| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Institute of Medical Microbiology and Hygiene, University of Mainz; and
* Institute of Physiology I, University of Tuebingen, Germany
Correspondence: Institute of Medical Microbiology and Hygiene, Hochhaus am Augustusplatz, D-55101 Mainz, Germany. E-mail: makowiec{at}mail.uni-mainz.de
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: neutrophils • shedding • bacterial toxin
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
(TNF-) [3
, 4
], and there is evidence to
suggest existence of further enzymes with related function
[5
]. The putative family of shedding enzymes is
characteristically inhibitable by peptide-hydroxamate metalloproteinase
inhibitors, e.g., TAPI, and to a lesser degree by conventional
metalloproteinase inhibitors such as 1,10-phenanthroline
[5
]. Several pathways converge to activate the shedding enzymes. One operates via activation of protein kinase C (PKC), so that exposure of cells to phorbol esters results in shedding of many proteins [2 ]. However, shedding can also be induced via PKC-independent pathways, of which at least one is Ca2+-dependent [6 ]. Knowledge on the sequence of events underlying sheddase activation, and the components involved in their regulation, has remained fragmentary.
It was recently found that pore-forming bacterial toxins invoke rapid and massive shedding of CD14 and the interleukin-6 (IL-6) receptor from monocytes. Enhanced sheddase function occurring in the wake of membrane permeabilization depended neither on PKC nor on Ca2+ [7 ]. Released IL-6 receptor retained biological activity and sensitized bystander cells to trans-signaling with IL-6. Hence, shedding events induced by membrane-damaging agents are potentially relevant in the context of disease pathogenesis.
In this work, we elected to study the shedding of L-selectin (CD 62L)
from granulocytes after treatment with streptolysin O (SLO). L-selectin
is constitutively expressed on lymphocytes, monocytes, and
granulocytes, and is responsible for their tethering to and rolling on
endothelial cells [8
, 9
]. L-selectin can be
cleaved by a sheddase to yield a soluble fragment that contains the
functional lectin and epidermal growth factor domains
[10
]. Various stimuli trigger L-selectin shedding,
including chemoattractants, phorbol esters, TNF-, C-reactive
protein, anti-inflammatory drugs, and cross-linking of the L-selectin
molecule itself [11
12
13
14
15
]. Analysis of L-selectin
shedding afforded the advantage that the process could easily be
followed and quantified by flow cytometry.
We report that the pore-forming toxin SLO induces rapid and massive shedding of L-selectin. A novel path to the shedding process is identified that involves the Ca2+-independent activation of neutral sphingomyelinase. Evidence is presented that ceramide, a major product of sphingomyelinase action, assumes a central role in this pathway.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The metalloproteinase inhibitor TAPI (N-{D,L-[2-(hydroxyaminocarbonyl)methyl]-4-methylpentanoyl}L-3-(2’-naphthyl)-alanyl-L-alanine, 2-amino-ethylamide) was prepared at Immunex [18 ]. Aprotinin, leupeptin, and Pefabloc [4-(2-aminoethylbenzenesulfonylfluoride)] were obtained from Boehringer Mannheim, Mannheim, Germany. The L-selectin enzyme-linked immunosorbent assay (ELISA) was supplied by R & D Systems, Wiesbaden, Germany.
Cell preparation and experimental protocol
Human granulocytes were isolated from heparinized blood of
healthy volunteers after conventional procedures. In brief, 1 vol of
dextran (Pharmacia Biotech, Uppsala, Sweden; 4.5%, in isotonic salt
solution, pH 7.4) was added to 5 vol whole blood, and cells were
allowed to sediment in tilted plastic centrifugation tubes for 40 min
at 37°C. The erythrocyte-depleted supernatants were transferred in
4-mL aliquots to 4-mL Ficoll-Hypaque gradients (Pharmacia Biotech) and
centrifuged for 20 min at 400 g, 20°C. Erythrocytes
contaminating the cell pellet were lysed in a buffer containing
NH4Cl (150 mM), KHCO3 (10 mM), and EDTA (10
mM), pH 7.4. Cells were then resedimented (100 g, 10 min),
washed, suspended in sterile Hanks’ buffered salt solution, and kept
on ice. The cell preparations contained <3% contaminating
lymphocytes, and <4% nonviable cells as determined by staining with
Trypan blue.
