Download Microbial Killing: Oxidants, Proteases Dispatch and Ions

Current Biology, Vol. 12, R357–R359, May 14, 2002, ©2002 Elsevier Science Ltd. All rights reserved.
Microbial Killing: Oxidants, Proteases
and Ions
Rene E. Harrison, Nicolas Touret and
Sergio Grinstein
There is accumulating evidence that two aspects of
the innate immune response, the respiratory burst
and secretion of proteases, are intimately intertwined. A recent study suggests that K+ may be the
missing link. Is it time to merge signal transduction
with biophysics?
Neutrophils, key players of the innate immune response,
deploy a prodigious arsenal of antimicrobial weapons
to eliminate invading organisms. Following engulfment
of the microorganisms into a phagocytic vacuole, neutrophils effect killing by a combination of oxidative,
proteolytic and other mechanisms, including poration
of the microbial membrane by defensins, scavenging
of iron by lactoferrin and alteration of the intraphagosomal pH [1–3]. The primary source of reactive oxygen
metabolites is the NADPH oxidase [4], a multimeric
enzyme complex which catalyzes the one-electron
reduction of oxygen to generate superoxide (Figure
1A). Superoxide is in turn the source of a variety of
other toxic oxidative species, including hydrogen peroxide formed by either spontaneous or catalyzed
dismutation, and hypochlorous acid generated by
myeloperoxidase (Figure 1B).
The importance of the oxidative burst to the
microbicidal response is most evident in chronic
granulomatous disease (CGD). CGD patients, who
have a defective NADPH oxidase, suffer from recurrent
life-threatening bacterial and fungal infections and
abnormal tissue granuloma formation [5,6]. Similarly,
defects in protease secretion compromise the innate
immune response. Elastase-deficient neutrophils
render the host sensitive to infection by gram-negative
bacteria such as Klebsiella pneumoniae and
Escherichia coli [7], while cathepsin G-deficient
neutrophils are reportedly sensitive to Staphylococcus
aureus [8].
While the oxidative and proteolytic arms of the
microbicidal response have been generally thought to
be independent, recent evidence suggests that these
pathways may be closely interlocked. Specifically,
Reeves et al. [8] have recently presented evidence
that optimal release of proteases and other granular
contents into the phagosomal lumen is dependent on
the activity of the NADPH oxidase. Their data are
consistent with earlier observations that degranulation
is abnormal in neutrophils from CGD patients [9,10].
The model presented by Reeves et al. [8] is based
on the notion, originally introduced by Uvnas and
Cell Biology Programme, The Hospital for Sick Children
Research Institute, 555 University Avenue, Toronto, Ontario,
M5G 1X8, Canada. E-mail: [email protected]
PII S0960-9822(02)00859-X
Dispatch
Aborg [11] and championed by Rahamimoff and Fernandez [12], that secreted materials can be packed at
high concentrations within granules without exerting
osmotic forces, by binding to a matrix composed of a
charged proteoglycan hydrogel. During exocytosis,
release of the contents from the matrix, which has the
properties of an ion-exchange resin, requires occupancy of the vacated sites by a counterion, in order to
maintain electroneutrality. Reeves et al. [8] propose
that K+ ions in the lumen of the phagosome function
as the counterion that enables solubilization of proteases and myeloperoxidase from the primary granule
matrix (Figure 2A).
Figure 1A illustrates the mechanism envisaged
by Reeves et al. [8] to account for K+ accumulation in
the phagosome. Briefly, electron transport by the
NADPH oxidase renders the lumen of the phagosome
electronegative with respect to the cytosol [2,6]
(Figure 1A). The presence of conductive pathways
would thus enable intraphagosomal accumulation of
K+ against its chemical gradient. Importantly, Reeves
et al. [8] argue that concentration of K+ well above the
cytosolic level occurs and is required for protease
liberation. In their model, osmotic swelling — which
would result in dilution of K+ — is precluded by
formation of a tight cytoskeletal mesh around the
phagosome, as suggested by the presence of vinculin
and paxillin. The model predicts that inhibition of the
NADPH oxidase would eliminate the electrical force
driving K+ ion accumulation, resulting in impaired protease secretion. Accordingly, inhibition of the oxidase
using diphenylene iodonium depressed the digestion
of bacteria by cellular proteases.
