 |
INTRODUCTION |
The three known selectins, L-, E-, and
P-selectin are cell adhesion molecules that initiate interactions
between leukocytes and endothelial cells during leukocyte extravasation
(1). The minimal ligand structure for all three selectins is the
tetrasaccharide sialyl Lewis X
(sLex).1 The
biosynthesis of sLex requires the sequential action of a
number of glycosyltransferases of which the final reaction is mediated
by 
(1, 3)-fucosyltransferases (2, 3). Of the six known human (1,
3)-fucosyltransferases (Fuc-TIII, Fuc-TIV, Fuc-TV, Fuc-TVI, Fuc-TVII,
and Fuc-TIX), Fuc-TIV and Fuc-TVII have been implicated in the
generation of selectin ligands (4). Fuc-TVII-deficient mice exhibit
severe defects in the transmigration of neutrophils into inflamed
peritoneum, demonstrating the central importance of this enzyme for the
generation of E- and P-selectin ligands on neutrophils (5). In
addition, the lack of Fuc-TVII severely reduced the expression of
L-selectin ligands on high endothelial venules in lymph
nodes, leading to strong inhibition of lymphocyte homing (5). The
ability of Th1 cells to bind to E- and P-selectin is induced during the
course of differentiation of these cells (6), and this correlates with
the induction of Fuc-TVII (7).
Human Fuc-TIV, also termed ELFT for ELAM-1
ligand fucosyltransferase, was
reported to generate E-selectin binding carbohydrate modifications (8).
However, this activity has been observed only when the enzyme is
expressed at a certain level and in cells with a certain glycosylation
phenotype (9-13). Transfection of Chinese hamster ovary (CHO) cells of
the DHFR
strain with human Fuc-TIV led to the expression
of CSLEX-1-reactive sLex ,although no sLex
could be generated by this enzyme in CHO-Pro5 cells (10).
In BHK-21 cells, human Fuc-TIV generates Lex seven times
more efficiently than sLex (14). In COS cells, human as
well as mouse Fuc-TIV could not generate sLex-epitopes on
the cell surface as defined by the mAb CSLEX-1 (15). In contrast, in
in vitro enzyme assays mouse Fuc-TIV transfected in COS
cells, but not human Fuc-TIV, could efficiently accept 3'-sialyl-N-acetyllactosamine to form sLex (15).
The biochemical basis for the selective generation of sLex-like structures by Fuc-TIV is, at present,
unexplained. Yet, Fuc-TIV does participate in the generation of
physiologically relevant selectin ligands, although it clearly plays a
minor role compared with Fuc-TVII. Although the fraction of rolling
leukocytes in non-inflamed venules of the skin was normal in
Fuc-TIV/ animals, an increase of the rolling velocity
was observed (16). Furthermore, leukocytes rolling at velocities below
10 µm/sec were absent in Fuc-TIV/ mice, and Fuc-TIV
deficiency in the context of Fuc-TVII deletion extinguished residual
selectin ligand activities observed in Fuc-TVII/ mice
(17). Thus, it is important to further analyze the relative contribution of Fuc-TVII and Fuc-TIV to the generation of
sLex-modified glycoconjugates that might serve as selectin ligands.
We have recently analyzed the selective contribution of Fuc-TVII and
Fuc-TIV to the generation of the glycoprotein ligands E-selectin
ligand-1 (ESL-1) (18, 19) and P-selectin glycoprotein ligand-1 (PSGL-1)
(20-23). Each of these two ligands requires sLex for
binding, with the difference that ESL-1 requires sLex on
N-glycans (18), whereas PSGL-1 requires it on O-glycans that
carry a core-2 branch (21, 24). In addition, PSGL-1 requires sulfation
of the tyrosine residues within its N terminus for binding to
P-selectin (25-29). The physiological relevance of PSGL-1 in leukocyte
extravasation is well established (23). Antibodies against mouse PSGL-1
inhibit neutrophil recruitment into inflamed peritoneum (30), and the
migration of Th1 cells into inflamed skin (6). Similar results were
obtained with mice deficient for the PSGL-1 gene (31, 32).
After analyzing mouse neutrophils from Fuc-TIV and Fuc-TVII deficient
mice, we have shown recently that Fuc-TVII exclusively directs the
expression of PSGL-1 glycoforms that bind with high affinity to
P-selectin (33). In contrast, Fuc-TIV preferentially directs expression
of ESL-1 glycoforms that exhibit high affinity for E-selectin. We could
mimic this selectivity in transfected CHO-Pro5 cells that
expressed PSGL-1, ESL-1, and core-2
-1,6-N-acetylglucosaminyl-transferase (C2GnT) (33). The
molecular mechanisms in the Golgi that are the basis for this in
vivo selectivity are unknown.
In addition to glycoprotein counter-receptors, glycolipids have also
been described as carriers of sLex and as selectin ligands
(34-36). However, selectin-mediated cell binding to glycolipid
counter-receptors has always been tested with immobilized glycolipids.
Hence, the relevance of glycolipids as selectin ligands in the
physiological context of a cell surface is still unclear. Here, we have
analyzed whether Fuc-TVII and Fuc-TIV would differ in their ability to
generate sLex-carrying glycolipids or glycoproteins in
transfected CHO-Pro5 cells. Surprisingly, we found that
Fuc-TIV preferentially generates sLex on glycolipids,
whereas Fuc-TVII preferentially decorates sLex on
glycoproteins. A comparison of the E-selectin binding efficiency of
these transfectants suggests the possibility that this unexpected acceptor specificity could be a reason for the lower efficiency with
which Fuc-TIV generates E-selectin ligands as compared with Fuc-TVII.
Furthermore, our results demonstrate that the Golgi environment of
CHO-Pro5 cells provides conditions under which each
enzyme is able to generate sLex-epitopes. However, the
enzymes preferentially synthesize these epitopes on different classes
of carrier moieties in vivo.
| |
EXPERIMENTAL PROCEDURES |
Cell Lines and Cell Culture--
CHO-Pro5 cells
(11) were obtained from Dr. A. Hasilik (University of Marburg) and
grown in -minimal essential medium (MEM) (Invitrogen)
containing 10% fetal bovine serum, 100 µg/ml
L-glutamine, and 100 units/ml penicillin/streptomycin at
37 °C in a humidified atmosphere of 10% CO2.
Generation of Stable Transfectants--
Stable transfection was
performed as described previously (37) with slight modifications.
