Tài liệu Báo cáo khoa học: Characterization of ICAM-4 binding to the I domains of the CD11a/CD18 and CD11b/CD18 leukocyte integrins pptx

Characterization of ICAM-4 binding to the I domains
of the CD11a/CD18 and CD11b/CD18 leukocyte integrins
Eveliina Ihanus
1
, Liisa Uotila
1
, Anne Toivanen
1
, Michael Stefanidakis
1
, Pascal Bailly
2
,
Jean-Pierre Cartron
2
and Carl G. Gahmberg
1
1
Department of Biosciences, Division of Biochemistry, University of Helsinki, Finland;
2
INSERM U76, Institut National de
Transfusion Sanguine, Paris, France
Intercellular adhesion molecule-4 (ICAM-4, LW blood
group antigen), a member of the immunoglobulin super-
family expressed on red cells, has been reported to bind to
CD11a/CD18 and CD11b/CD18 leukocyte integrins. The
location of the ICAM-4 binding sites on CD11a/CD18 and
CD11b/CD18 are not known. CD11/CD18 integrin I
domains have been found to act as major binding sites for
physiological ligands and a negatively charged glutamic acid
in ICAMs is considered important for binding. ICAM-4

lacks such a residue, which is replaced by an arginine.
However, we demonstrate here that ICAM-4 in red cells and
transfected fibroblasts interacts specifically with the I
domains of CD11a/CD18 and CD11b/CD18 integrins. The
binding was inhibited by anti-I domain and anti-ICAM-4
antibodies and it was dependent on divalent cations. Inter-
estingly, ICAM-4 negative red cells were still able to bind to
the CD11b/CD18 I domain but the binding of these cells to
the CD11a/CD18 I domain was clearly reduced. Using a
solid phase assay, we were able to show that isolated I
domains directly and specifically bind to purified recom-
binant ICAM-4 in a cation dependent manner. Competition
experiments indicated that the binding sites in ICAM-4 for
the CD11a and CD11b I domains are different. However,
the ICAM-4 binding region in both I domains seems to
overlap with the regions recognized by the ICAM-1 and
ICAM-2. Thus we have established that the I domains
contain an ICAM-4 binding region in CD11a/CD18 and
CD11b/CD18 leukocyte integrins.
Keywords: adhesion; ICAM, integrin; I domain; red cell.
The five intercellular adhesion molecules ICAM-1–5 cur-
rently known in humans, form a family of related cell
surface glycoproteins, which mediate cell adhesion by
binding to the leukocyte CD11/CD18 integrins. All the
ICAM proteins have extracellular C-type immunoglobulin-
like domains ranging in number from two to nine making
them members of the immunoglobulin superfamily [1–4].
Despite their structural similarity and integrin binding
capability, these proteins have a differential pattern of
expression and cellular distribution. ICAM-1 which consists

of five Ig-like domains, is found on the surface of leukocytes,
endothelial cells and various other cells, and can be
up-regulated by several proinflammatory cytokines [5,6].
ICAM-2 has two Ig-like domains. It is constitutively
expressed by leukocytes, endothelial cells [6], and platelets
[7]. ICAM-3 is composed of five Ig-like domains, and it is
present at high levels on resting lymphocytes, monocytes,
and granulocytes. It is the only ICAM significantly
expressed on neutrophils [8]. The expression of ICAM-4 is
restricted to erythrocytes and erythroid precursor cells [9].
ICAM-5 is expressed by subsets of neurons, exclusively
within the telencephalon of the mammalian brain [10].
The predominant cellular ligands for the ICAMs are the
leukocyte CD11/CD18 integrins, which consist of four
heterodimeric glycoproteins with specific a chains (CD11a,
-b, -c, -d) and a common b
2
chain (CD18). They play an
essential role in mediating adhesion of cells in the immune
system [1–4]. All five ICAM molecules are able to bind to
CD11a/CD18 (LFA-1, a
L
b
2
) which is expressed on all
leukocytes. The first NH
2
-terminal Ig domain of each
ICAM seems to be most important for binding [11–15].
ICAM-1, -2 and -4 have been shown to interact also with

CD11b/CD18 (Mac-1, a
M
b
2
, CR3), which is expressed
primarily on the cells of the myelo-monocytic lineage. The
thirdIg-likedomaininICAM-1[16]andthefirstNH
2
-
terminal Ig domain in ICAM-2 seem to mediate CD11b/
CD18 binding [12].
Extensive work has been carried out to localize the ligand
binding sites in the leukocyte b
2
integrins. The a chains of
the CD11/CD18 integrins contain an inserted approxi-
mately 200-amino-acid intervening domain (I domain),
which is homologous to the A domains of von Willebrand
factor, repeats in cartilage matrix protein and collagen [17].
The I domains of CD11/CD18 integrins have been shown
to contain recognition sites for most ligands of leukocyte
integrins. ICAM-1, -2 and -3 as well as several soluble
proteins such as fibrinogen, and the complement compo-
nent iC3b bind to the I domains of their receptor integrins.
Correspondence to C. G. Gahmberg, Department of Biosciences,
Division of Biochemistry, P.O. Box 56, Viikinkaari 5,
FIN-00014 University of Helsinki, Finland.
Fax: + 358 9 191 59068, Tel.: + 358 9 191 59028,
E-mail: Carl.Gahmberg@helsinki.fi
Abbreviations: ICAM, intercellular adhesion molecule; VCAM,

vascular cell adhesion molecule; CD11a/CD18, LFA-1, leukocyte
function associated antigen; CD11b/CD18, Mac-1; LW, Landsteiner–
Wiener blood group antigen; GST, glutathione S-transferase.
(Received 14 January 2003, accepted 20 February 2003)
Eur. J. Biochem. 270, 1710–1723 (2003) Ó FEBS 2003 doi:10.1046/j.1432-1033.2003.03528.x
By using monoclonal antibodies reacting with integrin
I domains and by mutational analysis of the I domains,
evidence has been obtained that the binding sites on the
I domains for different ligands are overlapping but not
identical [18–22]. Integrins need divalent cations for their
activity, and they have been shown to bind Ca
2+
and Mg
2+
[23,24]. Importantly, the I domains have been shown to
bind divalent cations [20,24].
ICAM-4 was originally identified as a 42-kDa red cell
membrane glycoprotein called the LW (Landsteiner–
Wiener) blood group antigen [25]. The LW protein has
been reported to require intramolecular disulfide bonds
and the presence of divalent cations, notably Mg
2+
,for
antigenic activity [26]. The LW and Rh blood groups
show an interesting phenotypic relationship, as the level of
LW expression is greater in RhD-positive than in RhD-
negative cells, and extremely rare Rh
null
cells, which lack
all Rh antigens, are also deficient in the LW protein.

However, individuals lacking LW antigens have been
found among RhD-positive individuals [27]. The LW
glycoprotein has been renamed ICAM-4 based on strong
sequence similarities with the ICAM family and subse-
quent work, which showed that it binds CD11a/CD18
and CD11b/CD18 integrins [9,28]. The ICAM-4 protein
contains two immunoglobulin domains of which the first
domain is 30% identical to the first domains of ICAM-1,
-2 and -3. CD11a/CD18 has been found to bind to the
first Ig domain, whereas CD11b/CD18 binding sites
encompass both domains of ICAM-4 [14]. Unlike the
other ICAMs, ICAM-4 does not contain the conserved
glutamate residue in the first domain, which is replaced by
an arginine 52 residue (Table 1). Mutation of arginine 52
back to glutamate did not affect CD11a/CD18 binding
and even reduced the interaction with CD11b/CD18.
Instead, site-directed mutagenesis studies identified other
residues on the CFG face of the first domain, which are
involved in CD11a/CD18 recognition. These data suggest
that the b
2
integrin binding motifs of ICAM-4 differ from
those of other ICAMs [14]. However, the adjacent
residues at arginine 52 in ICAM-4 are identical or similar
in the five ICAMs, suggesting a potential role for these
residues. The noncharged LLG sequence present on
ICAM-1 has been reported to function as a ligand-
binding motif for CD18 integrins [29]. A recent study has
suggested that ICAM-4 can also bind through novel
motifs to a

4
b
1
and a
v
family integrins [30].
In the present study, we wanted to define the role of the
CD11a and CD11b I domains in ICAM-4 binding. Our
results show that ICAM-4 binds specifically to the CD11a
and CD11b I domains.
Materials and methods
Antibodies
The b
2
integrin specific mAbs used in these studies include
TS1/22, MEM83, MEM30, MEM25, MEM177, 7E3, 60.1,
LM2/1, MEM170, 44, 107 and 904. TS1/22 (American
Type culture Collection, Rockville, MD), MEM83,
MEM30, MEM25 and MEM177 recognize the a chain of
CD11a/CD18. TS1/22, MEM83, MEM30 and MEM25 has
been mapped to the I domain of the CD11a/CD18 [31,32].
The anti-CD11b mAbs 7E3, 60.1, LM2/1 (American Type
culture Collection, Rockville, MD), MEM170, 44, 107 and
904 have been described previously [33,34] and are specific
for the I domain. The ICAM-1 antibodies have been
described: GP8911, GP8914 and GP8923 (the Leukocyte
Typing Workshop V), LB-2 [35] and RR1/1 [36]. The
ICAM-1 mAb B-H17, was a generous gift from C. Vermot-
Desroches, Diaclone, France. The three ICAM-2 mAbs
(B-T1, B-R7 and B-S9) have been described previously [37].

