Báo cáo khoa học: The chaperone and potential mannan-binding lectin (MBL) co-receptor calreticulin interacts with MBL through the binding site for MBL-associated serine proteases pdf

co-receptor calreticulin interacts with MBL through the binding site for MBL-associated serine proteases Rasmus Pagh1, Karen Duus1, Inga Laursen2, Paul R.. Calreticulin showed saturable

co-receptor calreticulin interacts with MBL through the binding site for MBL-associated serine proteases

Rasmus Pagh1, Karen Duus1, Inga Laursen2, Paul R Hansen3, Julie Mangor2, Nicole Thielens4, Ge´rard J Arlaud4, Leif Kongerslev5, Peter Højrup6and Gunnar Houen1

1 Department of Autoimmunology, Statens Serum Institut, Copenhagen, Denmark

2 Department of Clinical Biochemistry, Statens Serum Institut, Copenhagen, Denmark

3 Department of Natural Sciences, Faculty of Life Sciences, University of Copenhagen, Frederiksberg, Denmark

4 Laboratoire d’Enzymologie Mole´culaire, Institut de Biologie Structurale Jean-Pierre Ebel, Grenoble, France

5 NatImmune, Copenhagen, Denmark

6 Institute of Biochemistry and Molecular Biology, University of Southern Denmark, Odense, Denmark

Mannan-binding lectin (MBL) is an important

compo-nent of the mammalian innate immune system and a

member of the collectin family, which, among others,

also includes lung surfactant proteins A and D [1–6]

MBL is a homopolymer composed of 26-kDa

polypep-tides The protomers contain a short N-terminal

cyste-ine-rich domain, capable of forming inter-chain

disulfide bonds, a collagen-like region and a

C-termi-nal globular carbohydrate recognition domain (CRD) These associate as homotrimeric subunits by formation

of collagen-like triple-helical fibers for subsequent assembly into higher-order oligomers containing up to six subunits [7–12] In the mature MBL oligomer, the CRD is separated from the collagen-like triple-helical domain by a short coiled-coil sequence, called the neck region MBL recognizes patterns of neutral

Keywords

calreticulin; chaperone; collectin;

mannan-binding lectin; serine protease

Correspondence

G Houen, Department of Autoimmunology,

Statens Serum Institut, Artillerivej 5,

DK-2300 Copenhagen, Denmark

Fax: +45 32683149

Tel: +45 32683276

E-mail: gh@ssi.dk

(Received 7 August 2007, revised 19

October 2007, accepted 3 December 2007)

doi:10.1111/j.1742-4658.2007.06218.x

The chaperone calreticulin has been suggested to function as a C1q and collectin receptor The interaction of calreticulin with mannan-binding lectin (MBL) was investigated by solid-phase binding assays Calreticulin showed saturable and time-dependent binding to recombinant MBL, pro-vided that MBL was immobilized on a solid surface or bound to mannan

on a surface The binding was non-covalent and biphasic with an initial salt-sensitive phase followed by a more stable salt-insensitive interaction For plasma-derived MBL, known to be complexed with MBL-associated serine proteases (MASPs), no binding was observed Interaction of calreti-culin with recombinant MBL was fully inhibited by recombinant MASP-2, MASP-3 and MAp19, but not by the MASP-2 D105G and MAp19 Y59A variants characterized by defective MBL binding ability Furthermore, MBL point mutants with impaired MASP binding showed no interaction with calreticulin Comparative analysis of MBL with complement compo-nent C1q, its counterpart of the classical pathway, revealed that they display similar binding characteristics for calreticulin, providing further indication that calreticulin is a common co-receptor/chaperone for both proteins In conclusion, the potential MBL co-receptor calreticulin binds to MBL at the MASP binding site and the interaction may involve a confor-mational change in MBL

Abbreviations

AP, alkaline phosphatase; CRD, carbohydrate recognition domain; pNPP, para-nitrophenyl phosphate; MAp19, MBL-associated protein of

19 kDa; MASP, MBL-associated serine protease; pMBL, plasma-derived MBL; rMBL, recombinant MBL; TTN, Tris-Tween-NaCl.

carbohydrates on the surface of micro-organisms, and

the binding avidity is correlated to the degree of

oligo-merization [1,3–6]

