medium 1640 and resuspended in RPMI medium 1640 plus 10%
FBS plus 5 ?g?ml penicillin and 10 ?g?ml gentamicin to kill
extracellular GBS. At specified time points thereafter, the cells
were washed two times with PBS and lysed with 0.02% Triton
X-100, and dilutions were plated on THA to count intracellular
bacterial cfu. To assess oxidative burst function, supernatant was
aspirated and replaced with 0.5 mg?ml nitroblue tetrazolium
(NBT) solution (Sigma). After 30 min at 37°C, the supernatant
was removed, and adherent cells were washed and extracted with
200 ?l of DMSO and assayed colorimetrically at 580 nm.
Oxidant Killing Assays.Earlystationary-phaseGBSwasspundown
and resuspended in PBS or fresh THB. H2O2was added at a final
concentration of 0.03% and GBS was incubated at 37°C for 2 h,
at which time 1,000 units of catalase was added to quench
remaining H2O2. Bacteria were plated on THA medium to
GBS was incubated in a 24-well culture plate at 37°C in the
presence or absence of 0.01–0.2 mg?ml methylene blue and 10
cm from a 100-watt light source. Control plates were handled
identically but wrapped in foil. Bacterial viability was assessed
after 3–5 h by plating on THA. Paraquat (30 mM) and sodium
hypochlorite (0.2%, Sigma) were used as sources of superoxide
and hypochlorite in similar assays. For rescue studies, bacteria
were mixed in THA supplemented with 1?10th to 1?80th vol of
mice were inoculated in the tail vein with GBS WT and cylE mutant strains. (A)
Kaplan–Meier survival plot. (B) Bacterial load in mouse blood at 24 h. (C)
Human whole blood killing assay (representative of three experiments). Sur-
vival index ? (cfu at end of assay)?(cfu at time 0).
Deletion of cylE renders GBS less pathogenic in vivo. Groups of 10–13
www.pnas.org?cgi?doi?10.1073?pnas.0406143101Liu et al.
freshly prepared carotenoid extract from WT or ?cylE mutant
(control) and centrifuged for 5 min at 10,000 ? g before use.
Statistical Analyses. Differences in murine mortality after GBS
injection were analyzed by ?2test. Differences in cfu or viable
macrophage cell count were evaluated by Student’s t test.
GBS cylE Mutants Are Less Virulent and More Easily Cleared in Vivo.
The highly hemolytic?pigmented V-wt and moderately hemo-
lytic?pigmented Ia-wt GBS strains are clinical isolates belonging
to common serotypes causing invasive neonatal infections. The
importance of the cylE-encoded phenotypes in virulence and
bacterial survival in vivo was tested in a murine i.v. infection
model (Fig. 1A). WT GBS caused significantly higher mortality
than was seen in the corresponding isogenic cylE mutants, with
the highly hemolytic?pigmented V-wt causing death in 10 of 13
mice by 24 h. Overall, surviving mice in both WT groups were
less active, less responsive, and more emaciated. The 24-h blood
bacterial load in surviving mice paralleled the clinical trend with
significantly higher levels in WT GBS vs. cylE mutants (Fig. 1B).
WT and cylE mutant GBS were also assessed for survival in
freshly drawn blood from human donors (Fig. 1C). Consistent
with findings in the mouse model, V-wt and Ia-wt strains
proliferated in human blood, whereas their isogenic cylE mu-
tants were cleared.
GBS cylE-Encoded Phenotypes Confer a Survival Advantage Against
Phagocyte Killing. Neutrophils and macrophages play a leading
role in control of GBS disease. We hypothesized as an extension
of the in vivo findings that cylE contributed to GBS survival
during their encounter with phagocytes. To test this hypothesis,
we evaluated growth of WT and cylE-deficient GBS in the
presence of purified human neutrophils or thioglycolate-elicited
WT and cylE mutant strains at the indicated moi. (A) Viable bacterial count after incubation with neutrophils. (B) Bacterial count after incubation with
macrophages (Upper) and macrophage viability by trypan blue stain (Lower, % vs. uninfected control). (C) Representative trypan blue stain of macrophages
exposed to GBS at 4 h. (D) Effect of DPPC (2 mg?ml) on bacterial cfu and viable macrophages. All experiments were performed at least three times with similar
GBS ?H?C triggers lysis of phagocytes and promotes GBS survival. Murine macrophages or human neutrophils were incubated in the presence of GBS
moi with or without DPPC (2 mg?ml). Cells were washed and antibiotics were added after 2 h. At 24 h, macrophages were harvested for DNA fragmentation
(A) or TUNEL (B) assays. (C) TUNEL staining of macrophages exposed for 24 h to various concentrations of GBS ?H?C extract. (D) TUNEL assay on splenic sections
from C57BL mice 24 h after i.p. challenge with GBS.