The mAbs BS46 and BS56 react with the first domain of the
ICAM-4 [14,38]. A mouse IgG
1
negative control was
purchased from Silenius (Hawthorn, Australia) and poly-
clonal goat anti-GST antibodies from Pharmacia Biotech
Inc. The goat antihuman IgG specific antibody was
obtained from Sigma and the peroxidase-conjugated anti-
GST mAb from Santa Cruz Biotechnology (Santa Cruz,
CA).
Purification of the CD11a/CD18 and CD11b/CD18
integrins
CD11a/CD18 and CD11b/CD18 integrins were purified
from human blood buffy coat cell lysates as described
previously [12].
Expression and purification of the CD11a and CD11b I
domains
The Escherichia coli strain JM109 transformed with the
pGEX-2T (Pharmacia Biotech Inc.) plasmid containing the
cDNA fragment of the CD11b I domain was a generous
gift from A. Arnaout (Massachusetts General Hospital,
Boston, MA) [20]. The cDNA fragment of CD11a I domain
inserted into the expression vector pGEX-5X-3 (Pharmacia
Biotech Inc.) was kindly provided by D. Altieri and was
constructed as reported by Muchowski et al.[39].The
pGEX-5X-3 plasmid construct containing the cDNA of
CD11a I domain was transformed into the E. coli strain
BL21. The glutathione S-transferase fusion proteins of the
I domains were expressed in Escherichia coli cells (strain
BL21 for CD11a I domain; strain JM109 for CD11b

I domain) as described previously [20,39]. Cell pellets
derived from 1 to 4 L culture were thawed and lyzed by
resuspending in 30 mL of lyzing buffer (10% sucrose, 0.5%
Triton X-100, 50 m
M
Tris, pH 8.0) containing 5 m
M
EDTA, 1 m
M
dithiothreitol, 1 m
M
phenylmethylsulfonyl
fluoride, 5 lgÆmL
)1
aprotinin and leupeptin. After lysozyme
(350 lgÆmL
)1
) treatment the cells were sonicated on ice.
Triton X-100 was then added to a final concentration of
1%, and the sonication was repeated. After centrifugation
at 12 000 g for 10 min, the supernatant was incubated with
Table 1. Sequence alignments of the first immunoglobulin domains of
ICAM molecules illustrating the glutamate to arginine (shown in bold)
difference in ICAM-4 and the surrounding conserved residues (shown as
italic).
ICAM-4
NSSLRTPLRQ
ICAM-1 LLGIETPLPK
ICAM-2 VGGLETSLNK
ICAM-3 KIALETSLSK

ICAM-5 RGGLETSLRR
Ó FEBS 2003 Interaction of ICAM-4 with leukocyte integrins (Eur. J. Biochem. 270) 1711
prewashed glutathione-Sepharose 4B (Pharmacia Biotech,
Uppsala, Sweden) for 2 h at 4 °C. The resin was then
washed with 50 mL of 50 m
M
Tris/HCl, pH 8.0, 50 mL of
500 m
M
NaCl,1%TritonX-100,50m
M
Tris/HCl, pH 8.0
andagainwith50mLof50m
M
Tris/HCl, pH 8.0. The
fusion proteins were eluted with 20 m
M
reduced gluta-
thione, 50 m
M
Tris/HCl, pH 8.0 and the samples were then
passed through a Bio-Gel P-6DG column to remove
glutathione.
For the experiments, which utilized the I domains
separated from the GST moiety, the purified GST fusion
proteins of the I domains of CD11a/CD18 and CD11b/
CD18 were cleaved with Factor Xa and thrombin, respect-
ively. The GST fusion protein of CD11a I domain was
cleaved with biotin-labeled restriction protease Factor Xa
(Boehringer Mannheim). The biotin-labeled Factor Xa was

removed using a streptavidin gel and the free GST using the
glutathione resin. To release the recombinant CD11b
I domain the fusion protein was treated with thrombin
(Sigma) and the cleaved sample was then further purified by
ion exchange chromatography on a Mono S HR5/5 column
(Pharmacia) using the FPLC system (Pharmacia). Analysis
of the purified recombinant I domains on 12% SDS/PAGE
revealed a single band of the expected size after staining with
Coomassie blue.
Expression and purification of ICAM-Fc recombinant
proteins
The ICAM-2Fc and ICAM-4Fc fusion proteins were
produced by transient transfection of COS-1 cells by the
DEAE-dextran method (Pharmacia) and isolated from the
culture supernatants by protein A-Sepharose CL-4B (Phar-
macia) chromatography essentially as previously described
[14,40]. ICAM-1Fc and VCAM-1Fc fusion proteins were
obtained from R & D Systems. All the recombinant
proteins were checked by SDS/PAGE and Western blotting.
The ICAM-2Fc cDNA vector was kindly provided by D.
Simmons (John Radcliffe Hospital, Oxford, UK) and the
ICAM-4Fc cDNA has been described previously [14].
Cells and cell lines
Blood samples from common LW and Rh phenotypes
(ICAM-4 positive red cells) were obtained from normal
volunteers using heparin as an anticoagulant and the
LW(a
–
,b
–

) blood sample (ICAM-4 negative red cells) was
kindly provided by Kathy Burnie (Hematology University
Hospital, Ontario, Canada). The LW(a
–
,b
–
) cells and the
control cells were stored at )70 °C until further used.
The L929 mouse fibroblast cell line was maintained in
IMDM medium supplemented with 10% fetal bovine serum,
100 UÆmL
)1
penicillin, and 100 lgÆmL
)1
streptomycin. The
full-length ICAM-1 or ICAM-2 cDNAs in pEF-BOS vector
and the full-length ICAM-4 cDNA in pcDNA I vector were
separately cotransfected with pCDM8-neo stuffer into L929
mouse fibroblast cells according to standard procedures
using either the Lipofectamine reagent kit (Life Techno-
logies, Gaithersburg, MD) or the calcium phosphate preci-
pitation method. Stable transfectants were selected in
medium containing 0.5 mgÆmL
)1
G418. The G418-resistant
cell populations were analyzed for ICAM expression with a
Becton Dickinson (Immunocytometry Systems, San Jose,
CA) FACScan flow cytometer. The L929 cells expressing
either ICAM-1, ICAM-2 or ICAM-4 were cloned by limiting
dilution. The SV40-transformed African green monkey

kidney cell line COS-1 (ATCC) was grown in DMEM
supplemented with 10% fetal bovine serum, 100 UÆmL
)1
penicillin, and 100 lgÆmL
)1
streptomycin.
Flow cytometry studies
Wild type, ICAM-1-, ICAM-2-, or ICAM-4-transfected
L cells and ICAM-4-positive or -negative red cells were
washed and resuspended in NaCl/P
i
, pH 7.4. Aliquots of
1 · 10
6
L cells or 1–3 lL of packed red cells were
incubated with 25 lgÆmL
)1
of different anti-ICAM mAbs
for 30–60 min on ice. The cells were washed with NaCl/P
i
and incubated with FITC-conjugated rabbit antimouse
F(ab¢)
2
(Dakopatts a/s, Copenhagen, Denmark) for 30 min
on ice. After washing, 1 · 10
4
cells were analyzed with a
Becton Dickinson (Immunocytometry Systems, San Jose,
CA, USA) FACScan flow cytometer.
Red cell binding assays

Indicated amounts of the purified CD11b/CD18, recom-
binant CD11a and CD11b I domains or control proteins
were coated on plastic 96-well plates (Nunc, Roskilde,
Denmark) in 25 m
M
Tris, pH 8.0, 150 m
M
NaCl, 2 m
M
MgCl
2
by incubation overnight at 4 °C. The wells were
blocked with 1% BSA for 2 h at room temperature and
washed. ICAM-4 positive or negative red cells (0.7 · 10
6
per
well) in binding buffer (RPMI 1640 supplemented
with 50 m
M
Hepes, pH 7.4, 2 m
M
MgCl
2
,2m
M
CaCl
2
,
and 5% fetal bovine serum) were added to the wells and
the plates were then briefly centrifuged (900 g,2· 1min)

and incubated for 2 h at 37 °C. The input of red cells
was quantitated by counting cells in four randomly
chosen fields from duplicate wells. To remove non-
adherent cells, the wells were gently filled with binding
buffer, and the microplate was placed floating upside
down for 40 min in NaCl/P
i
solution before microscopic
observation and scoring the number of attached cells in
four randomly chosen fields from duplicate wells. The
data was presented as a percentage of bound cells
(amount of bound red cells divided by input of cells). For
blocking experiments, the cells or protein-coated wells were
pretreated with different mAbs (25 lgÆmL
)1
) or soluble
GST/I domain GST (0.2–3 l
M
)for10minatroom
temperature before starting the adhesion. For the binding
study with or without divalent cations, the adhesion assays
were performed with buffers containing 4 m
M
EDTA,
4m
M
EGTA and 2 m
M
MgCl
2

