Báo cáo khoa học: CD91 interacts with mannan-binding lectin (MBL) through the MBL-associated serine protease-binding site doc

CD91 interacts with mannan-binding lectin (MBL) through
the MBL-associated serine protease-binding site
Karen Duus
1
, Nicole M. Thielens
2
, Monique Lacroix
2
, Pascale Tacnet
2
, Philippe Frachet
2
,
Uffe Holmskov
3
and Gunnar Houen
1
1 Department of Clinical Biochemistry and Immunology, Statens Serum Institut, Artillerivej 5, Copenhagen, Denmark
2 Laboratoire d’Enzymologie Mole
´
culaire, Institut de Biologie Structurale Jean-Pierre Ebel, Commissariat a
`
l’Energie Atomique, CNRS UMR
5075, Universite
´
Joseph Fourier, 41 rue Jules Horowitz, Grenoble Cedex 1, France
3 Department of Cardiovascular and Renal Research, Institute of Molecular Medicine, University of Southern Denmark, Odense, Denmark
Introduction
CD91 (also known as low-density lipoprotein receptor-
related protein) is a pattern-recognition receptor that is
highly expressed on human macrophages and involved

in the recognition and phagocytosis of over 30 differ-
ent ligands. It consists of two noncovalently bound
polypeptide chains: a 515 kDa alpha-chain with four
Keywords
CD91; clearance; ficolin, complement;
mannan-binding lectin (MBL); scavenging
Correspondence
K. Duus, Department of Clinical
Biochemistry and Immunology, Statens
Serum Institut, Artillerivej 5, DK-2300
Copenhagen, Denmark
Fax: +45 32683149
Tel: +45 32688241
E-mail:
(Received 12 August 2010, revised
20 September 2010, accepted 4 October
2010)
doi:10.1111/j.1742-4658.2010.07901.x
CD91 plays an important role in the scavenging of apoptotic material, pos-
sibly through binding to soluble pattern-recognition molecules. In this
study, we investigated the interaction of CD91 with mannan-binding lectin
(MBL), ficolins and lung surfactant proteins. Both MBL and L-ficolin were
found to bind CD91. The MBL–CD91 interaction was time- and concentra-
tion-dependent and could be inhibited by known ligands of CD91. MBL-
associated serine protease 3 (MASP-3) also inhibited binding between MBL
and CD91, suggesting that the site of interaction is located at or near the
MASP–MBL interaction site. This was confirmed by using MBL mutants
deficient for MASP binding that were unable to interact with CD91. These
findings demonstrate that MBL and L-ficolin interact with CD91, strongly
suggesting that they have the potential to function as soluble recognition

molecules for scavenging microbial and apoptotic material by CD91.
Structured digital abstract
l
MINT-8040679, MINT-8040706: MBL (uniprotkb:P11226) binds (MI:0407)toCD91 (uni-
protkb:
Q07954)byenzyme linked immunosorbent assay (MI:0411)
l
MINT-8040690: L-ficolin (uniprotkb:Q15485) binds (MI:0407)toCD91 (uniprotkb:Q07954)
by enzyme linked immunosorbent assay (
MI:0411)
l
MINT-8040663: C1q B (uniprotkb:P02746), C1q C (uniprotkb:P02747), C1q A (uniprotkb:
P02745) and CD91 (uniprotkb:Q07954) physically interact (MI:0915)byenzyme linked immu-
nosorbent assay (
MI:0411)
l
MINT-8040821, MINT-8040869, MINT-8040928: CD91 (uniprotkb:Q07954) binds (MI:0407)
to MBL (uniprotkb:
P11226)bysurface plasmon resonance (MI:0107)
l
MINT-8040880: CD91 (uniprotkb:Q07954) binds (MI:0407)toL-ficolin (uniprotkb:Q15485)
by surface plasmon resonance (
MI:0107)
Abbreviations
k
a,
association rate constant; K
D,
apparent equilibrium dissociation constant; k
d,

dissociation rate constant; MASP, MBL-associated serine
protease; MBL, mannan-binding lectin; pNPP, para-nitrophenyl phosphate; RAP, receptor-associated protein; SP-A, surfactant protein A;
SP-D, surfactant protein D; SPR, surface plasmon resonance; TTN, Tris ⁄ Tween ⁄ NaCl.
4956 FEBS Journal 277 (2010) 4956–4964 ª 2010 The Authors Journal compilation ª 2010 FEBS
ligand-binding clusters and an 85 kDa beta-chain
involved in endocytosis and comprising the transmem-
brane and intracellular domains. Ligand interaction is
mediated by 31 homologus ligand-binding repeats dis-
tributed unequally between the four ligand-binding
clusters. Each repeat has a bound calcium ion and a
pattern of six cysteine residues forming three intramo-
lecular disulfide bonds [1–3].
In a number of reports, CD91 and calreticulin have
been described as a receptor complex for C1q, mediat-
ing removal of apoptotic cells and immune complexes
[2,4–6]. Other immune-recognition molecules, includ-
ing mannan-binding lectin (MBL) and the lung surfac-
tant proteins surfactant protein A (SP-A) and
surfactant protein D (SP-D), have also been proposed
to enhance phagocytosis through the calreticulin–
CD91 complex. Thus, MBL has been shown to recog-
nize and stimulate the ingestion of apoptotic cells, and
this could be inhibited by antibodies recognizing cal-
reticulin or CD91 [5]. Similarly, initiation of phagocy-
tosis by the collectin members SP-A and SP-D
through the CD91 ⁄ calreticulin pathway could be
inhibited by antibodies recognizing CD91 and calreti-
culin [2;6]. The more recently identified ficolins have
also been proposed to play a role in phagocytosis
[7–9], and L-ficolin and H-ficolin have been found