Upon binding to carbohydrate patterns, MBL

acti-vates the complement system The

complement-activat-ing function of MBL is dependent on its associated

serine proteases, the mannan-binding lectin-associated

serine proteases (MASPs), of which three forms

(MASP-1, MASP-2 and MASP-3) have been described,

together with a truncated form of MASP-2, named

MBL-associated protein of 19 kDa (MAp19) [11,13–

22] The MASPs form homodimers, which associate

with oligomeric MBL in a ratio of one MASP dimer

per MBL oligomer [9,23]

Several natural or site-directed MBL mutations

affecting MASP association and/or biological activity

have been described [24–31] These affect the

oligomer-ization of the protein, indirectly affecting the

associa-tion with the MASPs, or directly affecting the binding

site for the MASPs, which has been localized to the

C-terminal part of the collagen-like region MASP-1,

MASP-2 and MASP-3 have overlapping, but not

iden-tical binding sites [27,29,31]

C1q, the recognition molecule of the first

comple-ment component (C1) shows structural and functional

homology to MBL in many respects C1q is a hexamer

of heterotrimers, composed of homologous polypeptide

chains A, B and C These associate as N-terminal

disulfide-linked A–B and C–C dimers, which

subse-quently oligomerize into two heterotrimeric oligomers,

composed of two A–B dimers and one C–C dimer

Three sets of two heterotrimers assemble to form the

mature C1q hexamer, which in turn associates with a

tetrameric complex formed of two molecules each of

the serine proteases C1r and C1s [32,33]

The function of C1q is similar to that of the

collec-tins, and the role of these molecules in the immune

system relies on their ability to bind to repeating

patterns of certain carbohydrate residues and other

components on the surface of micro-organisms and

apoptotic cells, as well as to antigen-bound

immuno-globulins C1q recognizes IgG and IgM, bound to the

surface of invading pathogens, as well as blebs on the

surface of apoptotic cells, and MBL binds to

patho-gens and apoptotic cells [4,32–38] and changes

confor-mation upon binding [39] Target recognition activates

the associated proteases (MASPs or C1r/C1s), which

subsequently activate the complement system by

cleav-ing C4 and C2 to form the C3-convertase This leads

to the deposition of C3b on the target cell, formation

of the membrane attack complex and release of

ana-phylatoxins, thus killing pathogens and opsonizing

them for phagocytosis

Several receptors are involved in opsonization and phagocytosis (e.g the C3b receptor) Receptors for MBL and C1q are also assumed to play a role in opso-nization and clearance and have been the subject of intensive research Several candidate receptors have been suggested, including megalin, CD91 (a2 -macro-globulin receptor), CD35, CD93, gC1qR (hyaluronic acid binding protein) and cC1qR (calreticulin) [32,35,40–44]

Calreticulin is an abundant chaperone in the endo-plasmic reticulum, where it functions as a Ca2+ stor-age protein and a key component in the folding and quality control of glycoproteins and other specific pro-teins [45,46] Furthermore, it participates in the peptide loading of the major histocompatibility complex class I, for presentation on the surface of antigen-pre-senting cells [47] Calreticulin has also been reported to

be present at the surface of various cell types, in com-plex with cell surface receptors such as the general scavenger receptor CD91 The calreticulin/CD91 com-plex was shown to be present on the surface of phago-cytic cells and to function as a scavenger receptor complex for apoptotic cells and micro-organisms [48– 52] Thus, the calreticulin/CD91 complex has been sug-gested to recognize C1q and collectins bound to apop-totic target cells, and the interaction between C1q and calreticulin was shown to require a conformational change in C1q, such as that occurring upon binding

to aggregated immunoglobulins or to a hydrophobic polystyrene surface [53] To characterize the interaction

of calreticulin with MBL, we investigated the binding

of calreticulin to plasma-derived MBL (pMBL) and recombinant MBL (rMBL) under various conditions