Briefly, 1 × 107 CHO-Pro5 cells
harvested in PBS (containing 5 mM EDTA) at 90% confluency were electroporated in 0.6 ml of PBS either with 10 µg of pcDNA3 vector containing mouse Fuc-TIV cDNA (38) or 20 µg of pcDNA1 vector containing human Fuc-TIV cDNA (37) (both kindly provided by
Dr. John Lowe, University of Michigan, Ann Arbor, MI) in a 0.4 cm
cuvette at 950 µF and 0.25 kV. The human Fuc-TIV plasmid was
co-transfected with 5 µg of pAH58 vector (39) for hygromycin B
resistance. Cells transfected with mouse Fuc-TIV were selected with 800 µg/ml G418; cells transfected with human Fuc-TIV were selected with
300 µg/ml hygromycin. Mock transfected cells were generated with the
same vectors lacking the Fuc-T cDNA inserts. CHO-Pro5
cells co-transfected with mouse PSGL-1 and human C2GnT were called PC
and have been described (33). These cells, further transfected in a
second round either with mouse Fuc-TIV in pcDNA3 or mouse Fuc-TVII
in pcDNA3 and co-transfected with the pAH58 vector for hygromycin B
resistance, were also described previously and are named PC4 and PC7,
respectively (33).
A hybrid cDNA encoding the cytoplasmic, transmembrane domain and
stem region of mouse Fuc-TIV fused to the catalytic domain of mouse
Fuc-TVII was constructed according to a strategy taken from Ref. 40 as
follows. In a first PCR, a 303-bp fragment was generated using sense
primer 5'-TGG AAT TCT GCA GAT CA-3', antisense primer 5'-CCA GAT AAG
GAT GGT GAG CAG GCG TTG CGG AGC TGG-3', and the pcDNA3 vector
containing mouse Fuc-TIV cDNA as a template (PCR conditions were 3 min at 94 °C, 25 cycles with 15s at 94 °C, 20s at 55 °C, 30s
at 72 °C, and 10 min at 72 °C). The purified PCR product was then
used as a 5'-megaprimer in a second PCR together with antisense primer
5'-GTC AAG CCT GGA ACC AGC TT-3' and the pcDNA3 vector containing
mouse Fuc-TVII cDNA as a template (PCR conditions were 3 min at
94 °C, 30 cycles with 20s at 94 °C, 20s at 50 °C, 30s at
72 °C, and 10 min at 72 °C). The PCR product was cloned into a
pcDNA3 vector using the TOPO-TA system (Invitrogen). PC cells were
transfected with this construct (plus co-transfection with pAH58 vector
for hygromycin B resistance) and were referred to as PC 4/7 chimera cells.
Following transfection, CHO cells were seeded into 90-mm culture
dishes. After 6-10 days, individual clones were surrounded by glass
rings, sealed with sterile grease, and released with trypsin/EDTA
solution (Invitrogen). All cell lines were subcloned twice in
microtiter plates by seeding the cells at a statistical density of 0.5 cells/well followed by analysis for the expression of transfected
cDNAs using PSGL-1- and sLex-specific antibodies in
flow cytometry.
Antibodies and Selectin-IgG Chimeras--
The following
antibodies were used. HECA-452 (anti-sLex, rat IgM) (41)
and CSLEX-1 (anti-sLex, mIgM) (42) were purchased from the
ATCC. Anti-CD65s (a variant of sLex;
NeuAc2,3Gal1,4GlcNAc1,3Gal1,4(Fuc1,3)-GlcNAc (clone
VIM-2, mIgM)) (43) was purchased from Bio Research (Kaumberg, Austria); 2F3 (anti-sLex, mIgM) (44) and anti-CD15 (Lex)
mAb clone HI98 (mIgM) were purchased from BD Pharmingen.
FITC-conjugated rabbit anti-mouse IgG, FITC-conjugated rabbit anti-rat
IgG and IgM, FITC-conjugated rabbit anti-mouse IgM, DTAF-conjugated
goat anti-human IgG, TRITC-conjugated goat anti-mouse IgG and IgM, peroxidase-conjugated anti-rabbit IgG and peroxidase-conjugated anti-rat IgG and IgM were all purchased from Dianova (Hamburg, Germany). Polyclonal antibodies against ESL-1 (Affi-60) and polyclonal antibodies against mPSGL-1 (Affi-124) were generated as described previously (6, 19). Anti-mPSGL-1 mAb 4RA10 (rat IgG) was described
previously (45). The GM3-specific antibody GMR6 (mIgM) was purchased
from Seikagaku (Tokyo, Japan). Isotype-matched negative control
antibodies were R4-22 (rat IgM, BD Pharmingen), C48.6 (mIgM, BD
Pharmingen), and anti-mP-selectin antibody RB40.34 (rat IgG (46)).
E-selectin-IgG and vascular endothelial cadherin-IgG chimeras were
produced as described previously (47, 48). HECA-452 and CSLEX-1 were
prepared from supernatants of hybridomas (obtained from the ATCC)
cultured in Dulbecco's modified Eagle's medium (Invitrogen)
containing 10% fetal bovine serum.
Flow Cytometry Analysis--
Flow cytometry was essentially
performed as described (37, 38). Briefly, 5 × 105
cells were incubated with 10-20 µg/ml mAb in flow cytometry buffer (Hanks' buffer, 3% FCS, 0.04% azide) at 4 °C or 37 °C for 20 min. For detection of selectin-IgG binding, 20 µg/ml E-selectin-IgG construct in flow cytometry buffer was used. Cells were washed twice
with the same buffer and stained with FITC- or DTAF-conjugated secondary monoclonal antibodies at 1:100 dilutions. After 20 min of
incubation at 4 °C, cells were washed twice, counter stained with
propidium iodide, and analyzed by flow cytometry (FACScalibur, BD Life
Sciences). Data were collected by gating for propidium iodide-negative
cells and analyzed using the CellQuest program.
Fucosyltransferase Assays--
2 × 108 cells
were lysed in 1 ml of lysis buffer (1% Triton X-100, 20 mM
Tris-HCl, pH 8.4, 160 mM NaCl, 1 mM
CaCl2, 1 mM benzamidine, 1 mM
phenylmethylsulfonyl fluoride, and 0.1 units/ml
2-macroglobulin), and cell debris was pelleted at
10,000 × g for 10 min. The protein concentration was
determined by the BCA assay (Pierce). Fucosyltransferase activity
assays were carried out in a total volume of 40 µl containing 10 µg
of protein extract, 25 mM cacodylate (pH 6.2), 0.25%
Triton X-100, 10 mM MnCl2, 5 mM
GDP-fucose, 0.07 µCi of GDP-[3H]fucose, and 10 mM N-acetyllactosamine (NAL) as acceptor oligosaccharide (purchased from Dextra Laboratories). Blanks were prepared by omitting
the acceptor in the reaction mixture. After incubation at 37 °C for
2 h, 1 ml of a Dowex 1-X8 slurry (1:4 (w/v) in water) was added to
the reaction and vortexed. 500 µl of the supernatant was counted in 5 ml of scintillant (Ultima Gold XR, Packard). To obtain values solely
due to the fucosylation of the acceptor substrate, total counts of the
control (without acceptor substrate) were subtracted from total counts
of samples with the acceptor. The specific activity of the
fucosyltransferase was calculated as pmol/min/mg.