to interact with calreticulin [7,10]. CD91 and ⁄ or cal-
reticulin may therefore be likely candidates for a
ficolin receptor.
MBL is the recognition molecule of the MBL–lectin
pathway and is a homooligomer composed of 26-kDa
polypeptides. The protomers contain a short N-termi-
nal cysteine-rich domain that is capable of forming
interchain disulfide bonds, a collagen-like region and
a C-terminal globular carbohydrate-recognition
domain. MBL recognizes patterns of neutral carbo-
hydrates, such as mannan, and the binding avidity is
correlated to the degree of oligomerization. Upon
binding to carbohydrate patterns, MBL activates the
complement system, a function that is dependent on its
associated serine proteases, the MBL-associated serine
proteases (MASPs), which initiate the lectin comple-
ment cascade by cleavage of C2 and C4 [11].
The ficolins are recently discovered defence
proteins. Three types have been characterized in
humans: H-ficolin, L-ficolin and M-ficolin. L-ficolin is
a lectin with specificity for N-acetylated and neutral
carbohydrates [8]
.
It is synthesized in the liver and,
like the other ficolins, has the potential of initiating
the MBL ⁄ lectin complement cascade through the
MASPs [12–15].
SP-A and SP-D are lung surfactant proteins. Like
MBL and the ficolins they are pattern-recognition
molecules but do not share the property of binding the

MASPs and therefore do not activate the complement
cascade but function as opsonins. The lung surfactant
proteins share with MBL and the ficolins a unique
domain structure and organization with a collagen tail
linked to a carbohydrate-recognition domain. Each
polypeptide chain associates into homotrimers and
subsequently assembles into oligomers with a sertiform
or cruciform structure containing four to eight trimeric
subunits [16–18].
Previously, we have reported a direct, specific inter-
action between CD91 and C1q. The interaction could
be inhibited by the CD91 chaperone, receptor-associ-
ated protein (RAP), and partially by the C1q-binding
protein calreticulin, where inhibition experiments indi-
cated that half of the binding sites on C1q for CD91
were shared with calreticulin [19].
Here, it is shown that CD91 interacts directly with
MBL and, to a lesser extent, with L-ficolin and there-
fore has the potential to function as an MBL and ⁄ or
an L-ficolin receptor mediating the endocytosis of
apoptotic material. In contrast to previous reports,
CD91 showed direct recognition of both MBL and
L-ficolin in the absence of calreticulin. The interaction
appeared to be specific because it was time-dependent,
concentration-dependent and could be inhibited by
protein ligands of both CD91 and the lectins. The
interaction site of the CD91 molecule on MBL ⁄ L-fico-
lin seemed to be in close proximity to the MASP-asso-
ciation site because inhibition was observed with the
MASPs and reduced binding was seen for MBL point

mutants unable to bind the MASPs.
Results
CD91 binds MBL and L-ficolin directly
We have previously shown that CD91 interacts directly
with C1q. This prompted us to investigate the possible
role of CD91 as a scavenger receptor for other mole-
cules involved in apoptotic cell recognition. For this
purpose, we tested the interaction between CD91 and
several pattern-recognition molecules, namely MBL,
SP-A, SP-D and ficolins, which have previously been
shown to be involved in the phagocytosis of apoptotic
material [7,10,20–22].
When immobilized onto a polystyrene surface, MBL
interacted with soluble biotinylated CD91, to a level
comparable to that observed with C1q (Fig. 1). L-fico-
lin exhibited a lower, but significant, interaction,
whereas SP-A, SP-D, H-ficolin and the control pro-
tein, ovalbumin, showed no detectable binding. Fur-
ther analyses by surface plasmon resonance (SPR)
K. Duus et al. CD91 interacts with mannan-binding lectin
FEBS Journal 277 (2010) 4956–4964 ª 2010 The Authors Journal compilation ª 2010 FEBS 4957
provided no evidence for an interaction between
immobilized CD91 and SP-A, SP-D or H-ficolin (data
not shown).
The interaction between CD91 and MBL was time-
dependent, readily detectable after a few minutes and
reached a plateau after  400 min (Fig. 2A). Binding
was also dependent on the concentration of immobi-
lized MBL (Fig. 2B) and soluble CD91 (Fig. 2C).
Several ligands of CD91 inhibit the interaction

with MBL
Different ligands of CD91 were tested for their ability
to inhibit the interaction between CD91 and MBL.
RAP, a CD91 chaperone known to prevent access of
CD91 ligands [23,24], inhibited the interaction
between CD91 and MBL in a dose-dependent manner
(Fig. 3), whereas no inhibition was observed using the
control protein ovalbumin. We have previously char-
acterized the interaction between CD91 and C1q [19].
C1q was also tested for its ability to inhibit the
CD91–MBL interaction, and addition of this protein
decreased CD91 binding, with nearly complete inhibi-
tion at a 10-fold C1q:CD91 molar ratio (Fig. 3).
Alpha 2-macroglobulin, a known ligand of CD91 and
0
1
2
3
A
0 500 1000 1500 2000
Time (min)
CD91 binding (A405)
Tris, NaCl, Tween
Tris, NaCl, Calcium
B
C
CD91 binding (A405)
Immobilized MBL (nM)
0
1

