| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
*
Vascular Research Division, Department of Pathology, Brigham and Women’s Hospital,
Cardiovascular Research Center, Massachusetts General Hospital-East, Harvard Medical School, Boston, MA
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The cytoplasmic domains of a number of other adhesion molecules have been either implicated in or directly shown to activate intracellular signal transduction. Perhaps the best characterized of all cell adhesion molecules in this regard have been the integrins. The various members of the integrin superfamily have been demonstrated to activate tyrosine and serine/threonine kinase cascades, to induce cytoskeletal reorganization via interaction with cytoskeletal proteins, and to modulate their own affinities and adhesive functions by interaction with tail binding proteins such as cytohesin-1 and ß3-endonexin (11, 12). Members of the cadherin family also appear to support a signaling function. The cadherins are localized to adherens junctions at regions of cell-cell contact, and their cytoplasmic tails associate with a number of putative signaling proteins, including ß-catenin, plakoglobin, and p120 (13). A third example of cell adhesion molecules thought to be important signal transducers is the Ig superfamily, which includes ICAM-1 and platelet/endothelial cell adhesion molecule-1 (PECAM-1).5 In brain microvessel endothelial cells, ICAM-1 stimulates Src activity and thereby induces the tyrosine phosphorylation of a number of Src substrates, including the actin binding protein cortactin (14). The cytoplasmic domain of PECAM-1 serves as a substrate for Src and, when tyrosine-phosphorylated, can bind the protein tyrosine phosphatase SH2-containing phosphatase-2 (SHP-2) (15, 16). Recent data suggest that the selectins are also active participants in the process of signal transduction (6). To date, the best characterized family member in this respect has been L-selectin. The cytoplasmic domain of L-selectin has been shown to be tyrosine phosphorylated (17), and ligation of L-selectin can lead to mitogen-activated protein (MAP) kinase activation (17), association of Grb2 and Sos (17), and actin polymerization dependent upon Ras and Rac2 (18, 19) within leukocytes.
We have recently shown that leukocyte adhesion to vascular endothelium
can induce a biochemical and biophysical association of the cytoplasmic
domain of E-selectin with components of the endothelial actin
cytoskeleton, such as 
-actinin, vinculin, and filamin, as well as
certain regulatory elements, such as focal adhesion kinase (FAK) and
its substrate paxillin (20), thus suggesting a previously unrecognized
signaling role for E-selectin. In the current study we have
investigated the signaling function of E-selectin during leukocyte
adhesion. We report here that E-selectin in activated HUVEC in culture
is constitutively phosphorylated on serine residues within its
cytoplasmic domain, and that leukocyte adhesion can induce an
enzymatically regulated dephosphorylation of E-selectin. Interestingly,
a similar pattern of dephosphorylation of E-selectin is observed when
cell surface E-selectin is cross-linked using a mAb directed to the
extracellular ligand binding domain followed by a polyclonal IgG.
Moreover, the binding of beads coated with recombinant PSGL-1, a
glycoprotein ligand for E-selectin, to the endothelial cell surface can
induce a similar pattern of dephosphorylation of E-selectin. Thus, the
interaction of leukocytes, cross-linking Abs, or complex carbohydrate
ligands with the extracellular domain(s) of E-selectin can regulate the
phosphorylation state of the cytoplasmic domain of this molecule. These
data suggest a novel transmembrane outside-in signaling role for
E-selectin that may have implications for the orchestration of
endothelial cell responses in the leukocyte-endothelial adhesion
cascade.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Medium 199 and Dulbecco’s PBS were obtained from M. A. Bioproducts (Walkersville, MD). FBS and DMEM were purchased from Life Technologies (Grand Island, NY). Fifteen- and thirty-five-millimeter diameter culture plates and 96-well microtiter culture plates were purchased from Costar (Cambridge, MA). Endothelial cell growth factor was obtained from Biomedical Technologies (Stoughton, MA). Recombinant human IL-1ß was a gift from Biogen (Cambridge, MA). mAbs H18/7 and W6/32 are both of isotype IgG2a and were prepared as described previously (21, 22). The recombinant PSGL-1-human IgG Fc chimera was provided by Dr. Dale Cummings (Genetics Institute, Cambridge, MA). Biscarboxyethyl-carboxyfluorescein acetoxymethyl ester was purchased from Molecular Probes (Eugene, OR). [32P]orthophosphate was obtained from New England Nuclear (Boston, MA). Okadaic acid was purchased from Calbiochem (La Jolla, CA). Fluorescein-conjugated goat anti-mouse IgG was purchased from Caltag (Burlingame, CA).
Cell culture
The human kidney fibroblast cell line 293 was obtained from the American Type Culture Collection (Rockville, MD) and grown in DMEM supplemented with 10% FBS, 100 U/ml penicillin, 100 µg/ml streptomycin, and 2 mM L-glutamine. HL-60, a promyelocytic human leukocyte cell line, was obtained from American Type Culture Collection and grown in RPMI 1640 medium supplemented with 10% FBS, 100 U/ml penicillin, 100 µg/ml streptomycin, and 2 mM L-glutamine. HUVEC were isolated and established in culture as previously described (23). Primary cultures were serially passaged (<1:3 split ratio) and maintained in medium 199 buffered with 25 mmol/l HEPES and supplemented with 20% FBS, endothelial cell growth factor (25 µg/ml), 100 U/ml penicillin, 100 µg/ml streptomycin, and porcine intestinal heparin (50 µg/ml). For experimental purposes, HUVEC subcultured to passage 2 or 3 were used.