Semiquantitative RT-PCR--
Total cellular RNA was isolated
from 107 cells using the Trizol reagent (Invitrogen)
according to the manufacturer's protocol. 1 µl of total RNA was used
as template in a 20-µl reverse transcription reaction. To reflect the
initial mRNA expression levels, PCR amplification of cDNA was
carried out with cycle numbers that had been tested to be in the linear
range and well below the plateau phase of amplification. The following
number of cycles and primers were used. For -actin: 20 cycles,
sense-primer 5'-TGG GTC AGA AGG ACT CCT ATG-3', antisense-primer 5'-CAG
GCA GCT CAT AGC TCT TCT-3', product of a 591-bp fragment. For mouse
Fuc-TIV: 25 cycles, sense-primer 5'-GAC GCT AAC TGG CAA AGC CCT-3',
antisense-primer 5'-GGT GAT GTA ATC CAC GTG CCG-3', product of a 451-bp
fragment. For human Fuc-TIV: 25 cycles, sense primer 5'-TGG ATC TGC GCG
TGT TGG ACT-3', antisense primer 5'-CGG TCA CAT GTT GGC TCA GTT-3',
product of a 360-bp fragment. For mouse Fuc-TVII: 25 cycles, sense
primer 5'-CCG TCT GAG TGC TAA CCG GAG-3', antisense primer 5'-CGC CAG AAC TTC TCA GTG ATG-3', product of a 501-bp fragment. PCR reactions were performed in a final volume of 50 µl with 2.5 units of
Taq polymerase (Amersham Biosciences) and 4 µl of
single-stranded cDNA from the RT reaction as template. The PCR
amplification was carried out using a Biometra thermal cycler with the
following program: 95 °C for 3 min followed by 25 or 20 cycles of
94 °C for 50 s, 56 °C for 50 s, and 72 °C for
50 s followed by 72 °C for 7 min. To compare the mRNA
expression levels in more detail, PCR was carried out as above using
2-fold serial dilutions of the input cDNA. PCR products were
separated by 1.5% agarose gel electrophoresis, transferred to
nitrocellulose membranes, and hybridized with 32P-labeled
hFuc-TIV, mFuc-TIV, mFuc-TVII, or -actin cDNA probe (106 cpm/ml). The blots were exposed to Hyperfilm (Amersham Biosciences).
Enzyme Treatment of Cells--
For each experiment, cells were
harvested in PBS containing 5 mM EDTA and washed once with
MEM. For proteinase K treatment, cells were resuspended in PBS at a
density of 1 × 107 cells/ml. 1 × 106 cells (in 100 µl) were aliquoted to each well of
96-well v-bottom plates, and 1 µl of proteinase K stock solution
(Roche Molecular Biochemicals; 10 mg/ml in 50 mM Tris-HCl,
pH 8.0, and 1 mM CaCl2) was added to each well.
After 20 min of incubation at 37 °C, 100 µl of PBS containing 6%
FCS and a protease inhibitor mixture containing EDTA (Roche Molecular
Biochemicals) was added to inhibit proteinase K activity. After 5 min
of incubation at room temperature, cells were washed twice with Hanks'
buffer containing 3% FCS and 0.04% azide. In all assays, proteinase K
treatment resulted in less than 15% cell death. For trypsin treatment,
cells were resuspended in MEM and plated in V-bottom plates
(Greiner) with 7.5 × 105 cells/well. Trypsin (Roche
Molecular Biochemicals) was added to give a final concentration of 1.7 mg/ml. Cells were incubated for 1 h at 37 °C and washed three
times in MEM. Trypsin treatment resulted in less than 10% cell death.
Confocal Immunofluorescence Microscopy--
Cells were plated on
chamber slides (Nunc Lab-Tec) 1 day before the experiment, washed once
with Hanks' buffer/0.04% azide, and fixed with 4% paraformeldehyde
at room temperature for 10 min. To remove glycolipids, fixed cells were
incubated with 100 µl of chloroform/methanol (1:1) at room
temperature for 10 min. For staining controls, the extraction step was
omitted. Cells were then washed twice with Hanks' buffer/0.04% azide
and incubated for 1 h at room temperature with 100 µl of primary
antibody in DMEM containing 10% FCS and 0.04% azide. After washing
three times with Hanks' buffer/0.04% azide, FITC-, DTAF- or
TRITC-conjugated secondary antibodies were added and incubated at room
temperature for 1 h. Cells were washed five times with Hanks
buffer/0.04% azide and mounted with one drop of Dako fluorescent
mounting medium. Immunofluorescence images were captured and analyzed
by confocal microscopy (Leica, Heidelberg, Germany). Five different
areas on the microscope slide were scanned individually by using ×20 objective and 5-milliwatt laser output. Using Leica software, the
intensity of immunofluorescence on the image was quantified and divided
by the number of cells counted to make quantifications comparable (mean
fluorescence intensity).
Affinity Purification and Western Blot--
Cells were lysed in
lysis buffer (1 × 107 cells/ml lysis buffer)
containing 1% Triton X-100, 20 mM Tris-HCl, pH 8.0, 160 mM NaCl, 1 mM CaCl2, 1 mM benzamidine, 1 mM phenylmethylsulfonyl fluoride, and 0.1 units/ml 2-macroglobulin, at 4 °C for 30 min. Insoluble material was removed by centrifugation at 14,000 rpm for 15 min. The lysate was incubated with protein A-Sepharose beads loaded
with polyclonal mAbs against ESL-1 (Affi-60) or PSGL-1 (Affi-124) at
4 °C overnight. Immunoprecipitated proteins were separated on 6%
SDS-PAGE and transferred to nitrocellulose membranes (Schleicher & Schüll). Membranes were incubated with blocking buffer
(Tris-buffered saline, 1% Tween 20, 4% nonfat milk) at room
temperature for 2 h and probed with the hybridoma supernatant of
mAb HECA-452. After washing with blocking buffer, the blot was probed
with horseradish peroxidase-conjugated anti-rat IgM in blocking buffer
at room temperature for 1 h. The blot was washed with
Tris-buffered saline/0.1% Tween 20 and then visualized by chemiluminescence using the ECL reagent (Amersham Biosciences).
Cell Adhesion Assay under Static Conditions--
Adhesion assays
were performed in 96-well flat bottom plates (Maxisorp, Nunc) coated
with E-selectin-IgG, VE-cadherin-IgG, or human IgG1 diluted
in Hanks' balanced salt solution (Hanks' buffer, Biochem, Berlin,
Germany) or coated with 10% FCS (49). Subsequently, plates were
blocked with 10% FCS in DMEM for 1 h at 37 °C. Cells were
added at a density of 4 × 105 cells in 200 µl of
Hanks' buffer for 20 min at 4 °C or 37 °C under mild rotation
(80 rpm). Wells were washed three times with the same buffer, and the
remaining cells were fixed with Hanks' buffer containing 2%
glutaraldehyde (Sigma) at 4 °C for 30 min and evaluated by
computer-aided image analysis with the NIH Image 1.55 software (30).