Cell surface L-selectin (CD 62L) was quantified by flow cytometry using fluorescein isothiocyanate (FITC)-conjugated monoclonal antibodies (clone Dreg 56, Coulter-Immunotech, Hamburg, Germany). Granulocytes were incubated with anti-CD-62L (dilution 1:500) for 60 min at 4°C and washed twice.
After a 5-min pre-incubation at 37°C in the presence of inhibitors, cells were stimulated with N-formyl-methionyl-leucyl-phenylalanine (fMLP), SLO, sphingomyelinase from Staphylococcus aureus, C2-ceramide, or C2-dihydroceramide (Calbiochem-Novabiochem, Bad Soden, Germany) for the desired time. Flow cytometric analysis was performed using FACScan and Lysys II Software (Becton Dickinson).
The human myelomonocytic cell line JOSK-M was cultured in RPMI 1640 medium supplemented with 10% fetal calf serum, 10 mM HEPES (pH 7.4), 2 mM L-glutamine, 1 mM sodium pyruvate, 100 µM nonessential amino acids, 100 U/mL penicillin, 100 µg/mL streptomycin, and 50 µM ß-mercaptoethanol. The cells were differentiated in vitro by 4-day incubation with 50 nM retinoic acid and 10 nM bufalin.
Determination of neutral sphingomyelinase activity
Cells were lysed in a buffer containing HEPES (20 mM, pH 7.4),
EDTA (2 mM), dithiothreitol (5 mM), MgCl2 (10 mM),
Na3VO4 (0.1 mM), ß-glycerophosphate (10 mM),
ATP (7.5 mM), leupeptin (10 µM), and Triton-X 100 (Roth, Karlsruhe,
Germany; 0.2 %). Samples were sonicated three times for 10 s. The
substrate [14C]sphingomyelin (NEN Life Science Products,
Cologne, Germany; 0.5 µCi/sample, 54.5 mCi/mmol) was dried and
solubilized by sonication for 10 min in a water bath in HEPES (20 mM,
pH 7.4), MgCl2 (1 mM), and Triton-X 100 (0.2%). Aliquots
were added to the lysed samples.
After 30 min at 37°C, samples were extracted with chloroform/methanol/H2O (v/v/v, 4:2:1). The upper phase was collected, and radioactivity was determined in a liquid scintillation counter.
Activity of acidic sphingomyelinase
Cells were lysed in a sonication buffer containing Tris (50 mM,
pH 7.4), bacitracin, benzamidine, Na3VO4 (1 mM
each), soybean trypsin inhibitor (0.1 mg/mL), aprotinin, leupeptin (10
µg/mL each), and Triton X-100 (0.2%). Samples were immediately
sonicated three times for 10 s each.
After centrifugation at 600 g for 5 min, the supernatants were resuspended in a lysis buffer containing Tris (50 mM, pH 7.4), Na3VO4 (1 mM), aprotinin and leupeptin (10 µg/mL each), Triton X-100 (1%), and NP-40 (3%). The acidic sphingomyelinase was immunoprecipitated using a goat-anti-acidic sphingomyelinase serum [19 ] followed by protein A/G-coupled agarose (Santa Cruz Biotechnology, Santa Cruz, CA). Samples were washed three times each in the lysis buffer and in a buffer consisting of sodium acetate (50 mM, pH 5.0), Na3VO4 (1 mM), aprotinin and leupeptin (10 µg/mL each), and Triton X-100 (0.2%).
[14C]sphingomyelin (NEN-DuPont, 0.5 µCi/sample, 54.5 mCi/mmol) was dried and solubilized by 10-min water bath sonication in EDTA (1.3 mM), sodium acetate (250 mM, pH 5.0), and NP-40 (0.05%). Samples were processed as above.