[32P]Orthophosphate labeling of HUVEC
Confluent monolayers (consisting of approximately 3.0–6.0 x 106 cells) of cytokine activated or adenoviral vector-transduced HUVEC in 35-mm culture plates were preincubated in phosphate-free RPMI 1640 with 1% dialyzed FBS for 1 to 2 h, and then metabolically labeled in the same medium supplemented with 333 µCi/ml of [32P]orthophosphate for 90 min. The labeled monolayers were washed three times with regular RPMI 1640 with 1% FBS and incubated with various reagents, as described below. Cells were then rapidly lysed in 0.25 ml ice-cold lysis buffer (0.1 M Tris-HCl, 0.15 M NaCl, and 5 mM EDTA, pH 7.4) containing 1% Triton X-100, 10 µg/ml leupeptin, 60 U/ml aprotinin, and 1 mM PMSF. In some experiments, HL-60 cells (2 x 106/35-mm dish) or H18/7 mAb to E-selectin (10 µg/35-mm dish) were added to the labeled monolayers before lysis. In other experiments, the effects of phosphatase inhibitors were investigated by preincubation of the labeled monolayers with sodium vanadate (100 µM) and/or okadaic acid (100 nM) for 10 min before addition of HL-60.
Recombinant PSGL-1 beads
Recombinant PSGL-1 beads were prepared as follows. Recombinant
PSGL-1-human IgG Fc chimera was provided by Dr. Dale Cummings (Genetics
Institute). Tosyl-activated Dynabeads M-450 (5 x
106; diameter, 4.5 µm) were coated overnight with
100 µg of protein A (Zymed, South San Francisco, CA) in 0.1 M
carbonate buffer, pH 9.4. After washing twice with DPBS, 0.25 µg of
recombinant PSGL-1-human IgG Fc chimera was added to the beads and
incubated for 1 h. The beads were washed twice with DPBS
containing 1% BSA (DPBS/1% BSA) and then blocked with 200 µg/ml of
human IgG1 
in DPBS/1% BSA. After washing once with DPBS/1% BSA,
the beads were kept in this buffer. For binding to IL-1-activated HUVEC
monolayers, 106 beads were used/35-mm dish.
Immunoprecipitation and cell surface biotinylation
Total cell lysates of labeled HUVEC monolayers were centrifuged at 12,000 x g for 20 min, and the supernatants were precleared with 50 µl of cyanogen bromide-activated Sepharose beads for 1 h at 4°C. Mixtures were centrifuged at 12,000 x g for 5 min, and the supernatants were incubated with 2 or 3 µg of mAb H18/7, directed against human E-selectin for 1 h at 4°C. Fifty micrograms of goat anti-mouse IgG-coupled Sepharose beads, preincubated with unlabeled HUVEC lysates, were then added to the H18/7-lysate mixture and incubated for 1 h at 4°C. The beads were collected by centrifugation at 12,000 x g for 5 min and washed twice each with three different washing buffers (600 mM NaCl, 0.1 M Tris-HCl, and 5 mM EDTA, pH 7.4 containing 1% Triton X-100; 300 mM NaCl, 0.1 M Tris-HCl, and 5 mM EDTA, pH 7.4, containing 1% Triton X-100; 150 mM NaCl, 0.1 M Tris-HCl, and 5 mM EDTA, pH 7.4). Proteins were eluted from beads by boiling for 5 min in 30 µl of 0.5 M Tris-HCl (pH 6.8) containing 0.4% SDS and subjected to SDS-PAGE followed by autoradiography.
The amount of cell surface E-selectin present during the various experimental manipulations was monitored by biotinylation of HUVEC surface proteins, followed by immunoprecipitation of E-selectin from whole cell lysates as described above. Briefly, confluent HUVEC monolayers were washed three times with ice-cold DPBS before being incubated for 30 min with a solution of 0.5 mg/ml of sulfo-N-hydroxysulfosuccinimidyl biotin (Pierce, Rockford, IL) on ice. This reaction was quenched by washing the cells twice with cold DPBS containing 1 mg/ml lysine, followed by a single wash with cold DPBS. The cells were lysed, and E-selectin was immunoprecipitated and subjected to SDS-PAGE. Following SDS-PAGE, proteins were transferred to polyvinylidene difluoride membranes. Membranes were incubated for 1 h in blocking buffer (DPBS/3% BSA/0.1% Tween-20) and then incubated with a streptavidin-horseradish-peroxidase conjugate (Amersham, Arlington Heights, IL) diluted 1/1500 in blocking buffer for 1 h. The membranes were washed twice in 1x blocking buffer, twice in 0.1x blocking buffer, and twice in DPBS with 0.1% Tween-20 (5 min/wash). Peroxidase was detected using enhanced chemiluminescence (Amersham) according to the manufacturer’s protocols.