However attractive, the model proposed by Reeves
et al. [8] is not fully consistent with some crucial
observations. Their work relied heavily on the effects
of the ionophore valinomycin to implicate K+ in the
secretory cascade. They found that valinomcyin
impaired the digestion of bacteria and prevented the
phagosomal swelling observed at later times. They
attributed these effects to the ability of the ionophore
to dissipate the K+ concentration gradient generated
by the oxidase. Valinomycin is a conductive ionophore,
however, which should have enhanced, rather than
counteracted, the electrophoretic accumulation of K+
in the phagosomal lumen. Secondly, their microprobe
measurements indicate that the phagosomal and
cytosolic K+ concentrations are similar — 314±22 and
290 mMol g–1, respectively — failing to demonstrate
the proposed large increase in ionic strength. In this
regard, Reeves et al. [8] noted also that phagosomes
containing non-digestible particles, such as latex
beads, failed to swell even after the restraining
cytoskeletal mesh dissociates. Because the phagosomal membrane was shown to be readily water-permeable, failure to swell implies that the osmotic content
of the phagosomal lumen is not markedly different
from that of the cytosol.
Dispatch
R358
Figure 1. Ionic and chemical correlates of
the respiratory burst in neutrophil
ATP
phagosomes.
V-ATPa
V-ATPase
K+
(A) The engulfment of a microorganism
conductance
triggers the activation of the NADPH
.
oxidase, which results in the vectorial
HO + Cl+
+
2H
H + Cltransfer of electrons from NADPH into the
O2
.
O2 phagosome, leading to the development
K+
H+
H+
.SOD H2O2 MPO
HOCl
2 O2
of a transmembrane potential (inside negconductance
H2O2
.O2
O2
H2O
+ -ative) and the reduction of O2, generating
.
+ 2 NO
+ superoxide (O2–). The membrane potential
1O + Cl+ 2
2 ONOO
+
drives H+ and K+ ions into the lumen via
O2
+
NADPH
NADP
DPH
PH
conductive pathways; H+ ions are also
oxidase
pumped in by the V-ATPase. (B) Superoxide is a substrate for superoxide dismuNADPH NADP+ + H+
tase (SOD), which catalyzes the formation
Current Biology
of hydrogen peroxide. Peroxide can be
converted to hypochlorous acid (HOCl) by
myeloperoxidase (MPO). HOCl in turn can generate secondary reactive oxidants including hydroxyl radical (HO.) and singlet oxygen
(1O2). Note that H+ ions are consumed in the dismutation of superoxide and in HOCl formation. Superoxide can also combine with
nitric oxide to form peroxinitrite.
A
B
ADP + P
Other premises established by Reeves et al. [8] are
also difficult to reconcile with current views of phagocytosis and secretion. It is hard to envisage how secretory granules would gain access and fuse with
phagosomes that are surrounded by a tight cytoskeletal mesh that prevents them from swelling. In this
regard, persistence of a cytoskeletal coat around
phagosomes has been implicated in the inability of
lysosomes to fuse with phagosomes containing
mycobacteria [13]. Lastly, release of active proteases
by neutrophils stimulated with soluble agonists has
been amply documented by multiple laboratories [1].
A
B
+ +
+– – +
+ – –
+ – + +–
– +
+ +
+
+ +
– +
+– – +
+ +
+– –
+ +– +
+
+
+ K
+
++
+
+
+
+
C
++
+
ROS
+
+
+
+
+
+
+
+
MPO
Cathepsin G
Elastase
+
+
+ – +
–
–
+
– +
+ – –+ –
+ + +
Cation +
Current Biology
Figure 2. Models for the release of granule contents into the
phagosome.
(A) K+ ions entering from the phagosomal lumen via the fusion
pore leads to the release of proteases and myeloperoxidase
from the granule matrix. According to Reeves et al. [8], hypertonic concentration of K+ in the lumen is required for this
release. (B) Reactive oxygen species (ROS) may enter the
granule and contribute to release of contents. (C) K+ / H+ ions
entering across channels in the granule membrane contribute
to release of contents.