2
3
158 53 16 6 2 0.6 0.2 0.07 0.02 0.01
0
1
2
3
670 224 75 25 8 3 1 0.3 0.1 0.03
Immobilized CD91 (nM)
MBL binding (A405)
0
20
40
60
80
100
120
C1q MBL L-ficolin H-ficolin SP-D SP-A ovalbumin
CD91 binding (%)
Fig. 1. CD91 binds MBL and L-ficolin. Several molecules known to
enhance phagocytosis were tested for their ability to interact with
CD91. C1q, MBL, L-ficolin, H-ficolin, SP-A, SP-D and ovalbumin
(1 lgÆmL
)1
) as a control were immobilized on a polystyrene surface
and biotinylated CD91 (1 lgÆmL
)1
) was allowed to bind for 2 h.
Bound biotinylated CD91 was quantified using streptavidin-coupled
alkaline phosphatase and pNPP. Results are expressed relative to

C1q binding.
Fig. 2. Interaction between CD91 and MBL is time- and concentra-
tion-dependent. (A) MBL was immobilized on the polystyrene sur-
face and biotinylated CD91 was added for the indicated intervals of
time in 25 m
M Tris ⁄ HCl, 0.5% Tween 20, 150 mM NaCl (pH 7.5) or
25 m
M Tris ⁄ HCl, 150 mM NaCl, 5 mM CaCl
2
(pH 7.5). The level of
bound CD91 was quantified as described in the legend to Fig. 1.
(B) MBL was immobilized at the indicated concentrations and
CD91 (1 lgÆmL
)1
) was allowed to interact and quantified using
streptavidin-coupled alkaline phosphatase and pNPP. (C) CD91 was
immobilized at 1 lgÆmL
)1
. Biotinylated MBL was added at the indi-
cated concentrations, allowed to interact for 1 h, and the amount
of bound MBL was quantified using streptavidin-coupled alkaline
phosphatase and pNPP.
CD91 interacts with mannan-binding lectin K. Duus et al.
4958 FEBS Journal 277 (2010) 4956–4964 ª 2010 The Authors Journal compilation ª 2010 FEBS
MBL, also showed the ability to interfere with the
CD91–MBL interaction (Fig. S1). In addition, we
tested whether MBL ligands could inhibit the interac-
tion. No inhibition was observed using either man-
nose or GlcNAc. The proposed MBL receptor
calreticulin [5] was also tested but only inhibited the

interaction weakly at high calreticulin ⁄ CD91 ratios
(Figs 3 and S2).
Calcium-dependence of the CD91–MBL interaction
CD91 and MBL are both calcium-binding proteins.
CD91 holds 31 Ca
2+
ions that are important for main-
taining its structure and for ligand recognition [25]. As
a C-type lectin, the ligand-binding ability of MBL is
also dependent on calcium [26], in contrast to L-ficolin
which has been shown to recognize GlcNAc in the
absence of calcium ions [27]. We tested the interaction
between CD91 and MBL in the presence or absence of
5mm CaCl
2
, and surprisingly, no difference was
observed (Fig. 2A). To further investigate this ques-
tion, MBL was immobilized and CD91 was allowed to
interact at varying Ca
2+
and Mg
2+
concentrations
and in the presence of 5 mm EDTA. EDTA markedly
decreased the MBL–CD91 interaction but none of the
calcium concentrations tested had a significant effect
(Fig. 4A). The influence of calcium concentration
was also tested by SPR with CD91 immobilized on the
sensor chip surface and MBL in solution. In this
0

1
2
3
1010
C1q
Ovalbumin
Calreticulin
RAP
Molar excess of indicated inhibitor
CD91 binding (A405)
Fig. 3. The CD91–MBL interaction can be inhibited by CD91
ligands. MBL was immobilized on a polystyrene surface and
allowed to interact for 2 h with biotinylated CD91 alone or in the
presence of increasing concentrations of C1q or the control protein
ovalbumin, calreticulin or RAP. The level of bound CD91 was quan-
tified using streptavidin-coupled alkaline phosphatase and pNPP.
0
1
2
3
A
B
C
5 mM
EDTA
0
0.1
0.5
1
2