Phosphoamino acid analysis
HUVEC monolayers in 100-mm dishes were infected with the adenoviral E-selectin expression construct, AdRSV(wt-E). Forty-eight hours after infection, the cells were washed free of virus and preincubated for 1 h in phosphate-free RPMI 1640 containing 1% dialyzed FBS. The HUVEC were then incubated for 4 h (to maximize incorporation) with [32P]orthophosphate at a concentration of 200 µCi/ml in 5 ml of phosphate-free RPMI 1640 containing 1% dialyzed FBS. After incorporation of the radiolabeled phosphate, the cells were washed three times with RPMI 1640 supplemented with 1% FBS and subjected to HL-60 adhesion (1.2 x 107 HL-60/100-mm dish) for 10 min. The cells were then quickly washed with RPMI and lysed in 500 µl of ice-cold lysis buffer (0.1 M Tris-HCl, 0.15 M NaCl, and 5 mM EDTA, pH 7.4) containing 1% Triton X-100, 10 µg/ml leupeptin, 60 U/ml aprotinin, and 1 mM PMSF. The lysates were clarified and subjected to immunoprecipitation using the anti-E-selectin mAb H18/7. The resulting immunoprecipitates were resolved by SDS-PAGE and transferred to nitrocellulose. After a brief exposure to x-ray film, the bands corresponding to E-selectin were identified and excised from the nitrocellulose membrane. Next, the membrane slices were incubated in 200 µl of 5.7 M HCl for 1 h at 100°C to effect partial acid hydrolysis of the radiolabeled E-selectin. The hydrolysis products were then lyophilized, resuspended in 500 µl of dH2O, and lyophilized for a second time. The lyophilized material was resuspended in 5 µl of a solution containing 15 parts pH 1.9 buffer (2.2% formic acid and 1.4 M acetic acid, in dH2O) and 1 part nonradioactive phosphoamino acid standards (1.0 mg/ml each of phosphoserine, phosphothreonine, and phosphotyrosine). Finally, the samples were resolved by two-dimensional, high voltage, thin layer electrophoresis at pH 1.9 in the first dimension and at pH 3.5 in the second dimension using a Hunter thin layer electrophoresis system (HTLE-7000, CBS Scientific, Del Mar, CA) (24). The positions of the nonradioactive phosphoamino acid standards were identified by spraying the thin layer electrophoresis plate with ninhydrin followed by baking at 80°C for 10 min. The relative positions of the radiolabeled phosphoamino acids were then localized by autoradiography.
Construction of wild-type and mutant E-selectin recombinant adenoviral vectors
Three replication-defective recombinant type 5 adenoviruses were
used in these studies: AdRSV(wt-E), AdRSV(
Cyto-E), and
AdRSVßgal. AdRSVßgal was constructed as described previously (29).
Both AdRSV(wt-E) and AdRSV(Cyto-E) use the pJM17 backbone, contain
E1/E3 deletions, and were generated as follows. The cDNAs encoding
either a full-length human E-selectin (wt-E-selectin) or a cytoplasmic
deletion mutant of E-selectin (Cyto-E-selectin) (20) were ligated
into pAdRSV4 (provided by Dr. Beverly Davidson, University of Iowa,
Ames, IA) (25) between the RSV long terminal repeat and the SV40
polyadenylation signal. The resulting plasmids were cotransfected with
the plasmid pJM17 in 293 cells (26). Plaques were isolated and
characterized for protein expression by a fluorescent immunobinding
assay as described previously (27), using a saturating concentration of
H18/7 (anti-E-selectin mAb) and FITC-conjugated
F(ab')2 goat anti-murine IgG (Caltag, San
Francisco, CA). Viruses were propagated in 293 cells, purified by
double cesium gradient centrifugation, and titrated by plaque assay in
293 cells as previously described (28). Stock titers ranged from
1010 to 1012 plaque-forming units/ml, with a
particle to plaque-forming unit ratio of approximately 10. HUVEC were
infected as described previously (29).
Construction of serial cytoplasmic deletion mutants of E-selectin
To generate serial deletion mutants of E-selectin cDNA, a
PCR-directed mutagenesis approach was used. A cytoplasmic deletion
mutant (Cyto) E-selectin construct was constructed as previously
described (20). Three new mutants (E1953, E1925, and E1899) were
generated as follows (see Fig. 4
A). In the E1953
mutant, a stop codon was substituted in place of Lys605,
resulting in deletion of the cytoplasmic domain after
Lys605. In the E1925 mutant, a stop codon was introduced at
Gln595. In the E1899 mutant, a stop codon was substituted
for Lys587. A forward primer F2
(5-GGTTTGGTGAGGTGTGCTCATTC-3) and three different reverse primers,
R4 (5-GCGTTAACTTACTTTGCTTTCCGTAA-3), R5
(5-GCGTTAACTTAGCAGCTGCTGGCAGG-3), and R6
(5-GCGTTACTATTGGTAGCTTCCATCT-3), were used to amplify three
fragments from E-selectin cDNA. All three reverse primers (R4, R5, and
R6) contained both a stop codon and an HpaI site at their 3'
ends to make mutated E-selectin cytoplasmic domain fragments. In all
cases, the amplified fragments were ligated into a TA cloning vector
(Invitrogen, San Diego, CA), and the mutated sequences were confirmed
by the Sanger dideoxynucleotide chain termination method of sequencing.
These fragments were prepared by digestion with
HpaI/EcoRI and were used to replace the
corresponding HpaI/EcoRI fragments of the
wild-type E-selectin cDNA in the expression vector pCDM8
(Invitrogen).
|
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Immunoprecipitation of E-selectin from
[32P]orthophosphate-labeled, activated HUVEC
revealed constitutive phosphorylation of E-selectin in the absence of
additional stimuli (Fig. 1A). Adhesion of HL-60
cells induced a time-dependent dephosphorylation of this molecule (Fig. 1A). Prolonged adhesion of HL-60 for up to 35 min
resulted in dephosphorylation of E-selectin comparable to that
demonstrated at 10 min (data not shown). Radiolabeled HUVEC incubated
in parallel in the absence of leukocyte adhesion demonstrated no
spontaneous loss of the incorporated [32P]phosphate from
E-selectin during the same periods of incubation (data not shown).