Each experiment was done in triplicate. Bound cells from four areas of
each field were counted. Unspecific binding was tested by allowing
cells to bind in the presence of 5 mM EDTA.
Cell Attachment Assay under Flow--
Transfected CHO Cells were
cultured to 80-100% confluence and harvested in PBS containing 5 mM EDTA. After incubating in this buffer for 10 min at
37 °C, cells were pelleted, resuspended in the same buffer, and
incubated for another 10 min at 37 °C to obtain a single cell
suspension. Cells were finally resuspended in DMEM containing 10% FCS
and 0.04% azide at a concentration of 1 × 106
cells/ml. E-selectin-IgG and, for control, human-IgG1 were
immobilized by incubating glass cover slips for 3.5 h at room
temperature in Hanks' buffer containing the proteins at concentrations
of 0.1 or 0.5 µg/ml. Blocking of unspecific binding sites was done at
room temperature with 5% BSA in Hanks' buffer overnight. Adhesion under flow was essentially analyzed as described (30). Cells were
diluted to a final density of 1 × 106 cells/ml with
DMEM containing 10% FCS and 0.04% azide and perfused through a
rectangular transparent laminar flow perfusion chamber over the
protein-coated coverslip. The flow rate was adjusted to 0.52 or 1.58 dyn/cm2. Cell rolling was recorded by video camera
immediately after cell perfusion. Evaluation was started 90 s
after starting the peristaltic pump. The number of rolling cells was
counted in 10 different fields, and four areas were counted from each
0.5 mm2 field. Unspecific binding to E-selectin-IgG was
analyzed in the presence of 5 mM EDTA.
| |
RESULTS |
Fuc-TIV and Fuc-TVII Direct Expression of sLex Epitopes
on the Cell Surface of Transfected CHO-Pro5
Cells--
We have shown previously for mouse neutrophils that
Fuc-TVII, but not Fuc-TIV, exclusively directs the expression of
P-selectin binding glycoforms of PSGL-1 while Fuc-TIV preferentially
directs the expression of E-selectin binding glycoforms of ESL-1 (33). This selectivity could be mimicked in CHO-Pro5 cells that
express endogenous ESL-1 and have been stably transfected with mouse
PSGL-1 and the human core-2 branching enzyme C2GnT combined with
ectopic expression of either mouse Fuc-TIV or mouse Fuc-TVII (33). We
have now used these transfectants (called PC4 and PC7) to further
analyze their capacity to generate sLex epitopes on their
cell surface. Using flow cytometry, we found that both cell lines
expressed similar levels of sLex-epitopes as defined by the
mAbs CSLEX-1, HECA-452, and 2F3 (Fig. 1).
The expression of sLex was not detected in the absence of
Fuc-Ts (Fig. 1) and was sensitive to neuraminidase-treatment of the
cells (data not shown). If mouse Fuc-TIV was transfected alone,
i.e. in the absence of PSGL-1 and C2GnT, expression levels
of all three antibody-defined sLex epitopes on the cells
were in the same range (Fig. 1). The mRNA expression levels of
mouse Fuc-TIV and Fuc-TVII were similar as determined by
semi-quantitaive RT-PCR (Fig.
2A), and similar results were
obtained with six PC4-like and two PC7-like clones independently transfected in the same way as PC4 and PC7 cells, respectively (not
shown). Thus, mouse Fuc-TIV can efficiently generate
sLex-related structures defined by three mAbs on the cell
surface of CHO-Pro5 cells.
View larger version (36K):
[in this window]
[in a new window]
|
Fig. 1.
Flow cytometry of Fuc-T-transfectants with
mAbs against sLex-related carbohydrate epitopes.
CHO-Pro5 transfectants were analyzed by flow cytometry
with the anti-sLex mAbs CSLEX-1, HECA-452, 2F3, and the
anti-CD65s mAb VIM-2 (as indicated). Cells were transfected with mouse
PSGL-1, human C2GnT and mouse Fuc-TIV (PC4), mouse PSGL-1, human C2GnT
and mouse Fuc-TVII (PC7), only with mouse Fuc-TIV
(Pro5/mFuc-TIV), or only with human Fuc-TIV
(Pro5/hFuc-TIV). The thin lines
show staining of cells without transfected Fuc-Ts.
|
|
View larger version (33K):
[in this window]
[in a new window]
|
Fig. 2.
Expression levels of transfected Fuc-Ts.
A, mRNA expression levels of human Fuc-TIV
(hFuc-TIV) mouse Fuc-TIV (mFuc-TIV) or mouse
Fuc-TVII (mFuc-TVII) in the indicated transfectants were
determined by semiquantitative RT-PCR. PCR reactions were carried out
under non-saturating conditions using 2-fold serial dilutions of input
cDNA or no cDNA ( RT), as described under
"Experimental Procedures." PCR products were electrophoresed,
transferred to nitrocellulose, and hybridized with
32P-labeled cDNA probes for the respective
fucosyltransferase and -actin. B,
(1,3)-fucosyltransferase enzyme assays were performed with detergent
extracts of equal numbers of either CHO-Pro5 cells
transfected with human Fuc-TIV
(Pro5/hFuc-TIV) or mouse Fuc-TIV,
mouse PSGL-1, and human C2GnT (PC4) or mouse Fuc-TIV
(Pro-5/mFuc-TIV). In each case, measurements of transfected
clones (open bars) were compared with negative control
measurements performed with pcDNA3 mock transfected
CHO-Pro5 cells (closed bars). Data shown
correspond to fucosyltransferase activities (expressed as pmol/min/mg)
measured with samples containing acceptor substrate minus activities
measured without acceptor substrate and represent the mean and S.D.
from four similar assays.
|
|
Because the question of whether human Fuc-TIV can indeed generate
sLex on the cell surface has been controversial, we
compared mouse and human Fuc-TIV-transfected CHO-Pro5
cells. In agreement with previous reports (10, 11), human Fuc-TIV only
generated very small amounts of the CSLEX-1 and 2F3 epitopes on
CHO-Pro5 cells (Fig. 1). However, the
sLex-related structure defined by mAb HECA-452 was well
expressed and at similar levels as that generated by mouse Fuc-TVII and mouse Fuc-TIV (Fig. 1). Thus, human Fuc-TIV cannot express all, but at
least one, mAb-defined sLex-like epitope. In contrast,
mouse Fuc-TIV can generate all three sLex-like epitopes
with a similar efficiency as that of Fuc-TVII. Human and mouse Fuc-TIV
were expressed at similar levels as determined by semi-quantitative
RT-PCR (Fig. 2A) and in vitro enzyme assays (Fig.
2B). In addition, (1,
3)-fucosylation-dependent CD65s defined by the mAb VIM-2
were generated at similar levels by both transfectants as shown by
fluorescence-activated cell sorter analysis (Fig. 1). Similar results
were obtained with eight mouse Fuc-TIV and six human Fuc-TIV
transfected clones (not shown).