,or2m
M
MgCl
2
and 2 m
M
CaCl
2
. The results of antibody inhibition assays were
expressed as a relative percentage of attached cells, where
100% is given as the number of cells bound in the absence of
inhibitors. The significance was determined by unpaired
Student’s t-test.
Adhesion assays of ICAM transfectants
The control protein glycophorin A (1 lg per well) and the
purified CD11a/CD18 (1 lg per well) and CD11b/CD18
1712 E. Ihanus et al.(Eur. J. Biochem. 270) Ó FEBS 2003
(0.2 lg per well) were diluted in 25 m
M
Tris, pH 8.0,
150 m
M
NaCl, 2 m
M
MgCl
2
andattachedtoflat-bottom,
96-well microtiter plates (Nunc, Roskilde, Denmark) in
0.01–0.1% n-octyl glucoside by overnight incubation at
4 °C. The wells were blocked with 1% BSA for 2 h at

room temperature. For the experiments with coated GST
I domains the anti-GST antibodies (Pharmacia Biotech
Inc.)dilutedin25m
M
Tris, pH 8.0, 150 m
M
NaCl, 2 m
M
MgCl
2
were adsorbed overnight at 4 °C, at 0.2–1 lgper
well (50 lL per well in triplicate) on flat-bottom, 96-well
microtiter plates. After blocking nonspecific sites, indica-
ted amounts (0.4–2 lg per well) of GST or recombinant
purified GST I domains were added to the wells and
incubated for 2 h at room temperature. Wells were then
washed three times with the binding medium (Iscove’s
MDM with 50 m
M
Hepes, pH 7.4, 0.5% BSA, 2 m
M
MgCl
2
and 2 m
M
CaCl
2
) prior to addition of the cells
(1.5 · 10
5

per well, in the binding medium unless other-
wise indicated). For blocking experiments, the cells or
protein-coated wells were pretreated with different mAbs
(25 lgÆmL
)1
) or soluble GST/I domain GST (0.2–4 l
M
)
for 10 min at room temperature. After 20 min incubation
at room temperature the wells were filled with binding
buffer, and the microplate was put to float upside down
for 1 h in NaCl/P
i
solution to remove unbound cells. The
bound cells (or the total amount of added cells without
washing) were lyzed in 100 lL per well phosphatase
substrate-containing lyzis buffer (1% Triton X-100 and
50 m
M
sodium acetate, pH 5.0) and incubated at 37 °C
for 30 min. The reaction was terminated by adding 50 lL
per well of 1
M
NaOH and the absorbance at 405 nm was
measured [41].
Solid phase ELISA assay
The 96-well plates (Greiner) were coated overnight at 4 °C
with 400 ng per well of goat antihuman IgG Fc specific
antibody (Sigma) in 50 m
M

Tris/HCl, pH 7.4, 150 m
M
NaCl (TBS). After blocking nonspecific sites with 3%
BSA in TBS for 1 h at 37 °C the wells were washed three
times with TBS, 1 m
M
CaCl
2
,1m
M
MgCl
2,
0.05% Tween
20, 1% BSA. The recombinant proteins (200 ng per well)
diluted in TBS, 1 m
M
CaCl
2
,1m
M
MgCl
2
,1%BSAwas
then added to the wells and incubated for 2 h at room
temperature. The wells were washed as before, and 50 lL
per well I domain GST or control GST (0–20 lgÆmL
)1
)
diluted in TBS, 1 m
M

CaCl
2
,1m
M
MgCl
2
,1%BSAwas
added and incubated at room temperature for 2 h. Plates
were washed gently three times with TBS, 1 m
M
CaCl
2
,
1m
M
MgCl
2,
0.05% Tween 20, 1% BSA prior to the
addition of a peroxidase-conjugated anti-GST mAb
(1 : 500 dilution, Santa Cruz Biotechnology) to the wells.
After 1 h of incubation at 37 °C,theplateswerewashed
as before, and the bound proteins were detected with
100 lLperwell0.5mgÆmL
)1
o-phenylenediamine dihydro-
chloride added for 10 min, stopped by addition of 50 lL
of 12.5% H
2
SO
4

and plates were read in an ELISA
reader. For inhibition experiments, the soluble GST
fusion proteins or protein-coated wells were pretreated
with different mAbs (20–40 lgÆmL
)1
) or inhibitor proteins
(ICAMFc proteins, 200 n
M
; I domains, 1 l
M
) diluted in
TBS, 1 m
M
CaCl
2
,1m
M
MgCl
2
,1%BSAfor10minat
room temperature before the addition of the I-GST
protein to the wells.
Results
Purification of CD11/CD18 integrins, recombinant
I domains and ICAMFc proteins
The purified CD11a/CD18 and CD11b/CD18 preparations
were analyzed by SDS/PAGE and no major impurities were
observed (not shown). In contrast to the CD11b I domain
GST, the majority of the CD11a I domain GST was found
in the insoluble fraction of the E. coli lyzate. Modification

of the solubilization conditions and increasing the volumes
of the bacterial cultures yielded large enough amounts of
soluble CD11a I domain GST. The purified I domain GST
fusion proteins migrated as major bands of  50 kDa. After
Factor X or thrombin cleavage and removal of GST the
I domains appeared as single major bands of approximately
24 kDa (CD11a) and 26 kDa (CD11b). The purities of
ICAM-1Fc, ICAM-2Fc and ICAM-4Fc fusion proteins
were checked by SDS/PAGE. The preparations contained
the expected recombinant proteins and the purity of the
proteins was >90% (not shown).
Binding of red cells to purified I domains
The surface expression of ICAMs on ICAM-4 positive,
LW(a+, b–), and ICAM-4 negative, LW(a–, b–), red
cells was studied by flow cytometry. The only ICAM
expressed on LW(a+, b–) cells was ICAM-4, while none
of the ICAMs were found on LW(a–, b–) cells (not
shown).
By performing cell adhesion assays with the recombinant
I domains, we showed that red cells readily bound to
recombinant I domains of both CD11a and CD11b in a
concentration-dependent manner (Fig. 1A,C). The binding
of ICAM-4 positive erythrocytes to CD11a I domain was
effectively blocked by monoclonal antibodies to ICAM-4
and the CD11a I domain (Fig. 1B). The mAb 7E3 and the
mAbs BS46 and BS56 to ICAM-4 partially but significantly
inhibited the interaction between ICAM-4 positive red cells
and recombinant CD11b I domain (Fig. 1D). The ICAM-4
negative red cells retained most or all binding activity to the
recombinant CD11b I domain (Fig. 1D). The adhesion of

these cells was affected neither by mAbs to ICAM-4 nor the
activation-dependent mAb 7E3 [42]. However, the adhesion
of both ICAM-4 positive and negative erythrocytes to
coated CD11b I domain was almost completely abrogated
in the presence of mAb 60.1, which binds to the recombin-
ant CD11b I domain. These data indicate that there might
be an unidentified ligand in red cells that could mediate
binding to CD11b and probably also to CD11a.
CD11b I domain inhibits red cell binding to purified
integrin
Red cells bind poorly to CD11a/CD18 but more efficiently
to CD11b/CD18 [28]. Therefore, we tested inhibition of
erythrocyte binding to CD11b/CD18 by purified CD11b
I domain GST. Figure 2 shows that half maximal reduction
of cell adhesion was achieved at 0.2 l
M
I domain GST.
Ó FEBS 2003 Interaction of ICAM-4 with leukocyte integrins (Eur. J. Biochem. 270) 1713
Interaction between ICAM transfectants and purified
CD11a/CD18 and CD11b/CD18 integrins
To obtain further evidence that ICAM-4 binds to CD11a/
CD18 and CD11b/CD18 through the I domains we
generated stable mouse L cell transfectants expressing
recombinant human ICAM-4. Several ICAM-4 transfect-
ant clones were obtained and the ones expressing high levels
of ICAM-4 and strong binding to purified CD18 integrins
were chosen for further adhesion assays. The stable L cell
transfectants expressing human ICAM-1 and ICAM-2
have been established as previously described [41]. As
expected, the ICAM transfectants reacted only with the