|
B). Taken together, these data suggest that the
observed reduction of phosphorylation illustrated in Figure 1 reflects
changes in the phosphorylation state of cell surface E-selectin induced
by HL-60 adhesion rather than either spontaneous loss of phosphate or
changes in the amount of E-selectin protein at the cell surface. Ab cross-linking induces dephosphorylation of cell surface E-selectin in HUVEC
To better understand the nature of the stimulus required to induce
dephosphorylation of E-selectin, we next examined the effect of
cross-linking cell surface E-selectin with mAb. We incubated
[32P]orthophosphate-labeled, IL-1ß-activated HUVEC
with H18/7 (mAb to E-selectin) in the absence or the presence of a
secondary goat anti-mouse polyclonal IgG. H18/7 binds to the
extracellular lectin domain of E-selectin and is capable of blocking a
significant proportion of leukocyte adhesion (21). As shown in Figure 2, this Ab alone did not induce
dephosphorylation of E-selectin (97% of control). However, addition of
goat anti-mouse IgG (to physically cross-link E-selectin at the
cell surface) for 10 min induced dephosphorylation of E-selectin to a
comparable extent (48% of control) as HL-60 adhesion (see Fig. 1A). As was observed with leukocyte adhesion, the
biotinylated cohort of cell surface E-selectin was also not
significantly altered by Ab-mediated cross-linking (100% at time zero,
91.5 ± 12% at 5 min, 95.0 ± 13% at 10 min, and 93.0
± 10% at 15 min). As a control for the specificity of this stimulus,
the effect of W6/32 (a mAb to HLA-A,B), either by itself or
cross-linked, was also examined. The binding of this Ab and its
cross-linking at the endothelial surface had no effect on the state of
E-selectin phosphorylation/dephosphorylation (Fig. 2). These data
indicate that Ab-induced cross-linking of cell surface E-selectin on
HUVEC is a selective stimulus for its dephosphorylation.
|
Finally, we examined the ability of natural ligand binding to
induce E-selectin dephosphorylation. The glycoprotein PSGL-1 bears the
carbohydrate moiety sialyl-Lewis X (sLex) and has been
reported to bind both P- and E-selectin (30). A recombinant chimeric
form of PSGL-1 was chosen to test whether ligand-dependent interaction
with the extracellular domain of E-selectin can induce
dephosphorylation of its cytoplasmic tail. The PSGL-1-IgG chimera was
coupled to tosyl-activated magnetic beads, and human IgG-coupled beads
were used as a negative control. FACS analysis of the PSGL-1-coated
beads using an anti-PSGL-1 mAb confirmed that the recombinant
protein had uniformly coated the beads at a relatively high density
(data not shown), and when examined in a parallel plate flow chamber,
the PSGL-1-coated beads rolled on and adhered to activated HUVEC in a
fashion similar to leukocytes (31). These beads were added to
cytokine-activated HUVEC metabolically labeled with
[32P]orthophosphate. After 10 min of static adhesion, the
phosphorylation state of E-selectin was examined as described above. As
shown in Figure 3, incubation with PSGL-1
beads for a comparable period induced dephosphorylation of E-selectin
on activated HUVEC (43% of control). This result is similar in
magnitude to the dephosphorylation observed after both leukocyte
adhesion (see Fig. 1A) and mAb cross-linking (see
Fig. 2). As anticipated, the negative control IgG-coated beads failed
to induce dephosphorylation of E-selectin.
|
To investigate the functional significance of the cytoplasmic
domain in dephosphorylation of E-selectin, we generated two adenoviral
vectors: one (AdRSV(wt-E)) containing the wt E-selectin and the other
(AdRSV(Cyto-E)) a deletion mutant cDNA of E-selectin (Cyto)
lacking a cytoplasmic domain. These constructs were used to infect
cultured HUVEC, which were then used for experimentation 48 to 72
h postinfection. As shown in Figure 4A, both these adenovirus
constructs (AdRSV(wt-E) and AdRSV(Cyto-E)) resulted in the
expression of comparable levels of immunoreactive E-selectin protein in
infected HUVEC, as confirmed by fluorescence immunoassay, while a third
construct, AdRSVßgal, containing the lacZ gene (29), did
not lead to E-selectin expression. These cell surface levels of
E-selectin were similar in magnitude to those observed after standard
IL-1 stimulation of HUVEC. ICAM-1 expression was also analyzed to
monitor the activation state of the HUVEC during adenoviral
vector-mediated transfection. The expression level of ICAM-1 was
marginally, but significantly, increased by adenoviral infection
(66 ± 14 relative fluorescence units in cells expressing
E-selectin constructs vs 18 ± 3 relative fluorescence units in
AdRSVßgal-infected cells; p 
0.05; Fig. 4A), suggesting that this method of gene transduction
did not significantly activate cultured HUVEC. In addition, expression
of VCAM-1 and elaboration of IL-8 were not detected in these
E-selectin-transfected HUVEC (data not shown) (29).
The wt- and Cyto-E-selectin transduced HUVEC were metabolically
labeled with [32P]orthophosphate, and E-selectin was
immunoprecipitated from total cell lysates using anti-E-selectin
mAb H18/7. As shown in Figure 4B, wt-E-selectin, but not
Cyto-E-selectin, incorporated [32P]phosphate, thus
confirming that the cytoplasmic domain of E-selectin was the locus of
phosphorylation in cultured HUVEC, and that phosphorylation of
E-selectin could occur in the absence of cytokine activation. The
adhesion of HL-60 cells to wt-E-selectin-transduced HUVEC monolayers
induced a pattern of dephosphorylation similar to that observed in
cytokine-activated HUVEC monolayers (see Fig. 1A),
while adhesion of comparable numbers of HL-60 cells was without effect
on the Cyto-E-selectin monolayers. These data indicate that both
constitutive and regulated phosphorylation/dephosphorylation of
E-selectin take place in its cytoplasmic domain.