Inasmuch as release occurs into the extracellular
space, an open compartment of constant ionic
strength, it is clear that proper secretion can proceed
without hypertonic accumulation of K+ or other ions. Of
note, the abnormal secretion reported in CGD cells was
recorded under these conditions [9,10], where oxidase
activity cannot contribute to changes in K+ concentration of the medium that bathes the exocytic opening.
How else could the NADPH oxidase influence the
rate and extent of secretion? At least two other
mechanisms can be envisaged. The liberation of proteases from the granules may be stimulated by the
reactive oxygen products themselves (Figure 2B), or
by another ion driven into the phagosome by the electrical potential difference generated by the oxidase. It
is generally believed that protons are the main
counterions that neutralize the vectorial electron
transfer through the oxidase (Figure 1A). Proton
(equivalents) can either be pumped into the phagosome by the vacuolar-type ATPase, or enter passively
via conductive ‘channels’. The molecular nature of the
conductance remains the subject of debate, with
some authors favoring translocation across the
oxidase itself [14,15], while others believe that a separate pathway exists [16,17] (Figure 1A).
Regardless of the molecular identity of the conductive channel, the sum of the protons transported
into the lumen of the phagosome — or the extracelllar space in the case of stimulation by soluble
agonists — is not only required for the maintenance of
electroneutrality, but also provides an important
substrate for the dismutation and myeloperoxidation
reactions (Figure 1B). In this manner, proton translocation may contribute to the release of proteases. By
providing a parallel conductive pathway, valinomycin
is expected to depress the inward flow of protons.
This would in turn reduce the conversion of superoxide to relevant products which may be required for
protease secretion.
Another possible mode of coupling is illustrated in
Figure 2C. As has been suggested for neurosecretory
vesicles [12], it is conceivable that the counterions
required to neutralize the charges on the matrix gel
Current Biology
R359
enter the granules not through the exocytic opening,
but across the granule membrane itself. A high basal
ionic permeability of the granule membrane is consistent with their very acidic pH, as substantive inward
proton pumping requires an equivalent transmembrane
displacement of counterions. Even if the basal granular permeability were low, an increase may occur upon
fusion with the plasma or phagosomal membranes.
In any event, the magnitude of the flux of cations
into the granule lumen would be influenced by the
transmembrane potential. Upon fusion with the target
membrane, the main contributor to the electrical
potential across the membrane of granules would be
the NADPH oxidase, which would propel cations into
their lumen, favoring dissociation ofproteases from
the matrix gel. This model could also explain the
inhibitory effects of valinomycin: by opening cation
permeation pathways in a random fashion the
ionophore would minimize the entry of counterions
into the granule matrix. Clearly, other modes of coupling between the respiratory burst oxidase and
secretion, not involving electrical or ionic considerations, are also possible and should not be discounted.
In summary, the bactericidal mechanisms of
phagocytes appear to have unexpected synergistic
interactions. The ability of the NADPH oxidase to modulate intraphagosomal pH as well as the cytoplasmic
pH had already been documented and suggested to
influence various aspects of the bactericidal response
[6,18–20]. The oxidase was also shown to alter calcium
homeostasis through changes in membrane potential
[6], potentially modifying signalling of secretion,
chemotaxis and phagocytosis. A role for K+ ions is a
tantalizing possibility but, in our view, remains to be
confirmed. Lastly, direct molecular interactions via formation of multisubunit complexes that share signal
transduction modules is an attractive possibility. Thus,
Rac or other transducers recruited to the membrane
upon activation may perform simultaneous tasks in
various aspects of the immune response, such as the
respiratory burst, chemotaxis, phagocytosis and
secretion. Such multi-tasking would inevitably link the
various aspects of the microbicidal response.
References
1. Borregaard, N. (1997). Development of neutrophil granule diversity.
Ann. N.Y. Acad. Sci. 832, 62–68.
2. Hampton, M.B., Kettle, A.J. and Winterbourn, C.C. (1998). Inside the
neutrophil phagosome: oxidants, myeloperoxidase, and bacterial
killing. Blood 92, 3007–3017.