5
10
CD91 binding (A405)
Mg
2+
Ca
2+
mM Mg
2+
/Ca
2+
0
40
80
120
0 100 300
Time (s)
RU
NaCl/Tris + 2 mM EDTA
NaCl/Tris
NaCl/Tris + 2 m
M Ca
2+
200
0
1
2
3
Binding (A405)
b-MBL

b-ovalbumin
NaCl + Ca
2+
NaCl/Tris NaCl/Tris + mannose
Fig. 4. Calcium dependence of the CD91–MBL interaction. (A)
MBL was immobilized on a polystyrene surface and biotinylated
CD91 was allowed to interact in the presence of 5 m
M EDTA or at
different concentrations of CaCl
2
or MgCl
2
, as indicated. The level
of bound CD91 was quantified using streptavidin-coupled alkaline
phosphatase and pNPP. (B) CD91 (14 600 RU) was immobilized on
the surface of a sensor chip, and MBL was injected in 10 m
M
Tris ⁄ HCl, 150 mM NaCl, 0.005% P20 in the presence or absence of
2m
M EDTA or 2 mM CaCl
2
. (C) CD91 was immobilized on a poly-
styrene surface, and biotinylated MBL or the control protein (bioti-
nylated ovalbumin) was allowed to interact alone (NaCl ⁄ Tris) or in
the presence of either mannose (NaCl ⁄ Tris + 55 n
M mannose) or
CaCl
2
(NaCl ⁄ Tris + 5 mM CaCl
2

). RU, resonance units.
K. Duus et al. CD91 interacts with mannan-binding lectin
FEBS Journal 277 (2010) 4956–4964 ª 2010 The Authors Journal compilation ª 2010 FEBS 4959
configuration, the interaction was found to be highly
sensitive to calcium. Not only was the interaction abol-
ished in the presence of 2 mm EDTA but also in a buf-
fer without calcium ions (Fig. 4B). This may indicate
that immobilization of MBL on a polystyrene surface
restrains MBL in a conformation that, with respect to
CD91 binding, is not sensitive to calcium. This
hypothesis was confirmed by immobilizing CD91 on a
polystyrene surface and measuring the interaction
between MBL and L-ficolin, using ELISA, under dif-
ferent conditions. The interaction was markedly
increased in the presence of Ca
2+
, whereas mannose
(55 nm) had no detectable effect (Fig. 4C). In the case
of L-ficolin, the interaction with CD91 was abolished
in the presence of EDTA, but the presence or absence
of Ca
2+
ions in the buffer had no effect on the interac-
tion (data not shown).
Determination of the kinetic constants
To determine the kinetic constants of the interaction
between CD91 and MBL, increasing concentrations of
recombinant MBL, ranging from 3 to 30 nm, were
injected over a sensor chip on which CD91 has been
immobilized (Fig. 5A). The sensorgrams were properly

fitted using a Langmuir 1 : 1 reaction model, yielding
association rate constant (k
a
) and dissociation rate
constant (k
d
) values of 1.6 · 10
5
m
)1
Æs
)1
and
5.26 · 10
)4
s
)1
, respectively, and a resulting apparent
equilibrium dissociation constant (K
D
) value of 3.3 nm
(v
2
= 3.64).
The reverse configuration, whereby recombinant
MBL was immobilized on a sensor surface and soluble
CD91 was injected over the surface, revealed similar
binding characteristics (Fig. 5B) and kinetics. The
CD91 used contains residual attached RAP, which
influences the concentration of active CD91 and there-

fore the kinetics of the interaction when CD91 is in
solution. The K
D
value obtained was 9.33 nm with
a v
2
= 0.899, and k
d
and k
a
values of 5.74 · 10
)4
S
)1
and 6.15 · 10
4
MÆs
)1
, respectively.
In the case of L-ficolin, the kinetics for interaction
with CD91 was similarly determined by SPR using
plasma-derived L-ficolin concentrations ranging from
120 to 960 nm. This yielded a K
D
of 2.8 · 10
)7
m,
using a 1 : 1 binding model (Fig. 6), which was used
for comparison purposes with MBL; it should be
noted that the v

2
value was higher (6.52), indicating a
poorer fit of the experimental curves to the model than
in the case of MBL (Fig. 6). The K
D
is significantly
higher than that determined in the case of MBL, with
both higher k
a
and k
d
values (k
a
= 1.98 · 10
4
M
)1
Æs
)1
and k
d
= 5.51 · 10
)3
s
)1
).
CD91 interacts with MBL at or near the MASP
binding site
It has previously been shown that calreticulin inter-
acts with MBL at the MASP-binding site [28].

120 240 360 480
0
100
200
Time (s)
RU
3 n
M
30 n
M
6 n
M
9 n
M
12 n
M
18 n
M
24 n
M
300
A
B
150 300 450
0
100
Time (s)
RU
1 n
M

2 n
M
4 n
M
8 n
M
16 n
M
24 n
M
150
50
0
Fig. 5. SPR analysis of the MBL–CD91 interaction. (A) The kinetic
constants were determined by SPR using CD91 as the immobilized
ligand (14 600 RU) and concentrations of soluble MBL ranging from
3to30n
M, as indicated. (B) Similar binding data was obtained with
immobilized recombinant MBL (13 000 RU) and different concentra-
tions (1–24 n
M) of soluble CD91. RU, resonance units.
120 240 360 480
150
0
50
100
Time (s)
RU
120 nM
240 nM