To better define the region of the E-selectin cytoplasmic domain
containing the phosphorylation site(s), we constructed a series of
truncation mutant cDNAs of E-selectin (Fig. 5A). These mutants were
cloned into a mammalian expression vector, pCDM8, and transfected into
COS-7 cells. Comparable surface expression levels of the transfected
mutants of E-selectin were again confirmed by fluorescence immunoassay.
These COS-7 cells transfectants were metabolically labeled with
[32P]orthophosphate, and E-selectin was
immunoprecipitated using H18/7. As shown in Figure 5B,
constitutive phosphorylation was observed with wt E-selectin and the
E1953 mutant. However, the mutant E-selectins E1925, E1899, and Cyto
failed to show any phosphorylation, suggesting that the cytoplasmic
region from Gln595 to Gln604 (containing three
serines and one tyrosine) contains the E-selectin phosphorylation
site(s).
|
The adenoviral construct, AdRSV(wt-E) was used to infect quiescent
HUVEC, resulting in a high level of expression of wt E-selectin. These
cells were then radiolabeled and subjected to 10 min of HL-60 adhesion.
The E-selectin from these HUVEC was immunoprecipitated using the
specific mAb H18/7, and subsequent autoradiography revealed
adhesion-dependent dephosphorylation analogous to that demonstrated
with endogenous E-selectin. Phosphoamino acid analysis of this
immunoprecipitated material revealed that only phosphoserine was
present in E-selectin, both preadhesion (Fig. 6A) and postadhesion
(Fig. 6B), even though the cytoplasmic domain of the
protein contains two tyrosines that might have served as potential
phosphorylation sites.
|
The effect of phosphatase inhibitors also was examined to
determine whether the observed leukocyte adhesion-induced
dephosphorylation of E-selectin was enzymatically regulated.
[32P]orthophosphate-labeled, IL-1ß-activated HUVEC
were incubated for 10 min before HL-60 adhesion with either 100 µM
sodium vanadate, a broad spectrum inhibitor of protein tyrosine
phosphatases, or 100 nM okadaic acid, a specific inhibitor of the
serine/threonine phosphatases PP1 and PP2A. Both phosphatase inhibitors
completely blocked the leukocyte adhesion-induced dephosphorylation of
E-selectin in activated HUVEC (shown in Fig. 7). This strongly suggests that the
observed dephosphorylation of E-selectin was dependent upon the action
of an intracellular phosphatase(s).
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
-actinin in leukocytes (34). The cytoplasmic domain
of P-selectin (consisting of 35 amino acids, including seven potential
phosphorylation sites: two serines, two threonines, one tyrosine, and
two histidines) (8) appears to be important in trafficking to storage
granules (-granules in platelets, Weibel-Palade bodies in
endothelial cells) (35, 36) or to the lysosomal compartment and,
ultimately, in degradation (37). During cell activation,
phosphorylation of the cytoplasmic domain of P-selectin is rapidly
increased on serine, threonine (38), and tyrosine residues (39).
Recently, a novel histidine phosphorylation pathway was identified for
P-selectin (40), although the significance of this is as yet
unclear.
We recently have demonstrated that the cytoplasmic domain of E-selectin
(consisting of 32 amino acids, including six serines and two tyrosines)
is important for biomechanical linkage to the cytoskeleton during
adhesive interactions mediated by its extracellular (in particular,
lectin) domain. Certain actin-associated proteins, including
-actinin, filamin, vinculin, FAK, and paxillin, copurified with
E-selectin when the latter was selectively immunoextracted using
Ab-coated magnetic beads (20). A mutant of E-selectin (Cyto) lacking
a cytoplasmic tail failed to support these associations. Thus, we
concluded that E-selectin interacts with the cytoskeleton directly via
its cytoplasmic domain, rather than through an undefined auxiliary
protein (or analogous lipid or carbohydrate partner). The association
of FAK and paxillin with the cytoplasmic tail of E-selectin also raised
the question of an outside-in signaling function for E-selectin during
leukocyte-endothelial adhesion.
In this study, we have demonstrated that E-selectin is constitutively
phosphorylated in its cytoplasmic domain both when induced by cytokine
activation in cultured HUVEC and when expressed as a recombinant
protein by adenovirally mediated transduction in unactivated HUVEC.
Furthermore, we have shown that this phosphorylation occurs exclusively
on the amino acid serine (Fig. 6). The constitutive phosphorylation of
E-selectin is rapidly and significantly decreased in response to
stimuli that interact with the extracellular domain of this molecule.
These stimuli, which may result in E-selectin clustering in the plane
of the cell membrane, include 1) leukocyte adhesion, 2) mAb-mediated
cross-linking of cell surface E-selectin, and 3) the binding of beads
coated with an E-selectin glycoprotein ligand, namely PSGL-1. In
conjunction with our previously reported data regarding the association
of E-selectin with cytoskeletal proteins (20), these observations
provide the first experimental evidence for an outside-in signaling
function for E-selectin. Such a function may have important
implications for the orchestration of endothelial cell responses during
leukocyte-endothelial interactions involving E-selectin-dependent
adhesion.