Fuc-TVII, but Not Fuc-TIV, Preferentially Generates Cell Surface
sLex on Glycoproteins--
To test whether
sLex epitopes generated by Fuc-TIV or Fuc-TVII are
decorated on glycoproteins, PC4 and PC7 cells were treated with
proteinase K (100 µg/ml). As analyzed by flow cytometry, cell surface
sLex on Fuc-TVII transfectants was strongly reduced by 86%
after 20 min of proteinase K treatment at 37 °C (Fig.
3). Evaluation of three similar
experiments resulted in 87 ± 6% reduction (not shown). In
contrast, sLex on Fuc-TIV transfectants was not decreased
by proteinase K treatment (Fig. 3). Binding of the sLex
antibody to Fuc-TIV transfectants even slightly increased to 106%
(Fig. 3) after proteinase K treatment (114% with an S.D. of ± 17% obtained in four experiments, not shown). Efficient removal of the
glycoprotein PSGL-1 but not of the major glycolipid on CHO cells, GM3,
showed that proteinase K treatment was protein-specific (Fig. 3). The
increased binding of the GM-3 antibody in Fig. 3 was probably due to an
increased accessibility of glycolipids following the removal of
proteins from the cell surface. To rule out the possibility that
glycoproteins decorated with Fuc-TIV-generated sLex are
just more resistant to proteinase K than Fuc-TVII-modified glycoproteins, the cells were treated with up to 15 times higher concentrations of proteinase K at 37 °C for 2 h. Although these harsher protease conditions caused lysis of some cells, removal of
surface sLex on still intact cells was not stronger than
after milder protease digestion (data not shown). To confirm the
results obtained with proteinase K, PC4 and PC7 cells were also treated
with trypsin. The expression of sLex after trypsin
treatment was almost identical to that following proteinase K treatment
(data not shown). These results strongly suggest that Fuc-TVII
preferentially generates sLex on glycoproteins, whereas
Fuc-TIV-generated sLex is mainly found on
protease-resistant glycoconjugates. These could either be special
proteins of extreme resistance to proteases or proteins that are not
accessible for proteinase K and trypsin (although they would be
accessible for antibodies) or, and this is the most likely explanation,
they could be structures other than glycoproteins.
View larger version (41K):
[in this window]
[in a new window]
|
Fig. 3.
Protease sensitivity of sLex
epitopes on Fuc-T transfectants. The binding of
sLex-specific antibody CSLEX-1, anti-PSGL-1 antibody 4RA10,
and anti-GM3 antibody GMR6 to PC4 and PC7 cells was analyzed by flow
cytometry after cells had been treated with proteinase K
(PK(+), thick line) or without
proteinase K (PK(), shaded graph).
Staining with isotype-matched negative control antibodies is depicted
as a dotted line. The relative mean fluorescence intensity
is given as a percentage of the signal obtained with mock-treated
cells. One of three similar experiments is depicted.
|
|
Although this study shows that Fuc-TIV preferentially fucosylates
non-protein carriers for sLex, it is clear that at least a
few proteins are targets for this enzyme. We have previously shown that
in PC4 cells, Fuc-TIV generates an E-selectin binding form of the
glycoprotein ESL-1 (33), whereas in the same cells, PSGL-1 had no
selectin binding activity. In contrast, both ESL-1 and PSGL-1 were able
to bind to selectins in PC7 cells (33). In agreement with these
results, we now find that Fuc-TIV generates HECA-452 epitopes on ESL-1
but not on PSGL-1, whereas in the presence of Fuc-TVII HECA-452
epitopes are found on both selectin ligands (Fig.
4).
View larger version (38K):
[in this window]
[in a new window]
|
Fig. 4.
Expression of sLex epitopes on
ESL-1 and PSGL-1. Endogenously expressed ESL-1 (upper
panels) and ectopically expressed PSGL-1 (lower panels)
were immunoprecipitated from PC4 and PC7 cells (as indicated) or from
non-Fuc-T expressing PC cells () using affinity-purified polyclonal
antibodies against each antigen, electrophoresed on 6% polyacrylamide
gels under reducing conditions, and analyzed in immunoblots with the
mAb HECA-452 or as control with antibodies against ESL-1 or PSGL-1, as
indicated. The positions of ESL-1 and PSGL-1 are indicated by
arrows. PSGL-1 is only partially reduced and appears as the
dimeric 230 kDa and the monomeric 130 kDa form (see arrows).
Molecular mass markers (in kDa) are indicated on the
right.
|
|
Most of Fuc-TIV-generated sLex, but Only a Small
Portion of Fuc-TVII-generated sLexon the Cell Surface, Are
Found on Glycolipids--
We tested whether the large
protease-resistant portion of Fuc-TIV-generated sLex would
be sensitive to delipidation. To this end, PC4 and PC7 cells were
treated with methanol/chloroform in a 1:1 ratio, and residual
sLex-, GM3-, and PSGL-1-expression was then analyzed by
confocal immunofluorescence microscopy. We found that delipidation
strongly reduced the binding of HECA-452 as well as CSLEX-1 on PC4 but
not on PC7 cells (Fig. 5A).
For quantification, five different areas per slide were scanned, and
relative intensities were compared. This allowed us to determine that
74% (±6%) of Fuc-TIV-generated HECA-452 epitopes, but only 22%
(±8%) of Fuc-TVII-generated HECA-452 epitopes, could be removed by
delipidation (Fig. 5B). Staining with the sLex
antibody CSLEX-1 gave similar results (Fig. 5B). For
control, the glycoprotein PSGL-1 was stained with a monoclonal antibody following delipidation and was found to be unaffected, whereas the
glycolipid GM3 was completely removed by methanol/chloroform extraction. Residual GM3 staining with the -GM3 antibody was identical to background staining with an isotype-matched control antibody (Fig. 5A, and data not shown). These data show that
glycolipids are the main carriers of sLex generated by
Fuc-TIV. In contrast, Fuc-TVII-generated sLex epitopes are
primarily glycoprotein-bound.
View larger version (44K):
[in this window]
[in a new window]
|
Fig. 5.
Sensitivity of sLex epitopes on
Fuc-T transfectants to extraction with methanol/chloroform.