480 nM
720 nM
960 nM
200
Fig. 6. SPR analysis of the L-ficolin–CD91 interaction. Several con-
centrations of plasma-derived L-ficolin, ranging from 120 to 960 n
M,
were injected over a sensor chip on which CD91 was immobilized
(14 600 RU). RU, resonance units.
CD91 interacts with mannan-binding lectin K. Duus et al.
4960 FEBS Journal 277 (2010) 4956–4964 ª 2010 The Authors Journal compilation ª 2010 FEBS
To investigate whether CD91 binds in the same area
within MBL, we tested the ability of MASP-3 to
compete with CD91 for interaction. MBL was pre-
incubated with MASP-3 for 5 min before injection
over a CD91 surface. As shown in Fig. 7A, MASP-3
clearly inhibited the interaction between MBL and
CD91 in a dose-dependent manner. To further inves-
tigate whether interaction with CD91 takes place at
or near the MASP-binding site, the ability of the
MBL mutants K55A and K55Q, which lack the abil-
ity to associate with MASPs, to interact with CD91
was investigated. Compared with wild-type MBL,
both mutants showed strongly reduced binding
(Fig. 7B), providing further support to the hypothesis
that, as previously reported for calreticulin [28],
CD91 binds at or in the vicinity of the MASP-bind-
ing site of MBL.
Discussion
MBL and L-ficolin are both collectins that mediate

activation of the MBL–lectin complement pathway.
These molecules are pattern-recognition molecules of
the innate immune defence and function as opsonins
for the removal of pathogens and apoptotic material
[5,8]. We have previously shown that CD91 directly
recognizes C1q in vitro independently of calreticulin
[19] and here we present data showing that CD91
also directly recognizes MBL and L-ficolin and there-
fore has the potential to function as a phagocytotic
receptor for these proteins. CD91 interacted directly
with MBL and L-ficolin in solid-phase binding
assays, both when CD91 was immobilized and when
the collectins were immobilized. The interactions were
concentration-dependent, time-dependent and satura-
ble. The interaction between CD91 and MBL and
CD91 and L-ficolin could be inhibited by ligands of
both CD91 (RAP, C1q) and MBL (MASP-3) and
therefore fulfilled the requirements of a specific inter-
action. The binding between MBL and CD91 could
be inhibited by EDTA, similarly to the interaction
with other ligands of CD91 [29–31]. No requirement
for additional calcium ions was observed when MBL
was immobilized, whereas additional calcium was
obligatory for the CD91–MBL interaction when
MBL was in solution. This may suggest that calcium
maintains MBL in a conformation necessary for
interacting with CD91. Several concentrations of
L-ficolin and MBL were injected over a surface onto
which CD91 was immobilized to obtain kinetic con-
stants of the interactions. K

D
values of 3.3 and
280 nm were determined for MBL and L-ficolin,
respectively, indicating much higher affinity in the
case of MBL.
CD91 has been suggested to interact with C1q at a
site in close proximity to the C1r ⁄ C1s-binding site [19].
It was therefore of interest to investigate whether
CD91 also interacts with MBL near the site occupied
by the MASPs. MASP-3 induced dose-dependent inhi-
bition of the interaction between CD91 and MBL.
This result was confirmed by using MBL mutants with
a point mutation at residue 55, which rendered them
unable to bind the MASPs [32], showing that the
MBL mutations K55A and K55Q strongly inhibited
interaction with CD91.
Ogden and co-workers have previously reported
that MBL recognizes apoptotic cells and initiates
their phagocytosis through calreticulin and CD91,
with calreticulin acting as the MBL-recognition mole-
cule and CD91 as the internalizing receptor [5]. Here,
we describe direct interaction between CD91 and
MBL without the necessity of recognition by calreti-
culin. Additionally, we describe L-ficolin as a novel
CD91 ligand. As L-ficolin is known as an opsoniza-
tion molecule for apoptotic material, CD91 has the
potential to function as a receptor that mediates the
RU
Time (s)
30 110 190 270

0
100
200
300
MBL
MBL K55Q
MBL K55A
0
120
A
B
200100
0
300
40
80
RU
Time (s)
6 n
M
MBL
6 n
M
MBL +
0.50 n
M
MASP-3
6 n
M
MBL +

1 n
M
MASP-3
0.5 n
M
MBL +
2 n
M
MASP-3
Fig. 7. CD91 interacts with MBL at or near the MASP-binding site.
(A) Using SPR, 6 n
M of MBL was injected over a sensor chip on
which CD91 was immobilized (16 000 RU), with or without preincu-
bation with increasing concentrations of MASP-3, as indicated. (B)
Using SPR, 24 n
M of MBL, the MBL K55A mutant or the MBL
K55Q mutant were injected over a sensor chip on which CD91
was immobilized (14 600 RU). RU, resonance units.
K. Duus et al. CD91 interacts with mannan-binding lectin
FEBS Journal 277 (2010) 4956–4964 ª 2010 The Authors Journal compilation ª 2010 FEBS 4961
phagocytosis of L-ficolin-bound apoptotic cells and
debris.
Materials and methods
Chemicals
p-nitrophenyl-phosphate (pNPP), glycerol, Hepes, sodium
carbonate, dimethylsulfoxide and N-hydroxy-succinimido-
biotin were from Sigma (St Louis, MO, USA). Alkaline
phosphatase-conjugated streptavidin was from Dako
(Glostrup, Denmark). MaxiSorp microtitre plates were
from Nunc (Roskilde, Denmark). CM5 sensorchips, surfac-