corresponding ICAM mAb (Fig. 3A). None of the trans-
fectants reacted with mAbs to ICAM-3 or ICAM-5 (not
shown).
Using purified CD11a/CD18 and CD11b/CD18, we
studied the binding of ICAM transfectants to the integrins
coated on plastic (Fig. 3B). All the ICAM transfectants,
but not the wild-type L cells adhered to the coated integrins,
but not much to the control protein. We found that the
ICAM-4 tranfectants could bind much more efficiently to
CD11b/CD18 than to CD11a/CD18 coated on plastic
wells (approximately 60% and 15% of the total added
cells, respectively). Furthermore, the ICAM-4 transfectants
adhered more strongly to CD11b/CD18 than ICAM-1 and
ICAM-2 transfectants. Approximately 30% of these
other two ICAM transfectants bound to CD11b/CD18.
As the expression levels of ICAM-4 and ICAM-2 L cell
Fig. 1. Binding of red cells to purified CD11a/CD18 and CD11b/CD18 I domains. (A) Indicated amounts of CD11a/CD18 I domain (j)or
glycophorin A (GPA) (d) were coated per well. (B) Shows the effect of anti-ICAM-4 (BS46 and BS56) and anti-CD11a I domain (TS1/22 and
MEM83) mAbs on the binding of ICAM-4 positive and negative red cells to 1 lg of coated CD11a/CD18 I domain. (C) Wells were coated with
indicated amounts of CD11b/CD18 I domain (j)orICAM-3(d). (D) The effect of antibodies on binding of red cells to 0.4 lg of coated CD11b/
CD18 I domain was studied (anti-CD11b I domain mAbs: 7E3 and 60.1). The data in A and C is presented as a percentage of attached cells
(amount of bound cells divided by input of cells). The amounts of bound and added red cells were quantitated by counting cells in four randomly
chosen fields from duplicate wells. The results in B and D are expressed as a relative percentage of bound cells, where 100% is calculated from the
total number of ICAM-4 positive red cells bound to the I domain in the absence of pretreatment with mAbs. Controls included unrelated mouse
IgG antibody (not shown) and wells with coated control protein (GPA) or without coated protein (BSA only). The experiments were repeated 3–5
times with similar results. Data are expressed as mean ± SD and statistical significances are shown.wwwP <0.001,wP <0.1.
Fig. 2. Inhibition of red cell adhesion to purified CD11b/CD18 by
the CD11b I domain GST. The binding of erythrocytes to coated
purified CD11b/CD18 in the presence of indicated concentrations of
soluble CD11b I domain GST (s)orGST(m) are shown. Back-

ground binding of cells to BSA was substracted. The data is presented
as a percentage of attached cells. The amounts of bound and added red
cells were quantitated by counting cells in four randomly chosen fields
from duplicate wells. The experiment was repeated three times with
similar results.
1714 E. Ihanus et al.(Eur. J. Biochem. 270) Ó FEBS 2003
transfectants were clearly lower than ICAM-1 transfectants,
these results suggest that ICAM-4 might be an even more
potent ligand for CD11b/CD18 than the other two ICAMs.
To a certain extent (10–20%) the binding efficiences to
ICAM-4 varied between different preparations of CD11b/
CD18 integrins. The mAbs to CD11a/CD18, CD11b/CD18
and ICAMs clearly inhibited the binding of ICAM L cells
to coated CD18 integrins (data not presented).
To examine the role of I domains in ICAM-4 binding in
more detail we tested the ability of I domain GST fusion
proteins to block the interaction of ICAM transfectants
with purified CD11a/CD18 and CD11b/CD18 integrins
(Fig. 4). They efficiently inhibited the adhesion of ICAM
transfectants to CD11a/CD18 and CD11b/CD18 integrins.
The adhesion of ICAM-4 transfectants to CD11a/CD18
was inhibited by the CD11a I domain GST more efficiently
as compared to ICAM-1 and ICAM-2 transfectants. The
inhibition of ICAM-4 transfectant binding to CD11a/CD18
by soluble CD11a I domain GST was concentration-
dependent and 50% inhibition was obtained with an
inhibitor concentration of 0.4 l
M
. The soluble CD11b
I domain GST was a less active inhibitor of ICAM-4

transfectant adhesion to coated CD11b/CD18 integrin.
However the binding of all ICAM transfectants was
reduced to 50–60% in the presence of 1 l
M
CD11b
I domain GST.
Adhesion of ICAM transfectants to purified I domain
fusion proteins
The I domains of CD11a and CD11b contain binding sites
for ICAM-1 and ICAM-2. Further proof for the interaction
of ICAM-4 with these I domains was obtained by compar-
ing the ability of recombinant I domain fusion proteins to
support the adherence of L cell transfectants expressing
ICAM-1, ICAM-2 or ICAM-4. For our assays we immo-
bilized I domain GST fusion proteins via goat anti-GST
antibodies which presumably allowed the I domain GSTs
to be presented in favourable orientations and caused more
effective binding of the cells. As shown in Fig. 5, the
I domain fusion proteins supported the binding of all three
different ICAM transfectants in a concentration-dependent
manner. For this particular lot of I domains, 0.3–0.8 lgin
solution used to coat the wells resulted in good adherence of
ICAM L cell transfectants and low background binding of
wild-type L cells. After substraction of background binding,
approximately 30% of the total added ICAM-1 L cells
adhered to the I domain of CD11a, while approximately
20% of ICAM-2 and ICAM-4 L cells did so. However, at
high levels of CD11b I domain fusion proteins, wild-type
L cells were adherent as well, a phenomenon noticed also by
others for wild-type CHO cells [43]. The results indicate that

the I domain of CD11b interacts not only with ICAMs but
also with an unknown receptor on L cells. This interaction
with an L cell receptor is not unique for the recombinant
CD11b I domain, because wild-type L cells also adhered to
high levels of purified CD11b/CD18 integrin (data not
shown).
Effects of antibodies on binding of ICAM transfectants
to purified I domain fusion proteins
For further study, we investigated the effects of different
mAbs on the interaction of ICAM L cell transfectants
with I domain GST fusion proteins of CD11a and
CD11b (Figs 6 and 7). The CD11a I domain specific
TS1/22 mAb efficiently inhibited the binding of ICAM-1
and -2 transfected L cells to the I domain of CD11a, and
partially but significantly inhibited the interaction
between ICAM-4 L cells and CD11a I domain. The
adhesion of all ICAM transfectants to the I domain of
CD11a was almost completely blocked by the anti-
CD11a mAb MEM83 down to GST background level
(Fig. 6). In an ELISA assay both anti-CD11a mAbs
(TS1/22 and MEM83) and all the I domain specific anti-
CD11b mAbs (LM2/1, MEM170, 60.1, 44a, 107 and
904) reacted with the corresponding I domain GST
fusion proteins immobilized via goat anti-GST antibodies
(not shown).
Pretreatment of the coated CD11b I domain fusion
protein with the I domain specific mAbs resulted in efficient
inhibition of the adherence of all three different types of
ICAM L cells, except for mAb 904 which had no effect on
binding of ICAM-2 transfectants (Fig. 7).

Fig. 3. Cell surface expression of ICAM-1, ICAM-2 and ICAM-4 on
L cell transfectants and adhesion assay using purified CD11a/CD18 and
CD11b/CD18 integrins. (A) Parental L cells (dotted line) and L cells
transfected (dark line) with ICAM-1, ICAM-2 and ICAM-4 cDNAs
were stained with mAbs anti-ICAM-1 (LB-2), anti-ICAM-2 (B-T1),
and anti-ICAM-4 (BS46) followed by FITC-rabbit antimouse IgG
F(ab¢)
2
fragments and were analyzed by flow cytometry. (B) Cell
adhesion of parental L cells and ICAM transfectants to CD11a/CD18,
CD11b/CD18 and GPA proteins coated on plastic wells. The experi-
ments were repeated three times with similar results.
Ó FEBS 2003 Interaction of ICAM-4 with leukocyte integrins (Eur. J. Biochem. 270) 1715
Fig. 4. Inhibition of adhesion of ICAM transfectants to purified CD18 integrins by corresponding recombinant I domain GST. The binding of ICAM
transfectants to plastic coated purified CD11a/CD18 or CD11b/CD18 in the presence of indicated concentrations of soluble CD11a I domain GST
or CD11b I domain GST (s)orGST(m) are shown. Background binding of cells to BSA was substracted. The results are expressed as a relative
percentage of bound cells, where 100% is given as the total number of cells bound to the CD18 integrins in the absence of soluble competitors
(Materials and methods). The experiments were repeated three times with similar results.
Fig. 5. Adhesion of ICAM tranfectants to purified CD11a and CD11b I domain GST fusion proteins. Indicated amounts of anti-GST antibodies were
coated per well. After blocking with BSA, GST fusion proteins of I domains or purified GST were added in amount twice of the anti-GST. Both
binding of parental L cells to GST I domains (s)orGST(e) and binding of ICAM transfectants to GST I domains (j)orGST(d) are shown.
The data are presented as a mean percentage of attached cells (amount of bound cells divided by input of cells) ± SD.
1716 E. Ihanus et al.(Eur. J. Biochem. 270) Ó FEBS 2003
Several different ICAM-1 mAbs were tested in cellular
adhesion assays and variable degrees of inhibition were
detected. The most efficient inhibition of the binding of
ICAM-1 L cell transfectants to the I domains was
obtained with the ICAM-1 mAb LB-2 reacting with the
first domain of ICAM-1 and the mAbs GP8914 and
GP8923 which have been mapped to domains 4 and 5 of