Adherent PC4 and PC7 cells (as indicated) and non-Fuc-T
expressing PC cells (control cells) were first
fixed with 4% paraformaldehyde and then either extracted with
methanol/chloroform (1:1) (after delipidation) or not extracted
(without delipidation) and subsequently stained with the mAbs HECA-452,
CSLEX-1, GMR6 (-GM3), and 4RA10 (-PSGL-1) followed by incubation
with fluorescence-labeled secondary antibodies and analysis by
immunofluorescence microscopy. A, photomicrographs of
cells stained with HECA-452, CSLEX-1, -GM3, and -PSGL-1 were
captured and analyzed by confocal microscopy. B, five
different areas on each slide were scanned, and sLex
specific signals were quantified. Quantifications from three
independent experiments are shown. Relative mean fluorescence intensity
(MFI) ± S.D. is shown as a percentage of the
fluorescence signal obtained with non-extracted cells. Signals of
non-extracted cells are depicted as open bars,
signals of methanol/chloroform extracted cells as closed
bars. No signal was seen with control cells not expressing
Fuc-Ts.
|
|
Interestingly, comparison of the Fuc-TIV-generated acidic epitopes
sLex and CD65s (VIM-2) with the neutral epitope CD15
revealed that only the molecules decorated with the acidic epitopes
could be well removed by delipidation, whereas molecules decorated with the neutral CD15 epitopes were only partially removed by this treatment
(Fig. 6). Thus, Fuc-TIV fucosylates
acidic sialyl-polylactosamine preferentially on glycolipids, and
neutral lactosamine residues are also fucosylated on non-lipid carrier
molecules.
View larger version (58K):
[in this window]
[in a new window]
|
Fig. 6.
Sensitivity of sLex,
Lex, and the VIM-2 epitope on PC4 cells to extraction with
methanol/chloroform. PC4 cells were principally analyzed as
described for Fig. 7. Fixed cells were either not extracted
(without delipidation) or extracted with
methanol/chloroform (1:1) (after delipidation)
and subsequently stained with mAbs HECA-452 or CSLEX-1 against
sLex or with a mAb against Lex (CD15) or the
mAb VIM-2 against CD65, followed by incubation with
fluorescence-labeled secondary antibodies and analysis by
immunofluorescence microscopy.
|
|
Binding of Fuc-TIV and Fuc-TVII Transfectants to
E-selectin--
Using various binding assays, we tested whether
Fuc-TIV or Fuc-TVII-generated sLex on the surface of PC4
and PC7 cells would support binding to E-selectin. As analyzed by flow
cytometry with an E-selectin-IgG chimeric probe, PC7 and PC4 cells were
both stained by the construct at 4 °C, although PC7 cells gave
brighter signals (Fig. 7A).
Thus, Fuc-TIV-generated sLex does exhibit E-selectin
binding activity. Interestingly, all of this E-selectin binding
activity on PC7 as well as on PC4 cells could be removed by treatment
of the cells with proteinase K, although PC4 cells still expressed most
of its sLex on the cell surface (Figs. 3 and
7A). This suggests that E-selectin binding was almost
exclusively mediated by glycoproteins on PC7 and PC4 cells in these
assays. When the same experiments were performed at 37 °C,
E-selectin bound to PC4 cells equally well as it did to PC7 cells (Fig.
7B). After proteinase K treatment, PC4 and PC7 cells showed
12 and 4% residual binding activity, respectively (Fig.
7B).
View larger version (33K):
[in this window]
[in a new window]
|
Fig. 7.
Protease-sensitivity of the binding of
E-selectin-IgG to Fuc-T transfectants analyzed by flow cytometry.
PC4 and PC7 cells were treated with proteinase K
(PK(+), thick line) or without proteinase K
(PK(), shaded graph) and subsequently
analyzed by flow cytometry for binding of E-selectin-IgG at 4 °C
(upper panels) and at 37 °C (lower panels).
Negative control staining was performed with human IgG1
(dotted line). The depicted experiment is representative of
five similar experiments.
|
|
When E-selectin binding was tested by immunofluorescence microscopy, a
method less sensitive than flow cytometry, we found that Fuc-TIV
transfected cells could not be stained with soluble E-selectin-IgG,
although they were brightly positive for HECA-452 (data not shown). In
contrast, Fuc-TVII transfectants could be brightly stained with
E-selectin-IgG, and this staining was resistant to delipidation with
methanol/chloroform (data not shown), again suggesting that
Fuc-TVII-generated, sLex-bearing E-selectin ligands are
mainly formed by glycoproteins.
In static cell adhesion assays with E-selectin-IgG or control proteins
coated at different concentrations, the binding of PC4 and PC7 cells
differed considerably. At 4 °C only PC7 cells bound efficiently
(Fig. 8A), suggesting that
Fuc-TIV and Fuc-TVII differ dramatically in their ability to generate
E-selectin binding activity in cells, although they generate similar
levels of sLex. E-selectin binding of both cell types was
completely blocked by proteinase K treatment of the cells (Fig.
8A), implying that all of this binding is mediated by
sLex-like structures on glycoproteins.
View larger version (22K):
[in this window]
[in a new window]
|
Fig. 8.
Adhesion of Fuc-T transfectants to
immobilized E-selectin-IgG after treatment with proteinase K. Adhesion assays were performed with PC4 and PC7 cells in 96-well
microtiter plates coated with E-selectin-IgG. No binding was observed
in the presence of EDTA or when human IgG1 or vascular
endothelial cadherin-IgG were coated (not shown). Adhesion was measured
at different concentrations of E-selectin-IgG used for coating, as
indicated. Assays were performed at 4 °C (A), and at
37 °C (B and C). Prior to the assay, cells
were either mock treated (PC4 or PC7) or treated with proteinase K
(PK+). Each depicted experiment is representative of three
similar experiments.
|
|
Adhesion assays performed at 37 °C gave similar results when plates
were coated with 0.01-0.4 µg/ml E-selectin-IgG (Fig. 8B). However, at E-selectin-IgG concentrations of 0.8 µg/ml and higher, PC4 cells could adhere as well as PC7 cells (Fig. 8B). Up to
concentrations of 0.8 µg/ml, adhesion was fully protease sensitive
(Fig. 8B). Above 0.8 µg/ml the binding of both cell types
was mediated mainly by glycoproteins, because proteinase K treatment
reduced PC7 binding to 5% and PC4 cell binding to 17.5% (Fig.
8C). All of the binding shown in Fig. 8 could be completely
inhibited by EDTA, and there was no binding of cells to vascular
endothelial cadherin-IgG, human IgG1, or 10% FCS-coated
plates in parallel experiments (data not shown). These observations
allow us to conclude that Fuc-TIV and Fuc-TVII differ in their capacity
to generate ligands that support cell-binding to E-selectin.
Furthermore, most of the E-selectin binding activity is mediated by glycoproteins.
To further distinguish the ability of sLex-bearing cells
transfected with Fuc-TIV or Fuc-TVII to bind to E-selectin in a more physiological situation, adhesion assays were performed at different flow conditions in parallel plate laminar flow chambers (Fig. 9). E-selectin-IgG was coated at 0.1 and
0.5 µg/ml on coverslips, and cells were superfused at 0.52 and 1.58 dyn/cm2. Cell rolling was only observed at lower shear
stress and only if E-selectin-IgG was coated at a concentration of 0.5 µg/ml. Seven to eight times more PC7 cells than PC4 cells were
observed to roll on E-selectin (Fig. 9). Rolling was completely
abolished in the presence of EDTA or if cells had been treated with
proteinase K (Fig. 9). No rolling was observed on human
IgG1 or when cells lacking FucTIV or Fuc-TVII were used.