Smeets et al. have previously reported that in cultured HUVEC,
TNF--induced E-selectin was constitutively phosphorylated (41). To
investigate the effects of leukocyte adhesion mediated via E-selectin
on its phosphorylation state, we conducted experiments using HL-60, a
promyelocytic leukocyte cell line that expresses E-selectin ligands
(30). As shown in Figure 1A, adhesion of HL-60 induced a
time-dependent dephosphorylation of E-selectin in cytokine-activated
HUVEC. This is the first demonstration that leukocyte adhesion can
modulate the phosphorylation of E-selectin expressed on endothelial
cells. Dephosphorylation of E-selectin was again observed when cell
surface E-selectin was cross-linked by binding of a specific mAb
(H18/7) and polyclonal anti-mouse IgG (Fig. 2). We also
demonstrated that dephosphorylation of E-selectin can be induced by the
binding of PSGL-1-coated microspheres (Fig. 3), suggesting that a
physiologically relevant ligand could also trigger this process. In
parallel, we observed that a biotinylated cohort of cell surface
E-selectin did not change during the time course of either adhesion or
cross-linking, nor was the total cellular E-selectin content altered by
these stimuli. The observed decrease in the phosphate content of
E-selectin therefore presumably reflects a biochemical modification of
this protein in response to ligation and/or cross-linking, as opposed
to an increased rate of protein turnover and/or cell surface shedding.
Additionally, we found it interesting that the dephosphorylation of
E-selectin induced by our three distinct experimental stimuli reached a
maximum of approximately 50 to 60% of control values (Figs. 1, 2, 3, and 4B). There are at least two potential
explanations for this. First, certain of the phosphorylated amino acid
residues of E-selectin may be regulated by dephosphorylation, while
others may not. Second, while cell surface E-selectin is subject to
adhesion/cross-linking and subsequent dephosphorylation, the
intracellular cohort of E-selectin molecules (i.e., nascent molecules,
etc.) is not directly exposed to these input stimuli. Either of these
mechanisms alone or in conjunction could account for the results that
we observed.
Using adenoviral vectors to transduce E-selectin into cultured HUVEC,
we determined that phosphorylation of E-selectin is occurring within
the cytoplasmic domain of the protein (Fig. 4B). As
shown in Figure 4A, adenoviral gene transduction per se did
not significantly activate endothelial cells, as judged by expression
of the activation-responsive adhesion molecule ICAM-1. Previously we
have also shown that adenoviral gene transduction in HUVEC does not
up-regulate expression of either E-selectin or VCAM-1 (29). HUVEC
transduced with wild-type E-selectin exhibited a pattern of
phosphorylation in the absence of cytokine activation similar to that
observed in cytokine-activated HUVEC. Conversely, HUVEC transduced with
an E-selectin mutant lacking a cytoplasmic domain (Cyto) failed to
show any phosphorylation after metabolic labeling with
[32P]orthophosphate (Fig. 4B). Given
that the Cyto mutant was expressed at levels comparable to those of
the full-length E-selectin, we conclude that the cytoplasmic domain of
E-selectin is not required for targeting of the protein to the cell
surface.
Adhesion of HL-60 induced dephosphorylation of the adenovirally
transduced wt E-selectin in a time-dependent fashion (Fig. 4B). The truncated E-selectin mutant, Cyto, which
was not phosphorylated, did not appear to be affected by HL-60
adhesion. As we have previously reported, both wt and
Cyto-E-selectin-transfected HUVEC support HL-60 adhesion to the same
extent (20). Since the HUVEC were not activated by adenovirus infection
and yet dephosphorylation of the wt E-selectin did occur, it is likely
that this adhesion-dependent dephosphorylation does not require the
milieu of a cytokine-activated endothelial cell. Thus, the phosphatase
responsible for dephosphorylation of the serine(s) of E-selectin
appears to be constitutively expressed in endothelial cells.
Based upon our results using adenoviral transduction as a method of
gene transfer, it is clear that this is a particularly powerful
technique for analysis of E-selectin function. The infection of HUVEC
is highly efficient, resulting in up to 90% of cells expressing
E-selectin or the E-selectin mutants (data not shown). This is a stark
contrast to the difficulties encountered when using more traditional
methods of gene transfection (e.g., calcium phosphate coprecipitation,
lipofectin, or electroporation) with HUVEC. Additionally, infection
with these adenovirus constructs does not appear to significantly
perturb the basal state of cultured HUVEC, as judged by the lack of
induction of the endogenous E-selectin gene (a sensitive indicator of
activation in HUVEC) as well as other inducible genes, including VCAM-1
and IL-8 (29) (Fig. 4A) (J.-M. Kiely, unpublished
observations), thus allowing us to dissociate activation-dependent and
activation-independent phenomena.
Using transient transfection in COS-7 cells, we also tested a series of
truncation mutants to establish the importance of a particular region
as a potential phosphorylation site in the E-selectin cytoplasmic
domain. As shown in Figure 5B, only the E1953 mutant was
phosphorylated, following [32P]orthophosphate labeling,
to an extent similar to full-length E-selectin. The E1953 mutant
contains six potential phosphorylation sites: five serines and one
tyrosine (Fig. 5A). None of the remaining truncation
mutants, including E1925 with its two serines, incorporated any
appreciable amount of [32P]orthophosphate. Thus, a
site(s) of E-selectin phosphorylation must lie between
Gln595 and Gln604. Candidate residues in this
region are Ser596, Ser599, and
Ser602. Based upon our existing data we cannot determine
what role, if any, Ser607 plays in E-selectin
phosphorylation. Currently, we are exploring the relative importance of
the phosphorylation state of these candidate residues for E-selectin
association with the endothelial cytoskeleton and other leukocyte
adhesion-mediated processes.
The modulation of protein phosphorylation states plays a central role
in the regulation of numerous cellular processes, including cell
growth, differentiation, cell cycle progression, and cytoskeletal
reorganization (42). Protein kinases and protein phosphatases are the
enzymes that mediate this process of phosphorylation/dephosphorylation.