the ICAM-1 molecule [44]. None of these anti-ICAM-1
mAbs showed inhibitory effects on adhesion of ICAM-2
or ICAM-4 L cells. MAbs to ICAM-2 and ICAM-4
inhibited the binding of ICAM-2 and ICAM-4 L cells,
respectively. The three ICAM L cell transfectants used in
the cell adhesion assays were stained with all the above
mentioned ICAM mAbs and they reacted only with the
mAbs to transfected ICAM (data not shown).
Divalent cation requirements for ICAM/b
2
integrin
interaction
Divalent cations may have multiple effects on integrin-
mediated cell adhesion including enhancement, suppres-
sion, and modification of ligand binding activity. We
have previously shown that Ca
2+
and Mg
2+
are needed
for the maximal binding of CD11a/CD18 and CD11b/
CD18 to ICAM-4 [14,28]. Here we have investigated the
effect of divalent cations on the binding of ICAM-4
transfectants to the I domains of CD11a and CD11b and
compared the results to the divalent cation requirements
of ICAM-1 and ICAM-2 transfectants. We also analyzed
the cation dependence of red cell adhesion to the
I domains.
Fig. 6. Inhibition of adhesion of ICAM trans-
fectants to purified CD11a I domain GST

fusion protein. The effects of anti-ICAM and
anti-I domain mAbs on adhesion of ICAM
transfectants to 0.4 lg of I domain GST
captured with 0.2 lg of coated anti-GST Ab.
Controls include wells with captured GST and
the binding of wild-type L cells. Background
binding of cells to BSA was substracted. The
results are expressed as a mean relative per-
centage of bound cells, where 100% is given as
the total number of cells bound to the I
domain in the absence of pretreatment with
mAbs. The experiments were repeated 3–5
times with similar results. Standard deviations
and statistical significances are shown.
wwwP <0.002,wwP <0.02,wP <0.2.
Fig. 7. Inhibition of adhesion of ICAM trans-
fectants to purified CD11b I domain GST
fusion protein. The effects of anti-ICAM and
anti-I domain mAbs on adhesion of ICAM
transfectants to 0.4 lg of I domain GST
captured with 0.2 lg of coated anti-GST Ab.
Controls include wells with captured GST and
the binding of wild-type L cells. Background
binding of cells to BSA was substracted. The
results are expressed as a mean relative per-
centage of bound cells, where 100% is given as
the total number of cells bound to the I
domain in the absence of pretreatment with
mAbs. The experiments were repeated three to
five times with similar results. Standard devi-

ations and statistical significances are shown.
wwwP <0.002,wwP <0.02,wP <0.2.
Ó FEBS 2003 Interaction of ICAM-4 with leukocyte integrins (Eur. J. Biochem. 270) 1717
The most efficient binding of all the ICAM L cell
transfectants was observed in the presence of Mn
2+
(Fig. 8). In the absence of cations, the binding of ICAM-1
transfectants to the CD11a I domain fusion protein and the
binding of ICAM-2 transfectants to both I domain fusion
proteins were completely abrogated. In the presence of
EDTA the adhesion of ICAM-4 transfectants to both
I domains was efficiently, but not totally abolished, as was
also the binding of ICAM-1 transfectants to the CD11b
I domain fusion protein. The inhibitory effects of EDTA
were clearly significant. As can be seen in Fig. 8, MgCl
2
alone or in combination with CaCl
2
supported the interac-
tion of ICAM-1 L cells with both I domains and ICAM-2
L cell adherence to the I domain of CD11a. However, the
presence of both MgCl
2
and CaCl
2
seems to be required for
high affinity binding of ICAM-2 L cell transfectants to the
I domain of CD11b as well as for adhesion of ICAM-4
transfectants to both I domains. CaCl
2

alone was not
sufficient to support the maximal binding of any of the
ICAM transfectants to the I domain fusion proteins.
Figure 9 shows that the presence of Mg
2+
and Ca
2+
is
needed for efficient adhesion of red cells to the I domains of
CD11b and CD11a, while in the absence of Ca
2+
,chelated
using EGTA, the binding of red cells was partially but
significantly inhibited. A clear reduction of binding was
observed if divalent cations were omitted (not shown) or
when EDTA was included in the reaction mixture. MnCl
2
supported the binding of red cells to the I domains
efficiently (not shown).
Characterization of isolated I domain GST binding
to ICAM-4Fc using a solid phase ELISA assay
The binding of ICAM-4 directly to CD11a and CD11b
I domains was further investigated in a cell-free assay
(Figs 10 and 11). The specificity of the adhesion in the solid
phase assay was demonstrated by the ability of I domains to
bind immobilized ICAM-4Fc in a dose-dependent fashion
compared with a lack of binding to either BSA or another
closely related Ig-family protein, VCAM-1Fc (Fig. 10B,D).
The control GST fusion protein, LLG-C4-GST [29], did not
interact with ICAM-4Fc or BSA (Fig. 10A,C), showing

that the adhesion to ICAM-4Fc was mediated by the
I domain in the fusion protein. We tested the effects of
several mAbs on the binding of isolated I domain GST to
ICAM-4Fc (Fig. 11) and found that the blocking pattern
mostly reflected that observed in cellular assays for red
cells and ICAM-4 transfectants. The MEM83 antibody
Fig. 8. Adhesion of ICAM transfectants to CD11a and CD11b I domain
GST fusion proteins in the absence or the presence of divalent cations.
Stable transfectants were harvested and washed with cation-free Tris
buffer, and then resuspended in the Tris buffer containing either 2 m
M
MgCl
2
and 2 m
M
CaCl
2
,2m
M
MgCl
2
,2m
M
CaCl
2
,2m
M
MnCl
2
,or

4m
M
EDTA. Plastic wells precoated with 0.2 lg of anti-GST Ab,
were coated with 0.4 lg of CD11a (A) or CD11b (B) I domain GST
and washed three times with appropriate buffer before adding the cells.
Background binding of the cells was substracted. The experiments
were repeated three to five times with similar results.
Fig. 9. Binding of red cells to CD11a/CD18 and CD11b/CD18 I
domains in the absence or presence of divalent cations. Red cells were
washed with cation-free Tris buffer, and then resuspended in serum-
free buffer containing either 2 m
M
MgCl
2
and 2 m
M
CaCl
2
,4m
M
EDTA, or 4 m
M
EGTA and 2 m
M
MgCl
2
. Plastic wells coated with
1 lg of CD11a I domain or 0.4 lg of CD11b I domain were washed
three times with the appropriate buffer before adding the cells. The
results are expressed as a relative percentage of bound cells, where

100% is calculated from the total number of cells bound to the I
domains in the presence of divalent cations. Background binding of
red cells to BSA and the control protein (GPA) was substracted. The
experiments were repeated 3–5 times with similar results. Standard
deviations and statistical significances are shown. wwwP <0.001,
wwP<0.01.
1718 E. Ihanus et al.(Eur. J. Biochem. 270) Ó FEBS 2003
effectively inhibited the CD11a I domain/ICAM)4Fc in-
teraction while the TS1/22 blocked to a lesser degree as did
also the ICAM-4 mAb BS46. The mAbs MEM25 and
MEM30 substantially blocked the CD11a I domain GST
binding to captured ICAM-4Fc, whereas the MEM177 was
nonblocking.
Five CD11b I domain specific mAbs were tested for their
ability to inhibit in solid phase assay. MEM170, 44a and 107
were highly active inhibitors of the CD11b I domain GST
interaction with ICAM 4, whereas LM2/1 inhibited weakly
but significantly and mAb 904 had no effect. However, the
failure of mAb 904 to inhibit CD11b I domain/ICAM-4
interaction was unexpected, as the mAb was an efficient
blocker of adhesion between ICAM-4 L cell transfectants
and coated CD11b I domain GST. We checked that all the
mAbs to the I domains and the anti-ICAM-4 mAb bound
to corresponding coated and soluble recombinant proteins
(data not shown).
As a further approach to characterize the interaction
between ICAM-4 and the I domains we examined the effect
of soluble recombinant ICAMFc proteins and the isolated
I domains on the binding of I domain GST fusion proteins
to plastic captured ICAM-4Fc in solid phase ELISA assay