GBS ?H?C induces macrophages to undergo apoptosis. Macrophages were cultured in the presence GBS V-wt or V?cylE mutant strains at the indicated
Liu et al.
October 5, 2004 ?
vol. 101 ?
no. 40 ?
murine macrophages. WT strains and cylE mutants grew equally
well in media alone (data not shown). As shown in Fig. 2 A and
B, the number of cylE mutant cfu recovered after exposure to
human neutrophils or murine macrophages was significantly
lower than was seen with WT strains. The magnitude of these
differences correlated to the hemolytic titer of the WT strains,
because a notable survival advantage of the highly hemolytic
V-wt was seen at multiplicity of infection (moi) as low as 1:1
bacteria:phagocytes; a distinct survival advantage for the less
hemolytic Ia-wt was seen at high moi (20) but was absent at lower
moi (4:1 or 1:1).
GBS Survival Is Correlated to ?H?C-Mediated Cytotoxicity to Macro-
phages. The potent cytolytic property of GBS ?H?C against
human epithelial and endothelial cells has been reported in refs.
3, 14, and 15. We hypothesized that direct lysis of macrophages
by GBS ?H?C could contribute to the enhanced survival of WT
GBS compared with cylE mutants. Using trypan blue nuclear
staining, we observed an inverse relationship between the num-
ber of viable macrophages and the number of GBS surviving
exposure (Fig. 2B Lower). Shown in Fig. 2C are representative
images of peritoneal macrophages exposed to GBS and stained
with trypan blue and eosin. A marked decrease in viability of
macrophages challenged with the WT strain was seen, whereas
macrophages challenged with the cylE mutant were relatively
unaffected. If ?H?C was directly responsible for enhanced
survival of WT GBS, then DPPC inhibition of ?H?C activity
would be predicted to prevent macrophage cytotoxicity and
favor bacterial clearance. As shown in Fig. 2D, DPPC at 2 mg?ml
prevented macrophage cytotoxicity and led to enhanced clear-
ance of WT GBS to levels comparable with cylE mutants.
GBS ?H?C Activity Can Trigger Macrophage Apoptosis. The GBS
?H?C induces hepatocyte apoptosis in vivo (1), but conflicting
data exist on its contribution to GBS-mediated apoptosis of
macrophages (16, 17). We found that exposure of murine
macrophages to V-wt for 2 h, followed by 22-h incubation in the
presence of antibiotics to kill extracellular GBS, resulted in
macrophage apoptosis as evidenced by the presence of DNA
fragmentation and positive reactivity in a TUNEL assay (Fig. 3
A and B). Macrophages exposed to V?cylE mutant GBS under
the same conditions did not undergo apoptosis. Similar results
were found in apoptosis studies using the J774 macrophage-like
cell line (data not shown). The proapoptotic effect appeared to
be mediated by ?H?C, because greatly decreased apoptosis was
(Fig. 3B). Ability of GBS to induce apoptosis depended on the
?H?C titer of the GBS strain, because significant apoptosis was
documented only with the highly hemolytic V-wt strain but not
with Ia-wt GBS (data not shown). Macrophage apoptosis was
also produced by a GBS ?H?C extract in a dose-dependent
manner (Fig. 3C). Prolonged incubation of macrophages with
V-wt GBS in the absence of antibiotic killing resulted in cytolysis
without evidence of DNA fragmentation (data not shown). In
vivo studies confirmed the association of ?H?C expression with
increased apoptosis of host splenocytes. Extensive TUNEL
staining is seen within the follicles of spleens from animals
challenged with V-wt GBS in contrast with animals challenged
with the V?cylE mutant (Fig. 3D). Thus, cylE-encoded ?H?C
activity could contribute to macrophage death by direct cyto-
toxicity and?or apoptosis, in either case enhancing the potential
for GBS survival.