tant P20, 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide,
N-hydroxysuccinimide and ethanolamine were from GE
Healthcare (Uppsala, Sweden).
Proteins
Human C1q, alpha 2-macroglobulin, BSA and ovalbumin
were from Sigma. Human placenta-derived CD91 was from
BioMac (Chamalie
`
res, France). RAP was from Innovative
Research (Novi, Michigan, USA). Recombinant MBL was a
generous gift of NatImmune (København Ø, Denmark). The
MBL K55A and K55Q mutants were produced as described
by Teillet et al.[32]. Serum-derived MASP-free L-ficolin and
recombinant MASP-3 were purified as described by Teillet
et al. [32] with a final purification by high-pressure gel-
permeation chromatography on a TSK G-300 SWG column
(Tosohaas, King of Prussia, PA, USA) equilibrated in
145 mm NaCl, 1 mm CaCl
2
and 50 mm triethanol-
amine ⁄ HCl. Recombinant H-ficolin and L-ficolin was puri-
fied as described by Lacroix et al. [33]. Calreticulin was
purified as previously described [34,35]. SP-A was purified
from proteinosis lavage, as described by Madsen et al. [36],
and SP-D was purified from amniotic fluid, as described by
Leth-Larsen et al. [37]. The concentrations of MBL, L-ficolin
and MASP-3 were estimated using absorption coefficients
A1%, 1 cm at 280 nm of 7.8, 17.6 and 13.5, and protomer
molecular mass values of 25 300, 33 800 and 87 500 Da,
respectively [32]. Molar concentrations of CD91, C1q, calret-

iculin, RAP and ovalbumin were estimated using molecular
mass values of 595, 459, 39 and 44 kDa, respectively.
Protein biotinylation
CD91, MBL and ovalbumin were dialysed against
100 mm NaHCO
3
(pH 9.0) at 4 °C, followed by addition
of N-hydroxysuccinimidobiotin in dimethylsulfoxide
(10 mgÆmL
)1
) to a final ratio of 4 mgÆmg
)1
of protein. The
solution was incubated for 2 h at room temperature with
end-over-end agitation, and then dialysed against NaCl ⁄ P
i
(10 mm NaH
2
PO
4
⁄ Na
2
HPO
4
, pH 7.3, 150 mm NaCl) at
4 °C. The biotinylated pr oteins were stored at )20 °C un til use.
ELISA
Unless otherwise stated, incubations and washings were
performed at room temperature on a shaking table using
100 lLÆwell

)1
for incubation and 200 lLÆ well
)1
for washing
and blocking. TTN buffer (25 mm Tris ⁄ 0.5% Tween
20 ⁄ 150 mm NaCl, pH 7.5) was used for blocking, incuba-
tion and washing.
Proteins were coated at 1 lgÆ mL
)1
onto the microtiter
plate surface using 50 mm sodium carbonate (pH 9.6) as
the coating buffer. After coating overnight at 4 °C, plates
were washed three times, for 1 min each wash, followed by
blocking for 30 min. Subsequently, incubation with or with-
out biotinylated protein diluted to 1 lgÆmL
)1
was carried
out for 2 h, followed by another three washes. Finally, the
plates were incubated for 1 h with alkaline phosphatase-
conjugated streptavidin diluted 1:1000. Following another
three washes, the bound biotinylated protein was quantified
using pNPP (1 mgÆmL
)1
)in1m diethanolamine containing
0.5 mm MgCl (pH 9.8). The absorbance was read at
405 nm, with background subtraction at 650 nm, on a
VERSAmax microplate reader (Molecular Devices, Sunny-
vale, CA, USA). All experiments were repeated at least
twice and performed in duplicate. Error bars represent the
standard deviation of the mean from two wells in one

experiment.
SPR experiments
Analyses were performed using BIAcore 3000 or Biacore
X instruments (GE Healthcare). CD91 was diluted to
100 lgÆmL
)1
in 10 mm sodium formate (pH 3), and
immobilized on a CM5 sensor chip (GE Healthcare) in
10 mm Hepes, 150 mm NaCl, 3.4 mm EDTA, 0.005%
surfactant P20 using the amine-coupling chemistry recom-
mended by the manufacturer. Binding of MBL and
ficolins to immobilized CD91 were measured at a flow
rate of 20 lLÆmin
)1
in 10 mm Tris ⁄ HCl, 150 mm NaCl,
2mm CaCl
2
(pH 7.4), containing 0.005% surfactant P20.
Regeneration of the surface was achieved by injection of
10 lLof2m NaCl. As a control surface for background
subtraction, a quenched (activated ⁄ inactivated) surface
was used for plasma-derived proteins, whereas a BSA-
coated surface was used for recombinant proteins. Data
were analyzed by global fitting to a 1 : 1 Langmuir-bind-
ing model of both the association and dissociation phases
for at least five concentrations simultaneously, using the
BIAevaluation 3.2 software (GE Healthcare). The K
D
value was calculated from the ratio of the k
d