Results

The interaction of calreticulin with immobilized rMBL was studied using multi-well format solid-phase assays and showed the same characteristics as observed for its binding to immobilized C1q These included: (a) a time- and concentration-dependent saturable binding under conditions comprising a physiological salt con-centration and a relatively high detergent concentra-tion (25 mm Tris, 0.15 m NaCl, 0.5% Tween 20, pH 7.5), to avoid non-specific binding (Fig 1A) and (b) an initial salt-sensitive binding with maximal interaction

at physiological ionic strength, which is gradually changed to a salt-insensitive binding during interaction (Fig 1B) The binding could be disrupted by exposure

to high concentrations of urea (8 m) or SDS (10%) (results not shown), indicating that the interaction was based on non-covalent forces Binding experiments between calreticulin and MBL were performed both in

the presence and absence of Ca2+ ions (0–5 mm) as

well as in the presence of EDTA (5 mm), and no major

difference was observed except for a small stimulating

effect of 0.5–1 mm Ca2+ (Fig 2) A complication

related to these experiments was that Ca2+ was not

compatible with 0.5% Tween 20, and experiments with

Ca2+had to be conducted in the absence of detergent

Nevertheless, provided that the wells were preblocked

with Tris-Tween-NaCl (TTN) buffer, the omission of

Tween 20 only resulted in a minor increase in

back-ground signal Consequently, many control

experi-ments were carried out in both TTN buffer without

the addition of extra Ca2+(assuming that enough

cal-cium was naturally present to allow Ca2+-dependent

reactions to take place), in TTN buffer with EDTA added to test whether Ca2+was a limiting factor, and

in TN buffer (25 mm Tris, 0.15 m NaCl, pH 7.5) with

Ca2+ added Control experiments with non-coated wells and wells coated with control proteins (ovalbu-min, lysozyme or BSA) ruled out non-specific inter-actions between calreticulin and the solid phase (Fig 1A) In additional control experiments, BSA was used instead of Tween 20 as a blocking agent to reveal similar low non-specific binding of biotin-labelled calreticulin to non-coated wells, and binding between calreticulin and rMBL was also demonstrated using non-biotinylated calreticulin and antibodies recogniz-ing the C-terminus of calreticulin (results not shown) This ruled out the possibility that the binding was an artefact caused by biotinylation of calreticulin The calreticulin used to demonstrate binding was mono-meric but binding of oligomono-meric calreticulin to rMBL could also be observed (results not shown)

Preparations of rMBL and pMBL were analysed

by size-exclusion chromatography and showed nearly identical elution profiles, as measured by absorbance

at 280 nm (Fig 3) However, rMBL eluted slightly ear-lier from the column than pMBL SDS/PAGE analysis

of the fractions collected from the size-exclusion chro-matography revealed that rMBL contained somewhat higher oligomeric forms than pMBL when analyzed under non-reducing conditions, whereas only pMBL contained associated MASPs (appearing as a band of

70 kDa under reducing conditions), in agreement with the different origins and modes of production of these preparations (Fig 4) The comparison of pMBL and rMBL, with respect to oligomerization, is not straight-forward because pMBL originates from a pool of

0

1

2

3

Time (min)

C1q rMBL Control

0

1

2

3

150 100

50 0

150 100

50 0

C1q rMBL

Time before addition of salt (min)

A

B

Fig 1 (A) Comparative time-dependent binding of calreticulin to

immobilized rMBL and C1q For coating, rMBL and C1q were

diluted to a final concentration of 1 lgÆmL)1in carbonate buffer, pH

9.6, and the plate was incubated with shaking for 24 h, with

100 lL per well, at 4 C Control wells only received coating buffer

(negative control = background) Wells were then washed for

3 · 1 min and blocked for 1 h in TTN buffer (25 m M Tris, 0.15 M

NaCl, 0.5% Tween 20, pH 7.5) Biotin-labelled calreticulin

(0.33 lgÆmL)1) diluted in TTN was added and incubation was

con-tinued at room temperature for the indicated periods followed by

incubation with AP-labelled streptavidin The results are presented

as the mean ± SD of duplicate absorbance readings at 405 nm (B)

Time-dependent salt-sensitivity of the interaction of calreticulin with

rMBL and C1q rMBL and C1q were diluted to a final concentration

of 1 lgÆmL)1 in carbonate buffer pH 9.6 Wells were coated

and washed as described above followed by incubation with

0.33 lgÆmL)1biotin-labelled calreticulin diluted in TTN for different

time intervals, prior to the addition of 0.5 M NaCl to the wells.