Quantification of ceramide
Cellular lipids were extracted with chloroform/methanol/HCl
(v/v/v, 100:100:1), and the organic phase was dried under nitrogen.
Samples were re-extracted, dried again, and solubilized by sonication
for 10 min in a water bath in n-octyl-ß-glucopyranoside
(w/v, 7.5%), cardiolipin (5 mM), and DETAPAC (1 mM). Purified
Escherichia coli diacylglycerol-kinase (Calbiochem; 40
µg/mL) in 70 µL reaction buffer containing imidazole-HCl (100 mM,
pH 6.6), NaCl (100 mM), MgCl2 (25 mM), EGTA (2 mM) plus
[32P]
ATP (ICN Biochemicals, Eschwege, Germany) and ATP
(1 µm) was added. The samples were incubated for 30 min at 22°C,
after which the kinase reaction was stopped by adding
chloroform/methanol/HCl (v/v/v, 100:100:1, 1 mL), salt solution [170
µL, HEPES (10 mM, pH 7.2), NaCl (135 mM), CaCl2 (1.5 mM),
MgCl2 (0.5 mM), glucose (5.6 mM)], and EDTA (30 µL, 100
mM). The organic phase was dried, resuspended in chloroform/methanol
(v/v, 1:1), subjected to thin-layer chromatography on a silica 60 gel
plate (Merck, Darmstadt, Germany), using chloroform/methanol/acetic
acid (v/v/v, 65:15:5) as solvent. Autoradiography was performed, and
the spots were scraped into vials and counted in a liquid scintillation
counter.
To directly measure ceramide release JOSKM cells were labeled with [3H]serine (25 mCi/mmol, NEN) for 36 h, washed, and stimulated with SLO. Lipids were extracted with CHCl3/CH3OH/H2O/pyridine (60:160:6:1), the lower organic phase was dried, and phospholipids were degraded by 2-h incubation with methanolic NaOH at 37°C. After re-extraction, the lower organic phase was dried again, resuspended in CHCl3/CH3OH (95:5) and separated on silica G60 plates with CHCl3/CH3OH/CaCl2 (60:35:8, v/v/v). Ceramide was detected by co-migration with a C16 ceramide standard.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
70% (Fig. 2B)
.
|
|
shedding
proteinase, effectively inhibited shedding of L-selectin at a
concentration of 10 µM (Fig. 3
).
|
Shedding of L-selectin by SLO depends on membrane permeabilization
To differentiate whether L-selectin shedding was dependent on
toxin binding or on membrane permeabilization, we employed a
cysteine-substitution mutant (SLO N402C). After chemical modification
of the sulfhydryl group, this mutant retains cell-binding activity but
lacks pore-forming properties as a result of its inability to undergo
correct polymerization [16
, 21
]. It was
found that SLO N402C failed to trigger release of L-selectin
(Fig. 4
). Thus, shedding was dependent on membrane permeabilization. In a
further experiment, SLO N402C was first applied to the cells, and
wild-type SLO was subsequently added at 15 min. It was found that
L-selectin shedding occurred to the same extent as observed with WT SLO
alone. These results went further to show that L-selectin shedding
induced by SLO was not due to simple interaction of the toxin with
specific membrane binding sites.
|
The activities of neutral and acidic sphingomyelinase were measured in cell lysates of granulocytes after stimulation with SLO and fMLP. fMLP failed to activate either sphingomyelinase (data not shown). However, cells permeabilized with SLO showed a 1.5-fold increase in activity of neutral sphingomyelinase (Fig. 5 ). In contrast, no significant change in the activity of the acidic sphingomyelinase was observed.
|
|
|
|
The kinetics of shedding induced by ceramide resembled those observed after application of SLO, and results of inhibition experiments were also equivalent. Staurosporine did not block ceramide-induced shedding, and 1,10 phenanthroline only weakly suppressed L-selectin cleavage. In contrast, TAPI effectively inhibited shedding induced by C2-ceramide (Fig. 9 ) and by other ceramides (not shown). All other proteinase inhibitors tested were ineffective (not shown).