The constitutive serine phosphorylation of E-selectin (Fig. 6) and its
ligation-induced dephosphorylation led us to investigate the potential
effect of protein serine/threonine phosphatases. As shown in Figure 7, the phosphatase inhibitors okadaic acid and sodium vanadate both
prevented the dephosphorylation of E-selectin. Okadaic acid is a
specific inhibitor of both protein phosphatase-1 (PP1) and protein
phosphatase-2A (PP2A), although the concentration of okadaic acid that
we used (100 nM) was in the effective range for inhibition of PP1
(IC50 of 10–15 nM) (43). On the other hand, sodium
vanadate is a broad spectrum inhibitor of protein tyrosine
phosphatases. Little information is available regarding the action of
sodium vanadate on PP1 or PP2A, although PP2A is known to be activated
by tyrosine dephosphorylation (44). The results of sodium vanadate
inhibition of dephosphorylation would thus favor PP1 as a candidate
instead of PP2A. Another possibility is that the dephosphorylation of
E-selectin is accomplished via the action of a dual specificity protein
tyrosine phosphatase, such as either MAP kinase phosphatase-1 (MKP-1)
or the vaccinia H1-related human dual specificity phosphatase (VHR)
(45, 46). These enzymes are capable of hydrolyzing not only the
phosphate groups of tyrosines but also those of serines and threonines.
This might, then, account for the effectiveness of both okadaic acid
and sodium vanadate as inhibitors of E-selectin dephosphorylation. Our
results using phosphatase inhibitors thus favor a phosphatase with a
specificity similar to that of PP1. Its activity may be secondarily
regulated by tyrosine phosphorylation, or it may be a dual specificity
enzyme. Currently, efforts are in progress to determine the identity of
the phosphatase that is responsible for dephosphorylation of E-selectin
and to define the events required for its activation.
The signaling functions of adhesion molecules have drawn a great deal of attention in recent years. Of particular interest have been the integrins, the cadherins, and the Ig superfamily of adhesion proteins. Integrin receptors have been extensively characterized with respect to their outside-in and inside-out signaling functions in various cell types (including endothelial cells) as a response to binding to counter-receptors on interacting cells or to extracellular matrix ligands (11, 12). Signaling via the integrins can activate kinase cascades, induce cytoskeletal reorganization, and modulate integrin affinities and adhesive functions (11, 12). The cadherins, expressed at cell borders and cell-cell junctions, regulate both junctional permeability and cellular transmigration. The cytoplasmic domain of VE-cadherin, for example, has been shown to be an active participant in signal transduction via its interaction with ß-catenin, plakoglobin, and p120 (13). PECAM-1, a member of the Ig superfamily of adhesion molecules, has been recently reported to play a role in transmitting signals via modulation of the phosphorylation state of its cytoplasmic domain. In fact, PECAM-1 is a substrate of the tyrosine kinase, Src, and when tyrosine phosphorylated can bind the protein tyrosine phosphatase, SH2-containing phosphatase-2 (SHP-2) (15, 16).
It is becoming apparent that cell surface E-selectin is capable of transducing a transmembrane signal in the endothelial cell. Our previous data indicated that E-selectin can establish a direct molecular linkage to the cytoskeleton (20). Among the cytoskeletal proteins found to interact with E-selectin were the tyrosine kinase, FAK, and its substrate paxillin, which themselves have been shown to participate in signaling cascades involving Cas, Grb2, Crk, Ras, and MAP kinase (47). The present study extends this concept of a signaling function for E-selectin by demonstrating constitutive serine phosphorylation of the E-selectin cytoplasmic tail. Enzymatic dephosphorylation, as demonstrated by our phosphatase inhibitor data, provides a potential explanation for the regulation of E-selectin’s association with cytoskeletal components. In platelets, for example, PP1 and PP2A are differentially associated with the cytoskeleton depending upon whether the platelets have been activated by thrombin (48). As a second example, inhibition of PP1 in unactivated HUVEC results in marked changes in intracellular architecture, including perinuclear condensation of microtubules and F-actin (49). The extent to which E-selectin phosphorylation/dephosphorylation impacts upon the behavior of endothelial cell cytoskeletal components is the subject of ongoing investigation.
Our current observations suggest that biochemical and structural changes induced in and by E-selectin may propagate signals important for co-ordinating endothelial responses in the leukocyte-endothelial adhesion cascade. For example, initial leukocyte adhesion-induced E-selectin clustering in the membrane (reflecting localized association with the cortical cytoskeleton) could dynamically regulate the avidity of subsequent leukocyte adhesion events. Alternatively, ligation of E-selectin at the apical endothelial surface could trigger events that might influence lateral cell-cell junctional permeability, perhaps by altering the phosphorylation state of protein components of adherens junctions (i.e., VE-cadherin, ß-catenin, plakoglobin, and p120) or of tight junctions (i.e., occludin, cingulin, Zonula occludens-1 (ZO-1), and Zonula occludens-2 (ZO-2). Finally, an E-selectin-transduced signal might lead ultimately to gene activation, perhaps through MAP kinase signaling intermediates or via downstream interactions in the FAK-paxillin cascade. These putative E-selectin-mediated functions would have important implications for the role of this vascular selectin in the integrated pathophysiologic response of endothelial cells at sites of inflammation.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 These two authors contributed equally to this work.
3 Current address: Department of Molecular Genetics, Medical Research Institute, Tokyo Medical and Dental University, 1–5–45 Yushima, Bunkyo-ku, Tokyo, Japan.