EDTA (m M )

CaCl2 (m M )

0 1 2 3

Fig 2 Influence of calcium ions and EDTA on MBL calreticulin interaction Biotin-labelled calreticulin was incubated in rMBL-coated plates The interaction took place in incubation buffer (25 m M Tris, 0.15 M NaCl, pH 7.5) with addition of 0–5 m M of CaCl2

or 5 m M EDTA The interaction was quantified by incubation with AP-conjugated streptavidin and pNPP.

human plasma [54], and rMBL was produced using a

human embryonic kidney cell expression system [8]

However, when analyzing the MBL-containing

frac-tions for their ability to bind calreticulin, immobilized

rMBL from all fractions showed calreticulin binding,

whereas none of the pMBL fractions showed

detect-able binding (Fig 3)

Binding to calreticulin was also observed for rMBL

bound to immobilized mannan (Fig 5) By contrast,

although binding of pMBL to mannan is known to

activate the associated MASPs through a

conforma-tional change, no binding was observed to pMBL

immobilized in mannan-coated wells Binding to

pMBL was, however, observed after size-exclusion

chromatography at pH 5, conditions reported to cause

dissociation of the MASPs from MBL [55]

Neverthe-less, we did not obtain complete dissociation of the

bound MASPs (results not shown)

These results indicate that calreticulin is able to bind

directly to immobilized rMBL or to mannan-bound

rMBL through an initial ionic interaction, which,

pos-sibly through a conformational change in calreticulin,

gradually develops into a binding of higher strength,

presumably involving hydrogen bonds and

hydropho-bic interactions The results also suggest that

calreticu-lin may interact with MBL through the MASP binding

site because no significant binding was observed to

pMBL with associated MASPs, neither after direct

immobilization or after binding to mannan, indepen-dently of the degree of oligomerization (Fig 3B) This hypothesis was further investigated by performing vari-ous inhibition and binding assays Binding of calreticu-lin to rMBL could be inhibited by co-incubation with recombinant MASP-2, whereas a MASP-2 variant (D105G), defective in MBL binding ability [56], showed a decreased inhibitory activity (Fig 6A) When the immobilized rMBL was first pre-incubated with MASP-2 in the presence of calcium ions, complete inhibition was observed (Fig 6B) Calreticulin binding was also strongly inhibited by co-incubation with recombinant MASP-3 (Fig 6C) and the inhibitory effi-ciency increased as a function of the MASP-3 concen-tration used (Fig 6D) MAp19 was also inhibitory, whereas the Y59A MAp19 mutant, characterized by a reduced MBL binding activity [57] showed a signifi-cantly decreased inhibitory potential (Fig 6C) In line with these data, two MBL point mutants (K55A and K55E) with defective MASP-binding capability [31] showed no detectable interaction with calreticulin (Fig 6E)

Further experiments were conducted using a synthetic peptide, GLRGLQGPOGKLGPOG-NH2 (where O = hydroxyproline), spanning the putative MASP-binding region of MBL [29] As shown in Fig 7, this peptide was found to inhibit interaction of calreticulin with MBL to an extent of approximately 50% The binding of calreticulin to MBL as well as to C1q was also shown to be inhibited by fucoidan, a sul-fated polysaccharide known to bind C1q [58] (Fig 7) Monoclonal antibodies raised against pMBL and spe-cific for the CRD of MBL (Hyb 131-1) or its triple-helical collagen-like region (Hybs 131-10, 131-11), were also tested for their ability to inhibit the interaction between calreticulin and rMBL, and did not reveal any significant effect (results not shown), indicating that they bind to sites not involved in calreticulin binding,

in agreement with their ability to bind pMBL with associated MASPs

Taken together, these results indicate that the MASPs must dissociate from MBL to allow binding to calreticulin and that conformational changes may take place in MBL (e.g during ligand binding or immobili-zation) In support of this hypothesis, analysis of the interaction of immobilized calreticulin with soluble rMBL showed no binding either in the absence or presence of soluble mannan (results not shown) Using surface plasmon resonance analysis, calreticulin bound immobilized MBL with high on and off rates, indicat-ing that, in the absence of a conformational change in MBL, only the initial ionic interaction could occur (data not shown) Similarly, C1q did not bind to