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
SLO was used as the permeabilizing agent for several reasons. The toxin binds to membrane cholesterol and rapidly oligomerizes to form transmembrane pores [22 23 24 ]. In contrast to the pore-forming hemolysin of E. coli, SLO does not require Ca2+ for its activity [25 ]. Permeabilization experiments could therefore be conducted in the presence of EDTA, and these showed that L-selectin shedding occurring in response to SLO was not Ca2+-dependent. Utilization of SLO provided another asset because mutant toxins are now available that retain their binding function but cannot form pores due to their inability to correctly polymerize [16 , 21 ]. With the use of one such mutant, it was shown that L-selectin shedding was not induced by toxin binding alone, but required membrane permeabilization to occur.
The collective evidence indicates that SLO-induced L-selectin shedding occurs via activation of a physiological sheddase. Foremost, L-selectin liberation was effectively suppressed by TAPI, the classical sheddase inhibitor, whereas the conventional metalloproteinase inhibitor 1,10-phenanthroline, was relatively ineffective. This inhibition pattern is typical of the physiological sheddases [5 ]. The fact that toxin-induced shedding was entirely refractory to inhibition by all other proteinase inhibitors tested may, in retrospect, not appear surprising, but it is noteworthy nevertheless. In the presence of Ca2+, membrane permeabilization triggers granule secretion in neutrophils [26 ], so the possibility that lysosomal proteases might have been responsible for L-selectin cleavage required consideration. Our findings indicated that secreted cellular proteases were not involved, inasmuch as shedding also occurred in the presence of EDTA.
Surprisingly diverse pathways seem to converge to trigger sheddase activity. One pathway is Ca2+-dependent, but this clearly was not involved in our case. The second well-established sequence involves PKC; thus, phorbol esters are generally potent inducers of shedding, and activation of PKC via binding of physiological ligands to their receptors (e.g., fMLP, chemokines) naturally causes shedding [12 ]. We readily confirmed that fMLP induces L-selectin cleavage in neutrophils that was totally suppressed by staurosporine, the inhibitor of PKC.
In contrast, shedding provoked by membrane permeabilization was entirely refractory toward inhibition by staurosporine. Analogous findings were previously made in our shedding experiments with monocytes [7 ]. Thus, membrane permeabilization must be linked to sheddase activation via yet another pathway. In the literature, there have been reports on shedding processes that were apparently independent of PKC and Ca2+ [6 , 7 ]; however, little is known of the underlying mechanisms and putative mediators. The present work identifies a novel component that appears to play a pivotal role in this scenario. Our data indicate that neutral sphingomyelinase is important through its capacity to generate ceramide. The latter may be the molecule that directly causes shedding. This contention derives from the observation that shedding could be induced through application of an exogenous, cell-permeable ceramide analog. Furthermore, L-selectin cleavage occurred when cells were treated with exogenous sphingomyelinase from S. aureus. In contrast, other phospholipases (phospholipase A2, phospholipase D) induced little or no shedding.
Quantitative aspects of neutral sphingomyelinase stimulation warrant
comment. The finding that SLO caused an increase in activity of
40–50% within 1–5 min is comparable to results obtained in other
investigations using different agonists. Thus, Brenner et al. reported
an increase in neutral sphingomyelinase activity of 25% within 10 min
after cross-linking of L-selectin [27
]. Similarly,
Tomiuk et al. calculated a 20–25% activity increase within 10 min
after stimulation of U937 cells with TNF- [28
]. The
concomitant increase in ceramide generation observed in our study was
also compatible with the literature, although the kinetics were faster.
Thus, a 50% increase of ceramide in the cells was observed 10 min
after SLO stimulation. TNF treatment of MCF-7 cells resulted in a
comparable, but slower reaction [29
]. Yayadev et al.
similarly observed a 50% increase in cellular ceramide after 80 min of
stimulation of HLA60 cells with arachidonic acid [30
].
Pilot experiments in our laboratory indicate that arachidonic acid
indeed induces shedding of L-selectin in granulocytes, but at slower
rates compared with SLO.