We compared Ia-wt and Ia?cylE survival within macrophages
and neutrophils by using a lower moi (4:1) at which GBS did not
produce detectable cytolytic activity or trigger apoptosis. Under
these conditions, there remained a 10-fold reduction in viable
intracellular cfu of the cylE mutant compared with WT GBS
were obtained by using peritoneal macrophages and purified
human neutrophils (data not shown). Use of DPPC (2 mg?ml)
to neutralize ?H?C cytotoxicity at the start of the assay did not
affect intracellular survival of either Ia-wt or Ia?cylE GBS (data
not shown). These results implied that the presence of the cylE
gene conferred an additional survival advantage to WT GBS
phages. GBS WT and cylE mutant strains were incubated with J774 cells at
moi ? 4. (A) Survival of GBS within the macrophages in the presence or
absence of oxidative burst inhibitor diphenylene iodonium (DPI) (10 mM). (B)
Oxidative burst as assessed by nitroblue tetrazolium (NBT) during the course
of the macrophage assay. (C) Intracellular survival of WT and cylE mutant GBS
when coincubated with J774 at a ratio of 4:4:1 in the presence or absence of
DPI. (D) Impact of S. aureus carotenoid (CrtMN) expression on survival of GBS
cylE mutant within J774 cells. All experiments were performed at least three
times with similar results.
cylE mutants are more susceptible to oxidative killing by macro-
www.pnas.org?cgi?doi?10.1073?pnas.0406143101Liu et al.
beyond that attributable to ?H?C-mediated cytotoxicity or
GBS cylE Mutants Are More Susceptible to Oxidative Killing Within
Phagocytes. Enhanced intracellular survival is a property of
several pathogenic bacteria that escape the phagolysosome or
otherwise neutralize its antimicrobial properties. Earlier studies
that this prolonged survival is not a consequence of escape from
the phagolysosome (9). Cathelicidin antimicrobial peptides are
another important component of myeloid cell killing of strep-
tococci (18), but we found cylE mutants were no more suscep-
tible to the cathelin-related antimicrobial peptide (CRAMP), a
murine cathelidicin, than the WT serotype V or Ia strains (data
not shown). We next tested the relative susceptibility of WT and
cylE mutant GBS to oxidative burst killing by performing a
macrophage killing assay in the presence or absence of 10 mM
diphenylene iodonium (DPI), an inhibitor of reactive oxygen
species. As shown in Fig. 4A, the addition of DPI increased the
survival of GBS V-wt by ?2-fold and V?cylE by 10-fold. To
exclude the possibility that cylE encodes a function that blunted
activation of the macrophage oxidative burst, we performed a
time-course nitroblue tetrazolium assay that revealed a similar
profile upon phagocytosis of either WT and cylE mutant GBS
bacteria (Fig. 4B). As a further control, a mixing experiment was
performed in which WT and cylE mutant GBS were coincubated
with macrophages. In this experiment, WT bacteria had a
marked survival advantage that could be abrogated by addition
of DPI (Fig. 4C). These data strongly suggest that cylE confers
to GBS a specific resistance to oxidative killing in addition to its
GBS Pigmentation Confers Resistance to Killing by Reactive Oxygen
Species. The cylE gene is required for GBS production of an
orange pigment (3, 7), which can be appreciated in the compar-
ison of concentrated cultures of V-wt and V?cylE (Fig. 5A).
Spectral analysis of pigment extract from GBS showed a triple
peak characteristic of a carotenoid in the V-wt strain that was
absent in a similar extract from the V?cylE mutant (Fig. 5B).