(Fig. 11). The results indicated that recombinant CD11a
I domain lacking GST was a potent competitor of the
interaction between coated ICAM-4Fc and the CD11a
I domain GST fusion protein, whereas the CD11b I domain
only inhibited weakly. However, soluble CD11b I domain
readily inhibited the binding of captured ICAM-4Fc to
CD11b I domain GST, while the soluble CD11a I domain
did not have any effect. Furthermore, soluble ICAM-4Fc
was highly active in competing with the coated ICAM-4Fc
for the recombinant I domain GST fusion proteins and the
soluble ICAM-1Fc and ICAM-2Fc fusion proteins were
even more efficient competitors.
Discussion
In the present study, we have used several techniques to
show that the I domains of the CD11a/CD18 and CD11b/
CD18 leukocyte integrins contain binding sites for ICAM-4.
Our results show that ICAM-4 expressing red cells bound
specifically and dose-dependently to isolated recombinant
CD11a and CD11b I domains. The effective inhibition of
binding of ICAM-4 positive red cells by anti-ICAM-4
antibodies, indicate a major role for ICAM-4 in binding of
red cells to the I domains. The efficient inhibition of the
Fig. 10. Specific binding of purified recombinant I domain GST fusion
proteins to ICAM-4Fc in a solid phase assay. Dose-dependent binding
of CD11a I domain GST (A and B) or CD11b I domain GST (C and
D) to ICAM-4Fc, VCAM-1Fc and BSA. Recombinant ICAM-4Fc or
VCAM-1Fc was immobilized via antihuman IgG Fc specific antibody
to 96-well plates, which were blocked with BSA. Control wells were
only blocked with BSA. The I domain GST or the control GST were
diluted in Tris buffered saline containing 1 m

M
CaCl
2
,1m
M
MgCl
2
and1%BSAandincubatedfor2hatroomtemperature.Data
shown are from one representative experiment out of 3–5.
I domain GST + ICAM-4Fc (j), control GST + ICAM-4Fc (m),
I domain GST + BSA (s), control GST + BSA (e), I domain
GST + VCAM-1Fc (d).
Ó FEBS 2003 Interaction of ICAM-4 with leukocyte integrins (Eur. J. Biochem. 270) 1719
interaction between purified CD11b/CD18 and erythrocytes
by soluble CD11b I domain GST is consistent with the idea
of this subdomain being a ligand binding area for a red cell
receptor. However, it is probable that other erythrocyte
receptors for the CD11a and CD11b I domains exist
because of the residual binding of the ICAM-4 negative
cells. The mAb 60.1 reacting with the CD11b I domain
almost completely inhibited the binding of both ICAM-4
positive and negative red cells. To the best of our knowledge
the epitope for mAb 60.1 within the I domain has not been
mapped in detail but the epitope for activation dependent
mAb 7E3 has been localized to the amino-terminal region
of the CD11b I domain [42] overlapping partially with the
metal ion-dependent adhesion site. A clear reduction in
adhesion was also obtained with the CD11b I domain
specific LM2/1 mAb (not shown) which has been shown to
inhibit the binding of red cells to the CD11b/CD18 integrin

[28].
The soluble I domains also clearly reduced the binding of
all ICAM transfectants to purified CD11a/CD18 and
CD11b/CD18. However, the ICAM-4 binding to CD11b/
CD18 was reduced to 60% in the presence of the
corresponding I domain compared to 20% for the inter-
action between ICAM-4 and CD11a/CD18. This indicates
differences in binding of ICAM-4 to these two integrins.
The coated I domain fusion proteins were able to mediate
dose-dependent binding of all three different ICAM trans-
fectants.
Various anti-I domain mAbs were tested in cellular
binding assays to find out possible differences in binding of
ICAM-4 and the other two ICAMs for the I domains.
Many of these mAbs displayed similar inhibitory profiles in
assays investigating the effects of the mAbs on the
interaction of ICAM L cell transfectants with the I
domains. However, the CD11a I domain specific mAb
TS1/22 which efficiently blocked the binding of ICAM-1
and -2 L cells only partially inhibited the interaction
between ICAM-4 L cells and CD11a I domain. The
adhesion of all ICAM transfectants to the I domain of
CD11a was almost completely blocked by the I domain
specific MEM83. Previous reports have shown that
MEM83 stimulates cellular CD11a/CD18 binding to puri-
fied ICAM-1 [31,45] but inhibits the interaction between
ICAM-3 and CD11a/CD18 [46,47]. Using a flow cell assay
MEM83 has been reported to increase the number of rolling
cells expressing a membrane-anchored form of the CD11a
I domain on ICAM-1 bilayer membranes but to decrease it

on ICAM-3 bilayers. Furthermore, the MEM83 decreased
the rolling velocity of the cells on ICAM-1-containing
membranes indicating an enhanced avidity [48]. The differ-
ent results may be due to differences in assay systems. We
have investigated the static adhesion of ICAM expressing
transfected cells to the isolated recombinant I domains. The
previous results were performed with coated CD11a/CD18
or using cells expressing either a whole CD11a/CD18
integrin or a membrane-anchored form of the CD11a
I domain. Altogether, our data with both red cells and
ICAM-4 transfected L cells suggest that the epitope within
the CD11a I domain recognized by the MEM83 antibody is
involved in ICAM-4 binding.
Treatment of the coated CD11b I domain fusion protein
with I domain specific anti-CD11b antibodies resulted in an
efficient inhibition of the adherence of all three types of
ICAM L cells, except that mAb 904 had no effect on
binding of ICAM-2 transfectants. The LM2/1 mAb did not
inhibit the binding of ICAM-4 L cells as efficiently as did
the other anti-I domain mAbs. The degrees of inhibition
between mAbs may vary due to different mechanisms of
action. Several studies have pointed out that blocking mAbs
can inhibit ICAM/integrin interactions either by direct
competition for the ligand binding site or by an indirect
mechanism through binding to a regulatory site located
outside the actual ligand binding site [46,49,50]. Most
function-blocking mAbs directed against integrins may act
allosterically by stabilizing the low-affinity state of the
Fig. 11. Effect of monoclonal antibodies, soluble competitor proteins
and EDTA on binding of recombinant I domain GST to coated ICAM-

4Fc in a solid phase assay. CD11a I domain GST (A) and CD11b
I domain GST (B) binding to ICAM-4Fc was examined in the pres-
ence of mAbs (20–40 lgÆmL
)1
) against the ICAM-4 (BS46), CD11a
I domain (TS1/22, MEM83, MEM30, MEM25, MEM177), CD11b
I domain (LM2/1, MEM170, 44a, 107, 904) or soluble competitor
proteins; CD11a and CD11b I domains of which the GST part was
removed (1 l
M
), ICAM-Fc proteins (200 n
M
)or5m
M
EDTA. The
results are shown as a percentage of I domain GST binding in the
absence of soluble inhibitors or EDTA. The experiments were repeated
3–5 times with similar results. Data are from one representative
experiment. Standard deviations and statistical significances are
shown. wwwP < 0.005, wwP< 0.05, wP < 0.5. The significance
was determined by unpaired Student’s t-test.
1720 E. Ihanus et al.(Eur. J. Biochem. 270) Ó FEBS 2003
receptor and thus preventing the conformational change to
the active high affinity form. The activation-independent
mAbs LM2/1 and 904 have indeed been suggested to
recognize a common or overlapping epitope distant from
the ligand binding region [51]. MAbs 44a and 107 have been
reported to share the property of stabilizing the integrin in
the low affinity state [52,53]. However, the epitopes for
mAb 44a reside at considerable distance from the ligand

binding site on the opposite side of the metal ion-dependent
adhesion site (MIDAS) motif, whereas the binding interface
for mAb 107 seems to overlap that of physiologic ligands
and clearly engages the metal ion-dependent adhesion site
area [51–53].
The presence of Mg
2+
and Ca
2+
or Mn
2+
is required for
maximal binding efficiency of ICAM-4 transfectants and
red cells to the I domains of CD11a and CD11b. Our results
indicate that Mg
2+
or Mn
2+
, but not Ca
2+
are necessary
for the interaction of CD11a I domain with ICAM-1 and
ICAM-2 transfectants and for the CD11b I domain binding
to the ICAM-1 transfectants. The presence of both MgCl
2
and CaCl
2
gives better binding of ICAM-2 L cell transfect-
ants to the I domain of CD11b as well as for adhesion of
ICAM-4 transfectants and red cells to both I domains.

The finding that the soluble I domains were able to
compete only with the respective I domain GST fusion
protein for binding to the captured ICAM-4Fc strengthens
the earlier results that the binding sites for the CD11a and
CD11b I domains in ICAM-4 are different although they
could be partially overlapping [14]. The ICAM-4 binding
region in both I domains clearly overlaps with the region
recognized by ICAM-1 and ICAM-2. Soluble ICAM-1Fc
and ICAM-2Fc proteins prevented the binding of immobi-
lized ICAM-4Fc to the I domains almost completely
whereas the soluble ICAM-4Fc was a less active competitor,
perhaps due to lower affinity.
The observations described in this report have funda-
mental importance for the detailed understanding of the
recognition between ICAM-4 and CD11a/CD18 and
CD11b/CD18 integrins. The physiological role of
ICAM-4 is not known, but it may function as an
erythroid recognition protein for macrophage integrins in
erythroblastic islands in bone marrow during erythropoi-
esis. The expression level of ICAM-4 has been reported to
be higher on erythroid precursor cells and is gradually
decreased during the later stages of differentiation finally
downtothelevelobservedinmatureredcells[54].In
addition, the expression of the ICAM-4 integrin receptors
on macrophages suggests a possible function for this
interaction in red cell turnover to eliminate aged red cells
from the circulation by spleen macrophages. The ICAM-
4/b
2
-integrin interaction may also be important during

haemostasis where in the developing thrombus, erythro-
cytes interact with activated neutrophils and monocytes or
during wound healing to mediate removal of red cells
from the thrombus by phagocytic macrophages. The very
recent finding that ICAM-4 also binds to the platelet
integrin IIb/IIIa indicates an important role for this
molecule in haemostasis [55]. Another recent report
described the capture and adhesion of normal erythrocytes
to activated neutrophils and platelets, as well as to fibrin,
at depressed venous shear rates. In agreement with our
present and previous data the results suggested that the
binding of red cells to neutrophils might be mediated
through ICAM-4 and CD11b/CD18 [56].
Acknowledgements
We thank Leena Kuoppasalmi, Aili Grundstro
¨
m, Outi Nikkila
¨
,Tuula
Nurminen and Saija Ma
¨
kinen for excellent technical assistance, and
Yvonne Heinila
¨
for secretarial work. We also want to thank Erkki
Koivunen and Tanja-Maria Ranta for providing the LLG-C4-GST
protein. These studies were supported by the University of Helsinki, the
Academy of Finland, the Sigrid Juse
´
lius Foundation, the Magnus