and k
a
values.
Acknowledgement
Leif Kongerslev from NatImmune, Copenhagen,
Denmark is thanked for providing recombinant MBL.
CD91 interacts with mannan-binding lectin K. Duus et al.
4962 FEBS Journal 277 (2010) 4956–4964 ª 2010 The Authors Journal compilation ª 2010 FEBS
References
1 Brown MS, Herz J & Goldstein JL (1997) LDL-recep-
tor structure. Calcium cages, acid baths and recycling
receptors. Nature 388, 629–630.
2 Gardai SJ, McPhillips KA, Frasch SC, Janssen WJ,
Starefeldt A, Murphy-Ullrich JE, Bratton DL, Olden-
borg PA, Michalak M & Henson PM (2005) Cell-sur-
face calreticulin initiates clearance of viable or
apoptotic cells through trans-activation of LRP on the
phagocyte. Cell 123, 321–334.
3 Lillis AP, Van DuynLB, Murphy-Ullrich JE & Strick-
land DK (2008) LDL receptor-related protein 1: unique
tissue-specific functions revealed by selective gene
knockout studies. Physiol Rev 88, 887–918.
4 Gardai SJ, Xiao YQ, Dickinson M, Nick JA, Voelker
DR, Greene KE & Henson PM (2003) By binding
SIRPalpha or calreticulin ⁄ CD91, lung collectins act as
dual function surveillance molecules to suppress or
enhance inflammation. Cell 115, 13–23.
5 Ogden CA, deCathelineau A, Hoffmann PR, Bratton
D, Ghebrehiwet B, Fadok VA & Henson PM (2001)
C1q and mannose binding lectin engagement of cell

surface calreticulin and CD91 initiates macropinocyto-
sis and uptake of apoptotic cells. J Exp Med 194, 781–
795.
6 Vandivier RW, Ogden CA, Fadok VA, Hoffmann PR,
Brown KK, Botto M, Walport MJ, Fisher JH, Henson
PM & Greene KE (2002) Role of surfactant proteins A,
D, and C1q in the clearance of apoptotic cells in vivo
and in vitro: calreticulin and CD91 as a common collec-
tin receptor complex. J Immunol 169, 3978–3986.
7 Honore C, Hummelshoj T, Hansen BE, Madsen HO,
Eggleton P & Garred P (2007) The innate immune com-
ponent ficolin 3 (Hakata antigen) mediates the clearance
of late apoptotic cells. Arthritis Rheum 56, 1598–1607.
8 Matsushita M, Endo Y, Taira S, Sato Y, Fujita T,
Ichikawa N, Nakata M & Mizuochi T (1996) A
novel human serum lectin with collagen- and
fibrinogen-like domains that functions as an opsonin.
J Biol Chem 271, 2448–2454.
9 Teh C, Le Y, Lee SH & Lu J (2000) M-ficolin is
expressed on monocytes and is a lectin binding to
N-acetyl-D-glucosamine and mediates monocyte
adhesion and phagocytosis of Escherichia coli.
Immunology 101, 225–232.
10 Kuraya M, Ming Z, Liu X, Matsushita M & Fujita T
(2005) Specific binding of L-ficolin and H-ficolin to
apoptotic cells leads to complement activation. Immuno-
biology 209, 689–697.
11 Ip WK, Takahashi K, Ezekowitz RA & Stuart LM
(2009) Mannose-binding lectin and innate immunity.
Immunol Rev 230, 9–21.

12 Liu Y, Endo Y, Iwaki D, Nakata M, Matsushita M,
Wada I, Inoue K, Munakata M & Fujita T (2005)
Human M-ficolin is a secretory protein that activates
the lectin complement pathway. J Immunol 175, 3150–
3156.
13 Matsushita M, Endo Y & Fujita T (2000) Cutting edge:
complement-activating complex of ficolin and mannose-
binding lectin-associated serine protease. J Immunol
164, 2281–2284.
14 Matsushita M, Kuraya M, Hamasaki N, Tsujimura M,
Shiraki H & Fujita T (2002) Activation of the lectin
complement pathway by H-ficolin (Hakata antigen).
J Immunol 168, 3502–3506.
15 Gal P, Dobo J, Zavodszky P & Sim RB (2009) Early
complement proteases: C1r, C1s and MASPs. A struc-
tural insight into activation and functions. Mol Immunol
46, 2745–2752.
16 Haagsman HP, Hogenkamp A, van EM & Veldhuizen
EJ (2008) Surfactant collectins and innate immunity.
Neonatology 93, 288–294.
17 Chroneos ZC, Sever-Chroneos Z & Shepherd VL (2010)
Pulmonary surfactant: an immunological perspective.
Cell Physiol Biochem 25, 13–26.
18 Wright JR (2005) Immunoregulatory functions of sur-
factant proteins. Nat Rev Immunol 5, 58–68.
19 Duus K, Hansen EW, Tacnet P, Franchet F,
Arlaud GJ, Thielens N & Houen G (2010) Direct
interaction between CD91 and C1q. FEBS J 277, 3526–
3537.
20 Clark H, Palaniyar N, Strong P, Edmondson J,