Our results demonstrate that cells expressing Fuc-TIV- and
Fuc-TVII-generated sLex-like structures differ in their
ability to support rolling on E-selectin, although they express similar
levels of sLex on their surface. This may explain why the
expression level of sLex antibody epitopes on the cell
surface does not generally correlate with the ability of cells to
interact with E-selectin.
View larger version (16K):
[in this window]
[in a new window]
|
Fig. 9.
Rolling of Fuc-T transfectants on immobilized
E-selectin-IgG under flow. Flow adhesion assays were performed
with PC4 and PC7 cells and similar CHO-Pro5 transfectants
lacking Fuc-T (PC) in a planar laminar flow chamber
containing coverslips coated with 0.5 µg/ml E-selectin-IgG
(E-Sel-IgG) or human IgG1
(hIgG1). The number of rolling cells is
indicated. Cell rolling was recorded by video camera. Rolling cells
were counted in 10 different fields, and four areas were counted from
each field. No rolling cells were observed in the presence of 5 mM EDTA (not shown). Proteinase K (PK) treatment
of cells prior to the adhesion assay is indicated. The depicted
experiment is representative of three similar experiments.
|
|
The Non-catalytic Portion of Fuc-TIV Mediates the Specificity for
Glycolipid Substrates--
Finally, the surprising preference of
Fuc-TIV for glycolipid substrates prompted us to ask whether the
catalytic domain of this enzyme is responsible for the substrate
specificity. This might be expected, because this domain must get into
close contact with the substrate during the enzymatic reaction. To
address this question, we replaced the catalytic domain in Fuc-TIV by
the full catalytic domain of Fuc-TVII (Fig.
10A) and expressed this
chimeric construct in PC cells. Subsequently, the transfectants
(designated PC 4/7 chimera) were subjected to protease and delipidation
treatments. Interestingly, the chimeric fucosyltransferase generated
cell surface sLex on carriers that were largely proteinase
K-insensitive (Fig. 10B). Moreover, sLex on PC
4/7 chimera cells was nearly as susceptible to chloroform/methanol extraction as sLex generated by wild type Fuc-TIV (Fig. 10,
C and D). These data show that the non-catalytic
portion of Fuc-TIV, including the cytoplasmic, transmembrane, and stem
(CTS) region, is able to confer the preference for glycolipids onto the
catalytic domain of Fuc-TVII and that the catalytic domain of Fuc-TIV
is dispensable for the substrate specificity of this
fucosyltransferase.
View larger version (19K):
[in this window]
[in a new window]
|
Fig. 10.
The substrate specificity of Fuc-TIV is
independent from its catalytic domain. A, schematic
diagram of a chimeric enzyme in which the cytoplasmic (C),
transmembrane (T), and stem (S) regions of
mFuc-TIV were fused to the catalytic domain of mFuc-TVII. The fusion
was done without addition, deletion, or change of amino acids using a
megaprimer approach. The amino acids flanking the fusion site are
indicated. B, proteinase K sensitivity of sLex
on PC cells expressing mFuc-TIV (PC4), mFuc-TVII
(PC7), and the mFuc-TIV/VII chimera depicted in
panel A (PC 4/7
chimera). The relative mean fluorescence intensities
(MFI) ± S.D. from three flow cytometry assays using
antibody CSLEX-1 are shown and are given as the percentage of the
signal obtained with mock-treated cells. PK, proteinase K
treatment. C and D, sensitivity of
sLex on PC4, PC7, and PC 4/7 chimera cells to lipid
extraction with chloroform/methanol. The MFI-using antibody HECA-452
(C) and CSLEX-1 (D) are shown and are given as
the percentage of the signal obtained with mock-treated cells. Data are
from three assays (± range) in (C) and two assays (± range) in (D). A second PC 4/7 chimera clone was established
and gave equal results in the assays depicted in panels
B-D (not shown).
|
|
| |
DISCUSSION |
Both fucosyltransferases in mouse neutrophils, Fuc-TVII and
Fuc-TIV, contribute to the expression of functional selectin ligands in vivo, with Fuc-TVII being the more important one. The
lack of Fuc-TVII leads to severe defects in leukocyte extravasation and
the absence of E- and P-selectin ligands detectable by flow cytometry,
whereas the lack of Fuc-TIV alone only affects rolling velocity. In
addition, Fuc-TIV seems to be responsible for the residual low selectin
ligand activity that is still detectable in Fuc-TVII/
deficient mice (16, 17). The molecular basis for the selective contribution of each fucosyltransferase to the generation of
physiological selectin ligands is not known.
A simple explanation would be that Fuc-TVII might contribute more than
Fuc-TIV to the generation of sLex on neutrophils. This is
difficult to analyze, because none of the known antibodies against
sLex or sLex-related structures stain mouse
neutrophils. Reports on the ability of Fuc-TIV to direct expression of
sLex or related structures in transfected cells have been
controversial (8-13, 50). In vitro assays with recombinant
human Fuc-TIV demonstrated that this enzyme adds fucose effectively to
neutral N-acetyllactosamine units but is ineffective in transferring
fucose to the distal lactosamine of 2,3-sialylated polylactosamine
(51). Here we show that mouse Fuc-TIV can efficiently generate
sLex structures on the surface of transfected
CHO-Pro5 cells, which are recognized by the three mAbs
CSLEX-1, 2F3, and HECA-452. In contrast, sLex structures
that were generated by human Fuc-TIV on these cells were only
recognized by HECA-452 and not by CSLEX-1 or 2F3. Our results with
human Fuc-TIV are in agreement with previous results, demonstrating
that this enzyme could not generate CSLEX-1 epitopes in
CHO-Pro5 cells (11) but could direct expression of these
epitopes efficiently in CHO cells of the DHFR strain (11,
37). Thus, mouse and human Fuc-TIV can generate sLex
structures on the surface of CHO cells, but both enzymes differ in
their ability to generate sLex structures of various
complexity. It has been shown that CSLEX-1 loses its reactivity when
genuine sLex is modified by sulfate groups (52, 53),
whereas HECA-452 recognizes sulfated and non-sulfated sLex
equally well (54). It is conceivable that other structural elements
further modifying sLex or the presence of sLex
on more complex oligosaccharides might also selectively abolish CSLEX-1
but not HECA-452 binding. In conclusion, our results suggest that at
least mouse Fuc-TIV can efficiently generate sLex of
various complexity in vivo on the cell surface of CHO cells.
If Fuc-TVII and Fuc-TIV do not dramatically differ in their ability to
generate sLex epitopes in vivo, differences in
their ability to generate efficient selectin ligands would have to be
based on differences in the repertoire of carrier molecules that are
modified by each enzyme. Indeed, we have found previously, through the
analysis of neutrophils of gene-deficient mice, that Fuc-TVII
exclusively directs the expression of PSGL-1 glycoforms that bind with
high affinity to P-selectin. In contrast, Fuc-TIV preferentially
directs the expression of E-selectin binding glycoforms of ESL-1 (33).