Because carotenoids from plant and animal sources are known
to possess antioxidant properties, we hypothesized that the GBS
pigment may function to shield the bacterium from the action of
reactive oxygen species used in phagolysosomal killing. To
substantiate this hypothesis, we exposed WT GBS and cylE
mutant GBS to four principal oxidants of phagolysosomal kill-
ing: hydrogen peroxide (H2O2), hypochlorite, superoxide, and
singlet oxygen (Fig. 5 C–F). We found that WT GBS was
consistently more resistant to oxidant killing than isogenic cylE
singlet oxygen and hypochlorite and 5- to 10-fold for H2O2and
superoxide. The V?cylE mutant could be rescued from singlet
oxygen killing in a dose-dependent fashion by a filtered pigment
extract from the V-wt strain (Fig. 5G). Because all these
experiments were performed in a cell-free system, cytotoxic
effects of ?H?C are excluded. The relative resistance of WT
GBS to H2O2killing was not reversed by addition of the ?H?C
inhibitor DPPC (data not shown). The biosynthetic pathway for
GBS carotenogenesis is unknown (see Discussion), but we
recently discovered that genes encoding a carotenoid pigment
from S. aureus (crtMN) could be functionally expressed in GBS.
When these genes were expressed in the GBS Ia?cylE mutant,
a pigmented but ?H?C-deficient phenotype was generated, and
the transformed bacteria showed prolonged survival in macro-
phages compared with the Ia?cylE mutant transformed with
vector alone (Fig. 4D), lending support to our hypothesis that
carotenoid independently contributes to GBS resistance to
phagocyte oxidative killing.
The cylE gene is required for ?H?C production and pigmenta-
host phagocytic clearance mechanisms. Using two different GBS
WT strains, we show the contribution of cylE to survival in mice
and human whole blood, as well as the simplified setting of
bacterial coincubation with purified neutrophils or macro-
phages. The antiphagocytic properties encoded by cylE correlate
with decreased virulence and lethality of nonhemolytic, nonpig-
mented GBS observed in our mouse challenges and reported in
earlier in vivo studies (1, 2).
The cylE-encoded ?H?C of GBS produced direct cytolytic
injury to macrophages and could induce macrophage apoptosis
over a longer interval. With highly hemolytic strains or at high
bacterial inocula, GBS killing of the phagocyte appears to
outpace the phagocyte’s microbicidal mechanisms, allowing
extract (B) of late-logarithmic-phase GBS WT and cylE mutant strains. (C–F) Susceptibility of GBS WT and cylE mutant strains to oxidants. GBS was exposed to
singlet oxygen (methylene blue 0.05 mg?ml for 3 h), H2O2(0.03% for 2 h), superoxide (40 mM paraquat for 24 h), or hypochlorite (0.2% for 3 h). Survival index
of at least three experiments performed under each condition.
Liu et al.
October 5, 2004 ?
vol. 101 ?
no. 40 ?
bacterial proliferation. Addition of a ?H?C inhibitor blocked Download full-text
macrophage cytolysis and apoptosis and, thereby, restored
phagocytic killing. GBS is capable of triggering apoptosis by
?H?C-dependent (16) and ?H?C-independent mechanisms
(17). Our data suggest that a specific proapoptotic effect of the
toxin is appreciable in highly hemolytic GBS strains even at low
inoculum (e.g., V-wt at moi ? 1), whereas it is muted in weakly
hemolytic strains even at high inoculum (e.g., Ia-wt at moi ? 20).
At GBS inocula below the threshold for triggering macrophage
cytolysis or apoptosis, we found that cylE continued to promote
intracellular survival of the bacterium. We attribute this effect to
the other cylE-dependent GBS phenotype, the production of an
orange carotenoid pigment. Carotenoids are notable in biology for
their properties as free-radical scavengers (19), and we show that
the GBS carotenoid affords protection against H2O2, hypochlorite,
superoxide, and singlet oxygen, each an important component of
the oxidative burst killing mechanisms of host phagocytic cells. In
the presence of an oxidative burst inhibitor, the survival advantage
of WT GBS over the cylE mutant was diminished. The antioxidant
and explain the unexpected persistence of GBS within macrophage
phagolysosomes (8, 9) and high-level resistance of the organism to
Diminished oxidative burst function is present in the phago-
cytes of premature, stressed, or septic neonates (20), pregnant
women (21), and diabetics (22), the highest risk groups for
invasive GBS infections. We speculate that antioxidant effects of
the GBS carotenoid are sufficient to prevent bacterial killing by
phagocytes with a suboptimal oxidative burst. Neutrophils from
children with chronic granulomatous disease (CGD) are defi-
cient in NADPH oxidase and show diminished bactericidal
capacity for GBS (23). Interestingly, other bacterial and fungal
pathogens classically associated with infections in CGD patients
produce carotenoid (S. aureus) or melanin (Aspergillus fumigatus
and Burkholderia cepacia) pigments with antioxidant properties
(24–26), suggesting a common pathogenic theme and revealing
a potential selective advantage for pigment production against
phagocytic clearance mechanisms.