4 Address correspondence and reprint requests to Dr. Michael A. Gimbrone, Vascular Research Division, Department of Pathology, Brigham and Women’s Hospital, 221 Longwood Ave., LMRC-401, Boston, MA 02115. E-mail address:
5 Abbreviations used in this paper: PECAM-1, platelet/endothelial cell adhesion molecule-1; MAP, mitogen-activated protein; FAK, focal adhesion kinase; PSGL-1, P-selectin glycoprotein ligand-1; DPBS, Dulbecco’s phosphate-buffered saline; RSV, Rous sarcoma virus; wt, wild type; PP1, protein phosphatase-1.
Received for publication November 12, 1997. Accepted for publication March 12, 1998.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
6ß1 integrin. J. Leukocyte Biol. 61:397.[Abstract]
-actinin: receptor positioning in microvilli does not require interaction with -actinin. J. Cell Biol. 129:1155.This article has been cited by other articles:
|
|
 |
|
  H. Setiadi and R. P. McEver Clustering endothelial E-selectin in clathrin-coated pits and lipid rafts enhances leukocyte adhesion under flow Blood, February 15, 2008; 111(4): 1989 - 1998. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Osaka, S. Hagita, M. Haraguchi, M. Kajimura, M. Suematsu, and M. Yoshida Real-time imaging of mechanically injured femoral artery in mice reveals a biphasic pattern of leukocyte accumulation Am J Physiol Heart Circ Physiol, April 1, 2007; 292(4): H1876 - H1882. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. Nishiwaki, M. Yoshida, H. Iwaguro, H. Masuda, N. Nitta, T. Asahara, and M. Isobe Endothelial E-Selectin Potentiates Neovascularization via Endothelial Progenitor Cell-Dependent and -Independent Mechanisms Arterioscler Thromb Vasc Biol, March 1, 2007; 27(3): 512 - 518. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. Galkina and K. Ley Leukocyte Recruitment and Vascular Injury in Diabetic Nephropathy J. Am. Soc. Nephrol., February 1, 2006; 17(2): 368 - 377. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. Yu, K. S. Moulton, M. K. Khan, S. Vineberg, E. Boye, V. M. Davis, P. E. O'Donnell, J. Bischoff, and D. S. Milstone E-selectin is required for the antiangiogenic activity of endostatin PNAS, May 25, 2004; 101(21): 8005 - 8010. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J.-M. Kiely, Y. Hu, G. Garcia-Cardena, and M. A. Gimbrone Jr. Lipid Raft Localization of Cell Surface E-Selectin Is Required for Ligation-Induced Activation of Phospholipase C{gamma} J. Immunol., September 15, 2003; 171(6): 3216 - 3224. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Yoshida, Y. Takano, T. Sasaoka, T. Izumi, and A. Kimura E-Selectin Polymorphism Associated With Myocardial Infarction Causes Enhanced Leukocyte-Endothelial Interactions Under Flow Conditions Arterioscler Thromb Vasc Biol, May 1, 2003; 23(5): 783 - 788. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. S. Kluger, S. L. Shiao, A. L. M. Bothwell, and J. S. Pober Cutting Edge: Internalization of Transduced E-Selectin by Cultured Human Endothelial Cells: Comparison of Dermal Microvascular and Umbilical Vein Cells and Identification of a Phosphoserine-Type Di-leucine Motif J. Immunol., March 1, 2002; 168(5): 2091 - 2095. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. Hu, B. Szente, J.-M. Kiely, and M. A. Gimbrone Jr. Molecular Events in Transmembrane Signaling via E-selectin. SHP2 ASSOCIATION, ADAPTOR PROTEIN COMPLEX FORMATION AND ERK1/2 ACTIVATION J. Biol. Chem., December 14, 2001; 276(51): 48549 - 48553. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Yoshida, T. Sawada, H. Ishii, R. E. Gerszten, A. Rosenzweig, M. A. Gimbrone Jr, Y. Yasukochi, and F. Numano HMG-CoA Reductase Inhibitor Modulates Monocyte-Endothelial Cell Interaction Under Physiological Flow Conditions In Vitro : Involvement of Rho GTPase-Dependent Mechanism Arterioscler Thromb Vasc Biol, July 1, 2001; 21(7): 1165 - 1171. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. E. ERIKSSON, X. XIE, J. WERR, P. THOREN, and L. LINDBOM Direct viewing of atherosclerosis in vivo: plaque invasion by leukocytes is initiated by the endothelial selectins FASEB J, May 1, 2001; 15(7): 1149 - 1157. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Ishii, M. Yoshida, A. Rosenzweig, M. A. Gimbrone Jr, Y. Yasukochi, and F. Numano Adenoviral transduction of human E-selectin into isolated, perfused, rat aortic segments: an ex vivo model for studying leukocyte-endothelial interactions J. Leukoc. Biol., November 1, 2000; 68(5): 687 - 692. [Abstract] [Full Text]
|
|
|
|
|
|
Y. Hu, J.-M. Kiely, B. E. Szente, A. Rosenzweig, and M. A. Gimbrone Jr. E-Selectin-Dependent Signaling Via the Mitogen-Activated Protein Kinase Pathway in Vascular Endothelial Cells J. Immunol., August 15, 2000; 165(4): 2142 - 2148. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. R. Allport, W. A. Muller, and F. W. Luscinskas Monocytes Induce Reversible Focal Changes in Vascular Endothelial Cadherin Complex during Transendothelial Migration under Flow J. Cell Biol., January 10, 2000; 148(1): 203 - 216. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Cyto-E). HUVEC
plated in a 96-well microtiter plate were infected with either
AdRSV(wt-E) or AdRSV(