We could mimic this remarkable substrate specificity for PSGL-1 in
transfected CHO-Pro5 cells transfected with PSGL-1 and
the C2GnT branching enzyme and one of the two Fuc-Ts (named PC4 and
PC7) (33). Intrigued by this specificity, we have now used the same
cells, further analyzing the repertoire of carrier molecules modified
by each enzyme on the surface of these cells. Based on the following
evidence, we suggest that most of Fuc-TIV-generated sLex is
found on glycolipids. First, sLex generated by Fuc-TIV was
resistant to proteinase K and trypsin treatment. Second, 70-80% of
HECA-452 and CSLEX-1 epitopes were removed by lipid extraction with
methanol/chloroform. Based on similar experiments with reciprocal
results, we conclude that most of the Fuc-TVII generated
sLex is found on glycoproteins; ~87% of sLex
was sensitive to protease treatment, whereas only ~25% was removed by delipidation with methanol/chloroform. It should be noted that the
preference of Fuc-TIV for lipophilic substrates and especially the
preference of Fuc-TVII for protein substrates were also seen if PSGL-1
or C2GnT were not co-transfected and if CHO-DHFR cells
were used instead of CHO-Pro5 cells (data not shown).
The surprising selectivity of Fuc-TIV for lipophilic substrates and
Fuc-TVII for glycoprotein substrates raises the question of what
mechanism mediates this selectivity. Interestingly, the selectivity
with which Fuc-TVII modifies PSGL-1 in CHO-Pro5 cells was
not found in CHO-DHFR cells where Fuc-TIV as well as
Fuc-TVII could generate P-selectin binding glycoforms of PSGL-1 (33).
Thus, the factors or the mechanism that determine ligand selectivity of
Fuc-TIV and Fuc-TVII for PSGL-1 are present in CHO-Pro5
cells but not in CHO-DHFR cells. However, preferential
generation of sLex on lipophilic molecules by Fuc-TIV was
also found for CHO-DHFR cells (data not shown). Thus, the
mechanism that prevents Fuc-TIV from modifying PSGL-1 in
CHO-Pro5 cells differs from the one that determines
preferential modification of glycolipids by Fuc-TIV.
Because fucosylation is believed to be the final step in the
biosynthesis pathway of sLex, our results suggest that
Fuc-TIV preferentially interacts with or metabolizes acidic lactosamine
residues on lipophilic substrates in the Golgi. One explanation for
this would be that Fuc-TIV has a lower Km for such
substrates in the milieu of the Golgi than for acidic lactosamine on
glycoprotein carriers. In contrast with glycoprotein-based acidic
lactosamine residues, Fuc-TIV seems to be efficient in modifying
neutral lactosamine residues on non-lipid carriers, because the
product, CD15, was much more resistant to delipidation than
sLex. Thus, our results might indeed be based on
preferential enzyme-substrate interactions.
An alternative explanation for our results would be that Fuc-TIV might
be preferentially targeted to sites in the Golgi that are enriched in
acidic glycolipids like, for example, raft-like structures. This
hypothesis would be in agreement with our finding that it was not the
catalytic domain but rather the CTS region that determined the
substrate specificity of our chimeric Fuc-TIV/VII enzyme in the Golgi.
It has been suggested previously that the CTS regions of
glycosyltransferases can affect the targeting and thereby the substrate
specificity. Grabenhorst and Conradt (40) found that the CTS regions of
various glycosyltransferases fused to the catalytic domain of Fuc-TVI
determined the activity of these chimeric forms of Fuc-TVI to generate
sLex or Lex on a reporter glycoprotein. This
provided indirect evidence that CTS regions target glycosyltransferases
into different locations along the "assembly line" of
glycosyltransferases. In light of these experiments, our results
suggest that the CTS region of Fuc-TIV might target our chimeric enzyme
to glycolipid substrate-enriched Golgi membrane domains. It remains to
be tested whether Fuc-TIV is indeed enriched in such membrane domains
and whether Fuc-TVII would possibly be excluded from these areas.
Fuc-TVII- as well as Fuc-TIV-transfected CHO-Pro5 cells
displayed similar amounts of sLex on their cell surface;
however, the former bound much better to E-selectin than the latter.
The difference was not very pronounced when binding of soluble
E-selectin-IgG was analyzed by flow cytometry, but when binding was
analyzed by the less sensitive immunofluorescence method, signals were
only detected on Fuc-TVII-transfected cells (data not shown).
Furthermore, the different binding efficiency was evident if static
cell adhesion was analyzed at low concentrations of immobilized
E-selectin-IgG or if cell binding was tested under flow conditions.
This is in agreement with the dominant importance of Fuc-TVII in
leukocyte extravasation in vivo (5, 16, 17) and with other
studies analyzing the binding of human Fuc-TVII- and human
Fuc-TIV-transfected hemopoetic cell lines (12, 13).
In addition to the dominant role of Fuc-TVII over Fuc-TIV in the
generation of highly efficient E-selectin ligands, our results allowed
us to compare the efficiency with which sLex-carrying
glycolipids and sLex-carrying glycoproteins contribute to
efficient E-selectin binding. Numerous reports have described the
binding of E-selectin transfected cells to immobilized
sLex-carrying glycolipids (35, 55, 56), and it was also
shown that E-selectin-transfected cells can roll on immobilized
glycolipids under physiologic flow conditions (34). However, the
contribution of cell surface-expressed glycolipids in the context of a
cell surface glycocalyx to the binding of cells on immobilized
E-selectin has not yet been analyzed. Our data suggest that cell
binding to E-selectin is mostly mediated by sLex-carrying
glycoproteins and only weakly by glycolipids, because binding could be
dramatically blocked by proteinase K treatment. Furthermore, most of
the sLex on Fuc-TIV transfectants, but only about 25% of
total sLex on Fuc-TVII-transfectants, were found on
glycolipids; however the low residual binding after protease treatment
was only marginally better for Fuc-TIV-transfectants than for
Fuc-TVII-transfectants (Figs. 7-9). This suggests that
sLex-bearing glycolipids on the cell surface are less
efficient ligands than sLex-carrying glycoproteins. The low
efficiency of glycolipids as E-selectin-ligands and the preferential
generation of sLex by Fuc-TIV on glycolipids might provide
an explanation as to why Fuc-TIV is less important in the generation of
selectin ligands on mouse neutrophils than Fuc-TVII.
(1,3)-Fucosyltransferase Fuc-TIV, but Not
Fuc-TVII, Generates Sialyl Lewis X-like Epitopes Preferentially on
Glycolipids*
,
TOP
ABSTRACT
INTRODUCTION
Supported by a fellowship (Fortbildungsstipendium) from the
Max-Planck Society.