3. Lehrer, R.I. and Ganz, T. (2002). Defensins of vertebrate animals.
Curr. Opin. Immunol. 14, 96–102.
4. Babior, B.M. (1999). NADPH oxidase: an update. Blood 93,
1464–1476.
5. Curnutte, J.T. (1992). Molecular basis of the autosomal recessive
forms of chronic granulomatous disease. Immunodefic. Rev. 3,
149–172.
6. Geiszt, M., Kapus, A. and Ligeti, E. (2001). Chronic granulomatous
disease: more than the lack of superoxide? J. Leukoc. Biol. 69,
191–196.
7. Belaaouaj, A., McCarthy, R., Baumann, M., Gao, Z., Ley, T.J.,
Abraham, S.N. and Shapiro, S.D. (1998). Mice lacking neutrophil
elastase reveal impaired host defense against gram negative bacterial sepsis. Nat. Med. 4, 615–618.
8. Reeves, E.P., Lu, H., Lortat Jacobs, H., C.G.M., M., Bolsover, S.,
Gabella, G., Potma, E.O., Warley, A., Roes, J. and Segal, A.W.
(2002). Mechanism of action of neutrophils. Killing of bacteria by
microbicidal proteases following activation by potassium flux.
Nature 416, 291–297.
9. Gold, S.B., Hanes, D.M., Stites, D.P. and Fudenberg, H.H. (1974).
Abnormal kinetics of degranulation in chronic granulomatous
disease. N. Engl. J. Med. 291, 332–337.
10. Ellis, J.A., Mayer, S.J. and Jones, O.T. (1988). The effect of the
NADPH oxidase inhibitor diphenyleneiodonium on aerobic and
anaerobic microbicidal activities of human neutrophils. Biochem. J.
251, 887–891.
11. Uvnas, B. and Aborg, C.H. (1983). Cation exchange-a common
mechanism in the storage and release of biogenic amines stored in
granules (vesicles)? I. Comparative studies on the uptake of sodium
and biogenic amines by the weak cation (carboxyl) exchangers
Amberlite IRC-50 and Sephadex C-50 and by biogenic (granuleenriched) materials in vitro. Acta. Physiol. Scand. 119, 225–234.
12. Rahamimoff, R. and Fernandez, J.M. (1997). Pre- and postfusion
regulation of transmitter release. Neuron 18, 17–27.
13. Ferrari, G., Langen, H., Naito, M. and Pieters, J. (1999). A coat
protein on phagosomes involved in the intracellular survival of
mycobacteria. Cell 97, 435–447.
14. Demaurex, N., Schrenzel, J., Jaconi, M.E., Lew, D.P. and Krause,
K.H. (1993). Proton channels, plasma membrane potential, and respiratory burst in human neutrophils. Eur. J. Haematol. 51, 309–312.
15. Henderson, L.M. (2001). NADPH oxidase subunit gp91phox: a
proton pathway. Protoplasma 217, 37–42.
16. Nanda, A., Curnutte, J.T. and Grinstein, S. (1994). Activation of H+
conductance in neutrophils requires assembly of components of
the respiratory burst oxidase but not its redox function. J. Clin.
Invest. 93, 1770–1775.
17. DeCoursey, T.E., Cherny, V.V., Morgan, D., Katz, B.Z. and Dinauer,
M.C. (2001). The gp91phox component of NADPH oxidase is not the
voltage-gated proton channel in phagocytes, but it helps. J. Biol.
Chem. 276, 36063–36066.
18. Segal, A.W., Geisow, M., Garcia, R., Harper, A. and Miller, R. (1981).
The respiratory burst of phagocytic cells is associated with a rise in
vacuolar pH. Nature 290, 406–409.
19. Grinstein, S. and Furuya, W. (1986). Cytoplasmic pH regulation in
phorbol ester-activated human neutrophils. Am. J. Physiol. 251,
C55–65.
20. Jankowski, A., Scott, C.C. and Grinstein, S. (2002). Determinants of
the Phagosomal pH in Neutrophils. J. Biol. Chem. 277, 6059–6066.