Ehrnrooth Foundation and the Finnish Cancer Society.
References
1. Arnaout, M.A. (1990) Structure and function of the leukocyte
adhesion molecules CD11/CD18. Blood 75, 1037–1050.
2. Gahmberg, C.G., Tolvanen, M. & Kotovuori, P. (1997) Leukocyte
adhesion. Structure and function of human leukocyte b2-integrins
and their cellular ligands. Eur. J. Biochem. 245, 215–232.
3. Gahmberg, C.G. (1997) Leukocyte adhesion. CD11/CD18 inte-
grins and intercellular adhesion molecules. Curr.Opin.CellBiol.9,
643–650.
4. Hayflick, J.S., Kilgannon, P. & Gallatin, W.M. (1998) The inter-
cellular adhesion molecule (ICAM) family of proteins. New
members and novel functions. Immunol. Res. 17, 313–327.
5. Rothlein, R., Czajkowski, M., O’Neill, M.M., Marlin, S.D.,
Mainolfi, E. & Merluzzi, V.J. (1988) Induction of intercellular
adhesion molecule 1 on primary and continuous cell lines by pro-
inflammatory cytokines. J. Immunol. 141, 1665–1669.
6.Nortamo,P.,Li,R.,Renkonen,R.,Timonen,T.,Prieto,J.,
Patarroyo, M. & Gahmberg, C.G. (1991) The expression of
human intercellular adhesion molecule-2 is refractory to
inflammatory cytokines. Eur. J. Immunol. 21, 2629–2632.
7. Diacovo,T.G.,DeFougerolles,A.R.,Bainton,D.F.&Springer,
T.A. (1994) A functional integrin ligand on the surface of platelets:
Intercellular adhesion molecule-2. J. Clin. Invest. 94, 1243–1251.
8. de Fougerolles, A.R. & Springer, T.A. (1992) Intercellular adhe-
sion molecule 3, a third adhesion counter-receptor for lymphocyte
function-associated molecule 1 on resting lymphocytes. J. Exp.
Med. 175, 185–190.
9. Bailly, P., Hermand, P., Callebaut, I., Sonneborn, H.H.,
Khamlichi, S., Mornon, J P. & Cartron, J P. (1994) The LW

blood group glycoprotein is homologous to intercellular adhesion
molecules. Proc. Natl Acad. Sci. USA 91, 5306–5310.
10. Yoshihara, Y., Oka, S., Nemoto, Y., Watanabe, Y., Nagata, S.,
Kagamiyama, H. & Mori, K. (1994) An ICAM-related neuron
glycoprotein, telencephalin, with brain segment-specific expres-
sion. Neuron 12, 541–553.
11. Staunton, D.E., Dustin, M.L., Erickson, H.P. & Springer, T.A.
(1990) The arrangement of the immunoglobulin-like domains of
ICAM-1 and the binding sites for LFA-1 and rhinovirus. Cell 61,
243–254.
12. Li, R., Nortamo, P., Valmu, L., Tolvanen, M., Kantor, C. &
Gahmberg, C.G. (1993) A peptide from ICAM-2 binds to the
leukocyte integrin CD11a/CD18 and inhibits endothelial cell
adhesion. J. Biol. Chem. 268, 17513–17518.
13. Holness, C.L., Bates, P.A., Littler, A.J., Buckley, C.D.,
McDowall,A.,Bossy,D.,Hogg,N.&Simmons,D.L.(1995)
Analysis of the binding site on intercellular adhesion molecule 3
for the leukocyte integrin lymphocyte function-associated 1.
J. Biol. Chem. 270, 877–884.
14. Hermand, P., Huet, M., Callebaut, I., Gane, P., Ihanus, E.,
Gahmberg, C.G., Cartron, J P. & Bailly, P. (2000) Binding sites of
leukocyte b2 integrins (LFA-1, Mac-1) on the human ICAM-4/
LW blood group protein. J. Biol. Chem. 275, 26002–26010.
Ó FEBS 2003 Interaction of ICAM-4 with leukocyte integrins (Eur. J. Biochem. 270) 1721
15. Tian, L., Kilgannon, P., Yoshihara, Y., Mori, K., Gallatin, W.M.,
Carpe
´
n, O. & Gahmberg, C.G. (2000) Binding of T lymphocytes
to hippocampal neurons through ICAM-5 (telencephalin) and
characterization of its interaction with the leukocyte integrin

CD11a/CD18. Eur. J. Immunol. 30, 810–818.
16. Diamond, M.S., Staunton, D.E., Marlin, S.D. & Springer, T.A.
(1991) Binding of the integrin Mac-1 (CD11b/CD18) to the third
immunoglobulin-like domain of ICAM-1 (CD54) and its regula-
tion by glycosylation. Cell 65, 961–971.
17. Colombatti, A. & Bonaldo, P. (1991) The superfamily of proteins
with von Willebrand Factor type A-like domains: one theme
common to components of extracellular matrix, hemostasis, cell
adhesion, and defence mechanisms. Blood 77, 2305–2315.
18. Diamond, M.S., Garcia-Aguilar, J., Bickford, J.K., Corbi, A.L. &
Springer, T.A. (1993) The I domain is a major recognition site on
the leukocyte integrin Mac-1 (CD11b/CD18) for four distinct
adhesion ligands. J. Cell Biol. 120, 1031–1043.
19. Landis, R.C., McDowell, A., Holness, C.L.L., Littler, A.J.,
Simmons, D.L. & Hogg, N. (1994) Involvement of the ÔIÕ domain
of LFA-1 in selective binding to ligands ICAM-1 and ICAM-3.
J. Cell Biol. 126, 529–537.
20. Michishita, M., Videm, V. & Arnout, M.A. (1993) A novel diva-
lent cation-binding site in the A domain of the b2integrinCR3
(CD11b/CD18) is essential for ligand binding. Cell 72, 857–867.
21. Zhang, L. & Plow, E.F. (1996) Overlapping, but not identical, sites
are involved in the recognition of iC3b, neutrophil inhibitory
factor, and adhesive ligands by the aMb2integrin.J. Biol. Chem.
271, 18211–18216.
22. Lub, M., van Kooyk, Y. & Figdor, C.G. (1996) Competition
between lymphocyte function-associated antigen 1 (CD11a/CD18)
and Mac-1 (CD11b:/CD18) for binding to intercellular adhesion
molecule-1 (CD54). J. Leuk. Biol. 59, 648–655.
23. Gahmberg, C.G., Nortamo, P., Kotovuori, P., Tontti, E., Li, R.,
Valmu, L., Autero, M. & Tolvanen, M. (1992) Activation and

binding of leukocyte integrins to their ligands. In Leukocyte
Adhesion. Basic and Clinical Aspects. Excerpta Medica, Interna-
tional Congress Series 1010 (Gahmberg, C.G., Mandrup-Poulsen,
T., Wogensen Bach, L. & Ho
¨
kfelt, B., eds), pp. 205–219. Elsevier
Science Publishers, BV, Amsterdam.
24. Griggs, D.W., Schmidt, C.M. & Carron, C.P. (1998) Character-
istics of cation binding to the I domains of LFA-1 and MAC-1.
J. Biol. Chem. 273, 22113–22119.
25. Mallinson, G., Martin, P.G., Anstee, D.J., Tanner, M.J.A.,
Merry, A.H., Tills, D. & Sonneborn, H.H. (1986) Identification
and partial characterization of the human erythrocyte membrane
component(s) that express the antigens of the LW blood-group
system. Biochem. J. 234, 649–652.
26. Bloy, C., Hermand, P., Blanchard, D., Cherif-Zahar, B., Goos-
sens, D. & Cartron, J P. (1990) Surface orientation and antigen
properties of Rh and LW polypeptides of the human erythrocyte
membrane. J. Biol. Chem. 265, 21482–21487.
27. Cartron, J P. (1994) Defining the Rh blood group antigens.
Biochemistry and molecular genetics. Blood Rev. 8, 199–212.
28. Bailly, P., Tontti, E., Hermand, P., Cartron, J P. & Gahmberg,
C.G. (1995) The red cell LW blood group protein is an inter-
cellular adhesion molecule which binds to CD11/CD18 leukocyte
integrins. Eur. J. Immunol. 25, 3316–3320.
29. Koivunen, E., Ranta, T M., Annila, A., Taube, S., Uppala, A.,
Jokinen, M., van Willigen, G., Ihanus, E. & Gahmberg, C.G.
(2001) Inhibition of b
2
integrin-mediated leukocyte cell adhesion