Methods to correctly determine cellular ceramide levels have been the subject of recent debates [31 , 32 ]. As suggested by Perry and Hannun [32 ], we performed the DGK assay with an excess of enzyme. In addition, we validated the amount of ceramide by direct labeling of ceramide with [3H]serine and obtained similar results with both methods.
How ceramide promotes shedding remains to be clarified. Basically, it is possible to envisage three distinct mechanisms. First, ceramide may interact directly with the sheddase and enhance its activity. Second, ceramide may act as a second messenger to trigger further events that ultimately lead to shedding. Third, ceramide may alter the vertical or horizontal organization of membrane components and so promote access of the shedding substrates to the enzyme.
The cause of sphingomyelinase activation is unknown. One possibility was that membrane depolarization played a key role. However, experiments in which cells were immersed in 100 mM KCl yielded negative results, and no shedding of L-selectin was observed. Intracellular glutathion has been reported to suppress the activity of neutral sphingomyelinase [29 , 33 ]. Membrane permeabilization leads to rapid loss of glutathion (unpublished data), and it is therefore tempting to postulate that this directly results in a selective derepression of neutral sphingomyelinase. However, pilot experiments have not yet enabled us to verify this hypothesis, and further experimentation is required to generate a firm basis on which an explanatory model can be built. Independent of the resolution of this problem, our findings reveal a potentially significant novel consequence of bacterial pore-forming toxins and of sphingomyelinase action in general. The concentration of SLO required to induce shedding was more than a magnitude below the concentration found in an overnight culture supernatant of the bacteria. Further to its action from the fluid phase, SLO likely can provoke pathology when cell-adherent bacteria secrete the toxin. Hence, SLO-induced shedding may be of significance in both intravascular and extravascular compartments. Depending on the location, both diapedesis and migration through tissues could be impeded. The significance of SLO-induced shedding in relation to sheddase activation by other microbial products cannot be gauged at present.
Sphingomyelinase is produced by many bacterial pathogens, but the biological relevance of the enzyme is not well understood. Our results may be relevant in the context of microbial pathogenesis because shedding of L-selectin can be expected to adversely affect vascular adhesion and emigration of cells. Shedding of other biologically important molecules from various cell targets may contribute to pathological events encountered during infections with microorganisms that elaborate pore-forming toxins and sphingomyelinase.
|
ACKNOWLEDGEMENTS |
|---|
Received January 23, 2000; revised July 10, 2000; accepted July 11, 2000.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
from cells Nature 385,729-733[Medline]
Nature 385,733-736[Medline]
precursor J. Biol. Chem. 266,5769-5773 J. Biol. Chem. 269,5757-5763-induced cell death J. Biol. Chem. 273,11313-11320This article has been cited by other articles:
|
|
 |
|
  A. I. Iliev, J. R. Djannatian, R. Nau, T. J. Mitchell, and F. S. Wouters Cholesterol-dependent actin remodeling via RhoA and Rac1 activation by the Streptococcus pneumoniae toxin pneumolysin PNAS, February 20, 2007; 104(8): 2897 - 2902. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. Chen, A. Hayashida, A. E. Bennett, S. K. Hollingshead, and P. W. Park Streptococcus pneumoniae Sheds Syndecan-1 Ectodomains through ZmpC, a Metalloproteinase Virulence Factor J. Biol. Chem., January 5, 2007; 282(1): 159 - 167. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. W. Park, T. J. Foster, E. Nishi, S. J. Duncan, M. Klagsbrun, and Y. Chen Activation of Syndecan-1 Ectodomain Shedding by Staphylococcus aureus {alpha}-Toxin and {beta}-Toxin J. Biol. Chem., January 2, 2004; 279(1): 251 - 258. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
V. Matthews, B. Schuster, S. Schutze, I. Bussmeyer, A. Ludwig, C. Hundhausen, T. Sadowski, P. Saftig, D. Hartmann, K.-J. Kallen, et al. Cellular Cholesterol Depletion Triggers Shedding of the Human Interleukin-6 Receptor by ADAM10 and ADAM17 (TACE) J. Biol. Chem., October 3, 2003; 278(40): 38829 - 38839. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