Hawgood S & Reid KB (2002) Surfactant protein D
reduces alveolar macrophage apoptosis in vivo.
J Immunol 169, 2892–2899.
21 Nauta AJ, Castellano G, Xu W, Woltman AM,
Borrias MC, Daha MR, van Kooten C & Roos A
(2004) Opsonization with C1q and mannose-binding lec-
tin targets apoptotic cells to dendritic cells. J Immunol
173, 3044–3050.
22 Schagat TL, Wofford JA & Wright JR (2001) Surfactant
protein A enhances alveolar macrophage phagocytosis of
apoptotic neutrophils. J Immunol 166, 2727–2733.
23 Bu G (1998) Receptor-associated protein: a specialized
chaperone and antagonist for members of the LDL
receptor gene family. Curr Opin Lipidol 9, 149–155.
24 Striekland DK, Ashcom JD, Williams S, Battey F,
Behre E, McTigue K, Battey JF & Argraves WS (1991)
Primary structure of alpha 2-macroglobulin
receptor-associated protein. Human homologue of a
Heymann nephritis antigen. J Biol Chem 266,
13364–13369.
25 Moestrup SK, Kaltoft K, Sottrup-Jensen L & Gliemann
J (1990) The human alpha 2-macroglobulin receptor
contains high affinity calcium binding sites important
for receptor conformation and ligand recognition.
J Biol Chem 265, 12623–12628.
26 Weis WI, Drickamer K & Hendrickson WA (1992)
Structure of a C-type mannose-binding protein
K. Duus et al. CD91 interacts with mannan-binding lectin
FEBS Journal 277 (2010) 4956–4964 ª 2010 The Authors Journal compilation ª 2010 FEBS 4963
complexed with an oligosaccharide. Nature 360,

127–134.
27 Le Y, Tan SM, Lee SH, Kon OL & Lu J (1997) Purifi-
cation and binding properties of a human ficolin-like
protein. J Immunol Methods 204, 43–49.
28 Pagh R, Duus K, Laursen I, Hansen PR, Mangor J,
Thielens N, Arlaud GJ, Kongerslev L, Hojrup P &
Houen G (2008) The chaperone and potential mannan-
binding lectin (MBL) co-receptor calreticulin interacts
with MBL through the binding site for MBL-associated
serine proteases. FEBS J 275, 515–526.
29 Fass D, Blacklow S, Kim PS & Berger JM (1997)
Molecular basis of familial hypercholesterolaemia
from structure of LDL receptor module. Nature 388,
691–693.
30 Huang W, Dolmer K & Gettins PG (1999) NMR
solution structure of complement-like repeat CR8 from
the low density lipoprotein receptor-related protein.
J Biol Chem 274, 14130–14136.
31 van Driel IR, Goldstein JL, Sudhof TC & Brown MS
(1987) First cysteine-rich repeat in ligand-binding
domain of low density lipoprotein receptor binds Ca2+
and monoclonal antibodies, but not lipoproteins. J Biol
Chem 262, 17443–17449.
32 Teillet F, Lacroix M, Thiel S, Weilguny D, Agger T,
Arlaud GJ & Thielens NM (2007) Identification of
the site of human mannan-binding lectin involved
in the interaction with its partner serine proteases:
the essential role of Lys55. J Immunol 178, 5710–
5716.
33 Lacroix M, Dumestre-Perard C, Schoehn G, Houen G,

Cesbron JY, Arlaud GJ & Thielens NM (2009)
Residue Lys57 in the collagen-like region of human
L-ficolin and its counterpart Lys47 in H-ficolin play a
key role in the interaction with the mannan-binding
lectin-associated serine proteases and the collectin recep-
tor calreticulin. J Immunol 182, 456–465.
34 Duus K, Pagh RT, Holmskov U, Hojrup P, Skov S &
Houen G (2007) Interaction of calreticulin with CD40
ligand, TRAIL and Fas ligand. Scand J Immunol 66,
501–507.
35 Hojrup P, Roepstorff P & Houen G (2001) Human
placental calreticulin characterization of domain
structure and post-translational modifications. Eur
J Biochem 268, 2558–2565.
36 Madsen J, Tornoe I, Nielsen O, Koch C, Steinhilber W
& Holmskov U (2003) Expression and localization of
lung surfactant protein A in human tissues. Am J Respir
Cell Mol Biol 29, 591–597.
37 Leth-Larsen R, Holmskov U & Hojrup P (1999) Struc-
tural characterization of human and bovine lung surfac-
tant protein D. Biochem J 343 (Pt 3), 645–652.
Supporting information
The following supplementary material is available:
Fig. S1. Alpha 2-macroglobulin and MBL binds over-
lapping sites on CD91.
Fig. S2. Calreticulin has little inhibitory effect on the
CD91-MBL interaction.
This supplementary material can be found in the
online version of this article.
Please note: As a service to our authors and readers,

this journal provides supporting information supplied
by the authors. Such materials are peer-reviewed and
may be re-organized for online delivery, but are not
copy-edited or typeset. Technical support issues arising
from supporting information (other than missing files)
should be addressed to the authors.
CD91 interacts with mannan-binding lectin K. Duus et al.
4964 FEBS Journal 277 (2010) 4956–4964 ª 2010 The Authors Journal compilation ª 2010 FEBS