by leucine-leucine-glycine motif-containing peptides. J. Cell Biol.
153, 905–915.
30. Spring, F.A., Parsons, S.F., Ortlepp, S., Olsson, M.L., Sessions,
R., Brady, R.L. & Anstee, D.J. (2001) Intercellular adhesion
molecule-4 binds a
4
b
1
and a
v
-family integrins through novel
integrin-binding mechanisms. Blood 98, 458–466.
31. Randi, A.M. & Hogg, N. (1994) I domain of b
2
integrin lym-
phocyte function-associated antigen-1 contains a binding site for
ligand intercellular adhesion molecule-1. J. Biol. Chem. 269,
12395–12398.
32. Huang, C. & Springer, T.A. (1995) A binding interface on the I
domain of lymphocyte function-associated antigen-1 (LFA-1)
required for specific interaction with intercellular adhesion mole-
cule 1 (ICAM-1). J. Biol. Chem. 270, 19008–19016.
33. Zhou,L.,Lee,D.H.S.,Plescia,J.,Lau,J.Y.&Altieri,D.C.(1994)
Differential ligand binding specificities of recombinant CD11b/
CD18 integrin-I-domain. J. Biol. Chem. 269, 17075–17079.
34. Li,R.,Rieu,P.,Griffith,D.L.,Scott,D.&Arnaout,A.(1998)
Two functional states of the CD11b A-domain: Correlations with
key features of two Mn
2+
-complexed crystal structures. J. Cell

Biol. 143, 1523–1534.
35. Patarroyo, M., Clark, E.A., Prieto, J., Kantor, C. & Gahmberg,
C.G. (1987) Identification of a novel adhesion molecule in human
leukocytes by monoclonal antibody LB-2. FEBS Lett. 210, 127–131.
36. Rothlein, R., Dustin, M.L., Marlin, S.D. & Springer, T.A. (1986)
A human intercellular adhesion molecule (ICAM-1) distinct from
LFA-1. J. Immunol. 137, 1270–1274.
37. Xie, J., Li, R., Kotovuori, P., Vermot-Desroches, C., Wijdenes, J.,
Arnaout, M.A., Nortamo, P. & Gahmberg, C.G. (1995) Inter-
cellular adhesion molecule-2 (CD102) binds to the leukocyte
integrin CD11b/CD18 through the A domain. J. Immunol. 155,
3619–3628.
38. Sonneborn, H.H., Uthemann, H., Tills, D., Lomas, C.G., Shaw,
M.A. & Tippett, P. (1984) Monoclonal anti-LWab. Biotest. Bull.
2, 145–148.
39. Muchowski, P.J., Zhang, L., Chang, E.R., Soule, H.R., Plow, E.F.
& Moyle, M. (1994) Functional interaction between the integrin
antagonist neutrophil inhibitory factor and the I domain of
CD11b/CD18. J. Biol. Chem. 269, 26419–26423.
40. Simmons, D.L. (1993) Cloning cell surface molecules by transient
expression in mammalian cells. In Cellular Interactions in
Development. A Practical Approach (Stevenson, B. ed.) pp. 93–128,
Oxford University Press, Oxford.
41. Tian, L., Yoshihara, Y., Mizuno, T., Mori, K. & Gahmberg, C.G.
(1997) The neuronal glycoprotein telenchephalin is a cellular
ligand for the CD11a/CD18 leukocyte integrin. J. Immunol. 158,
928–936.
42. Plescia, J., Conte, M.S., Van Meter, G., Ambrosini, G. & Altieri,
D.C. (1998) Molecular identification of the cross-reacting epitope
on a

M
b
2
integrin I domain recognized by anti-a
IIb
b
3
monoclonal
antibody 7E3 and its involvement in leukocyte adherence. J. Biol.
Chem. 273, 20372–20377.
43. Feng,Y.,Chung,D.,Garrard,L.,McEnroe,G.,Lim,D.,Scar-
dina,J.,McFadden,K.,Guzzetta,A.,Lam,A.,Abraham,J.,Liu,
D. & Endemann, G. (1998) Peptides derived from the com-
plementarity-determining regions of anti-Mac-1 antibodies block
intercellular adhesion molecule-1 interaction with Mac-1. J. Biol.
Chem. 273, 5625–5630.
44. Berendt,A.R.,McDowall,A.,Craig,A.G.,Bates,P.A.,Stern-
berg, M.J.E., Marsh, K., Newbold, C.I. & Hogg, N. (1992) The
binding site on ICAM-1 for Plasmodium falciparum-infected ery-
throcytes overlaps, but is distinct from, the LFA-1-binding site.
Cell 68, 71–81.
45. Myones, B.L., Daizell, J.G., Hogg, N. & Ross, G.D. (1988)
Neutrophil and monocyte cell surface p150,95 has iC3b-receptor
(CR4) activity resembling CR3. J. Clin. Invest. 82, 640–651.
46. Binnerts, M.E., van Kooyk, Y., Edwards, C.P., Champe, M.,
Presta,L.,Bodary,S.C.,Figdor,C.G.&Berman,P.W.(1996)
Antibodies that selectively inhibit leukocyte function-associated
antigen 1 binding to intercellular adhesion molecule-3 recognize a
unique epitope within the CD11a I-domain. J. Biol. Chem. 271,
9962–9968.

1722 E. Ihanus et al.(Eur. J. Biochem. 270) Ó FEBS 2003
47. Woska, J.R., Jr, Morelock, M.M., Durham Jeanfavre, D., Cavi-
ness, G.O., Bormann, B J. & Rothlein, R. (1998) Molecular
comparison of soluble intercellular adhesion molecule (sICAM)-1
and sICAM-3 binding to lymphocyte function-associated antigen-
1. J. Biol. Chem. 273, 4725–4733.
48. Knorr, R. & Dustin, M.L. (1997) The lymphocyte function-
associated antigen 1, I domain is a transient binding module
for intercellular adhesion molecule (ICAM)-1 and ICAM-3 in
hydrodynamic flow. J. Exp. Med. 186, 719–730.
49. Huth, J.R., Olejniczak, E.T., Mendoza, R., Liang, H., Harris,
E.A., Lupher, M.L., Jr, Wilson, A.E., Fesik, S.W. & Staunton,
D.E. (2000) NMR and mutagenesis evidence for an I domain
allosteric site that regulates lymphocyte function-associated anti-
gen 1 ligand binding. Proc. Natl Acad. Sci. USA 97, 5231–5236.
50. Lupher, M.L., Jr, Harris, E.A.S., Beals, C.R., Sui, L., Liddington,
R.C. & Staunton, D.E. (2001) Cellular activation of leukocyte
function-associated antigen-1 and its affinity are regulated at the I
domain allosteric site. J. Immunol. 167, 1431–1439.
51. Zhang, L. & Plow, E.F. (1999) Amino acid sequences within the
alpha subunit of integrin alpha M beta 2 (Mac-1) critical for
specific recognition of C3bi. Biochemistry 38, 8064–8071.
52. Xiong, J P., Li, R., Essafi, M., Stehle, T. & Arnaout, M.A. (2000)
An isoleucine-based allosteric switch controls affinity and shape
shifting in integrin CD11b A-domain. J. Biol. Chem. 275, 38762–
38767.
53. Li, R., Haruta, I., Rieu, P., Sugimori, T., Xiong, J P. &
Arnaout, M.A. (2002) Characterization of a conformationally
sensitive murine monoclonal antibody directed to the metal ion-
dependent adhesion site face of integrin CD11b. J. Immunol. 168,

1219–1225.
54. Bony, V., Gane, P., Bailly, P. & Cartron, J P. (1999) Time-
course expression of polypeptides carrying blood group antigens
during human erythroid differentiation. Br.J.Haematol.107,
263–274.
55. Hermand,P.,Gane,P.,Huet,M.,Jallu,V.,Kaplan,C.,Sonne-
born, H.H., Cartron, J.P. & Bailly, P. Red cell ICAM-4 is a novel
ligand for platelet activated A
IIb
b
3
-integrin. J. Biol. Chem. 278,
4892–4898.
56. Goel, M.S. & Diamond, S.L. (2002) Adhesion of normal ery-
throcytes at depressed venous shear rates to activated neutrophils,
activated platelets, and fibrin polymerized from plasma. Blood
100, 3797–3803.
Ó FEBS 2003 Interaction of ICAM-4 with leukocyte integrins (Eur. J. Biochem. 270) 1723