Consistent with earlier findings (6), the GBS pigment exhib-
ited a triple peak at 455, 485, and 520 nm highly characteristic of
?-carotene and related pigments. We found the GBS carotenoid
chromophore degraded quickly to a single 425-nm peak when
heated to 40°C, and that its migration pattern on size exclusion
columns or HPLC suggested conjugation to a macromolecule
(data not shown). The carotenoid was insoluble in methanol,
hexane, and several other nonpolar solvents. These properties
appear to distinguish the GBS carotenoid chromophore from
other known bacterial carotenoids.
The GBS pigment is a product of de novo synthesis because a
shown). Moreover, carotenoid is extractable from GBS resus-
pended in saline buffers and grown aerobically or anaerobically
with only simple sugars as an energy source (6). The two
published GBS genome sequences (27, 28) lack homologues of
the phytoene synthases, phytoene or carotene dehydrogenases,
carotenogenesis (29), indicating that a previously uncharacter-
ized pathway and?or enzymes are involved in producing the GBS
pigment. Allelic exchange mutagenesis and single-gene comple-
mentation demonstrate the cylE gene product is required for
GBS to exhibit pigmentation (4, 5), but whether this effect is
direct or indirect remains to be elucidated.
In summary, we have shown that a unique combination of two
virulence phenotypes linked through a single genetic determinant
(cylE) allows the pathogen GBS to subvert host phagocytic clear-
ance mechanisms. At high bacterial inoculum such as seen in the
cfu per g of tissue), cytotoxic and proapoptotic effects of the ?HC
toxin may overwhelm the host phagocytes before killing can occur.
At all stages of infection, the antioxidant properties of the carot-
enoid pigment could potentially create an elevated threshold for
bactericidal activity that may not be achieved by phagocytic cells of
the immature or compromised host. Further understanding of the
precise molecular mechanisms underlying these two cylE-encoded
phenotypes can provide therapeutic targets to increase GBS sus-
ceptibility to normal innate immune defenses.
We thank William Fenical and Matthew Woolery for assistance with
carotenoid characterization and Nissi Varki for histopathology studies.
This work was supported by a Howard Hughes Medical Institute
Fellowship (to G.Y.L.), a Burroughs–Wellcome Career Award (to
K.S.D.), an Edward J. Mallinckrodt, Jr., Scholar Award (to V.N.),
National Institutes of Health Grant AI048694 (to V.N.) and University
of California at San Diego Institutional National Institutes of Health
training award A107036 (to G.Y.L.).
1. Ring, A., Braun, J. S., Pohl, J., Nizet, V., Stremmel, W. & Shenep, J. L. (2002)
J. Infect. Dis. 185, 1745–1753.
2. Doran, K. S., Liu, G. Y. & Nizet, V. (2003) J. Clin. Invest. 112, 736–744.
3. Nizet, V., Gibson, R. L., Chi, E. Y., Framson, P. E., Hulse, M. & Rubens, C. E.
(1996) Infect. Immun. 64, 3818–3826.
4. Spellerberg, B., Pohl, B., Haase, G., Martin, S., Weber-Heynemann, J. &
Lutticken, R. (1999) J. Bacteriol. 181, 3212–3219.
5. Pritzlaff, C. A., Chang, J. C., Kuo, S. P., Tamura, G. S., Rubens, C. E. & Nizet,
V. (2001) Mol. Microbiol. 39, 236–247.
6. Tapsall, J. W. (1986) J. Med. Microbiol. 21, 75–81.
7. Spellerberg, B., Martin, S., Brandt, C. & Lutticken, R. (2000) FEMS Microbiol.
Lett. 188, 125–128.
8. Cornacchione, P., Scaringi, L., Fettucciari, K., Rosati, E., Sabatini, R., Orefici,
G., von Hunolstein, C., Modesti, A., Modica, A., Minelli, F. & Marconi, P.
(1998) Immunology 93, 86–95.
9. Teixeira, C. F., Azevedo, N. L., Carvalho, T. M., Fuentes, J. & Nagao, P. E.
(2001) Microsc. Res. Tech. 54, 254–259.
10. Wilson, C. B. & Weaver, W. M. (1985) J. Infect. Dis. 152, 323–329.
11. Wieland, B., Feil, C., Gloria-Maercker, E., Thumm, G., Lechner, M., Bravo,
J. M., Poralla, K. & Gotz, F. (1994) J. Bacteriol. 176, 7719–7726.
12. Jeng, A., Sakota, V., Li, Z., Datta, V., Beall, B. & Nizet, V. (2003) J. Bacteriol.
13. Staali, L., Morgelin, M., Bjorck, L. & Tapper, H. (2003) Cell. Microbiol. 5,
14. Gibson, R. L., Nizet, V. & Rubens, C. E. (1999) Pediatr. Res. 45, 626–634.
15. Nizet, V., Kim, K. S., Stins, M., Jonas, M., Chi, E. Y., Nguyen, D. & Rubens,
C. E. (1997) Infect. Immun. 65, 5074–5081.
16. Fettucciari, K., Rosati, E., Scaringi, L., Cornacchione, P., Migliorati, G.,
Sabatini, R., Fetriconi, I., Rossi, R. & Marconi, P. (2000) J. Immunol. 165,
17. Ulett, G. C., Bohnsack, J. F., Armstrong, J. & Adderson, E. E. (2003) J. Infect.
Dis. 188, 1049–1053.
18. Nizet, V., Ohtake, T., Lauth, X., Trowbridge, J., Rudisill, J., Dorschner, R. A.,
Pestonjamasp, V., Piraino, J., Huttner, K. & Gallo, R. L. (2001) Nature 414,
19. Krinsky, N. I. & Yeum, K. J. (2003) Biochem. Biophys. Res. Commun. 305,
20. Drossou, V., Kanakoudi, F., Tzimouli, V., Sarafidis, K., Taparkou, A., Bou-
giouklis, D., Petropoulou, T. & Kremenopoulos, G. (1997) Biol. Neonate 72,
21. Crouch, S. P., Crocker, I. P. & Fletcher, J. (1995) J. Immunol. 155, 5436–5443.
22. Mazade, M. A. & Edwards, M. S. (2001) Mol. Genet. Metab. 73, 259–267.
23. Stroobant, J., Harris, M. C., Cody, C. S., Polin, R. A. & Douglas, S. D. (1983)
Infect. Immun. 39, 966–969.
24. Dahl, T. A., Midden, W. R. & Hartman, P. E. (1989) J. Bacteriol. 171, 2188–2194.
25. Hamilton, A. J. & Holdom, M. D. (1999) Med. Mycol. 37, 375–389.
26. Zughaier, S. M., Ryley, H. C. & Jackson, S. K. (1999) Infect. Immun. 67,
27. Glaser, P., Rusniok, C., Buchrieser, C., Chevalier, F., Frangeul, L., Msadek, T.,
Zouine, M., Couve, E., Lalioui, L., Poyart, C., et al. (2002) Mol. Microbiol. 45,
M. R., Paulsen, I. T., Nelson, K. E., Margarit, I., Read, T. D., et al. (2002) Proc.
Natl. Acad. Sci. USA 99, 12391–12396.
29. Sieiro, C., Poza, M., de Miguel, T. & Villa, T. G. (2003) Int. Microbiol. 6, 11–16.
www.pnas.org?cgi?doi?10.1073?pnas.0406143101Liu et al.