0 2 4 6 8 10 12 14

0

1

2

A

B

30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48

(mL)

0 2 4 6 8 10 12 14

0

1

2

30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48

A 405 nm

A 405 nm

(mL)

Fig 3 Elution profiles from size-exclusion chromatography of (A)

rMBL and (B) pMBL Hatched bars represent results from ELISA

analysis of the collected fractions for binding of biotin-labelled

cal-reticulin.

immobilized calreticulin, unless it was in complex with IgG as previously described [53]

Discussion

The results obtained in the present study demonstrate that calreticulin exhibits strong binding to rMBL with the following characteristics: (a) a fast, saturable, and salt-sensitive binding phase; (b) a slower binding phase that is resistant to high salt concentrations, but sensi-tive to 8 m urea and 10% SDS; (c) the interaction is inhibited in the presence of MASP-2, MASP-3 and MAp19, but not by mutant forms of MASP-2 and MAp19 with defective MBL binding abilities; (d) the interaction between calreticulin and rMBL may require conformational changes in MBL, which can be achieved by immobilization on a polystyrene surface

or through binding to a natural immobilized ligand such as mannan; (e) binding of calreticulin is inhibited

0

1

2

3

Coating agent:

1 layer:

2 layer:

3 layer:

Mannan

rMBL

b-calreticulin

AP-strep.

Mannan pMBL b-calreticulin AP-strep.

Mannan – b-calreticulin AP-strep.

rMBL – b-calreticulin AP-strep.

pMBL – b-calreticulin AP-strep.

Fig 5 Interaction of calreticulin with MBL bound to immobilized

mannan Wells were coated as indicated with rMBL and pMBL

(1 lgÆmL)1), or with mannan (1 mgÆmL)1) followed by incubation

with rMBL and pMBL Subsequently, wells were incubated with

biotin-labelled calreticulin (0.33 lgÆmL)1) in TTN followed by

incuba-tion with AP-conjugated streptavidin Results are presented as the

mean ± SD of duplicate absorbance readings at 405 nm.

*

Fraction number

kDa

250

150 100 75

60 37 25 20 15

kDa

250

100

75 60 37 25 20 15

kDa

250

150

100 75 60

25 20

37

kDa

250 150

100 75 60 37 25 20 15

Fraction number

Fig 4 SDS/PAGE analysis of peak fractions

from size exclusion chromatography of

rMBL (A1–A2) and pMBL (B1–B2) as shown

in Fig 2 (A1, B1) SDS/PAGE under reducing

conditions (A2, B2) Non-reducing SDS/

PAGE Gels (4–12%) were stained with

Coomassie Brilliant Blue *MASP-derived

bands.

by a short synthetic peptide mapping to the MASP-2

binding site on MBL; and (f) calreticulin does not bind

MBL point mutants with defective MASP interaction

Steinø et al [53] showed that calreticulin interacts

strongly with immobilized C1q, whereas pMBL

(asso-ciated with the MASPs) only exhibits a low level of

binding to calreticulin after prolonged heating at

57C In the present study, we provide experimental evidence that calreticulin can interact with MBL in a way similar to C1q, provided that no MASP is associ-ated The finding that inhibition of the MBL-calreticu-lin interaction was achieved with rMASP-2, rMASP-3 and rMAp19, but not with the D105G variant of rMASP-2 and the Y59A variant of rMAp19, is consis-tent with the fact that the variants lack the ability to associate with rMBL [56,57] In the same way, the fact that no binding was observed with pMBL is fully con-sistent with the latter being associated with MASP-1, MASP-2, MASP-3 and MAp19 [54] The most likely hypothesis, therefore, is that any associated MASP and MAp19 will sterically prevent binding to calreticu-lin However, it cannot be excluded that these may also bring about constraints preventing conformational changes necessary for calreticulin binding Taken together, the above observations, together with the observation that MBL point mutants with impaired ability to associate with the MASPs do not interact with calreticulin, provide strong experimental support for the hypothesis that calreticulin binds to the MASP binding site of MBL

0

1

2

3

Positive

control

rMASP-2 (D105G) mutant rMASP-2

75 50 150

1

A

B

C

D

E

2 3

0

1

2

3

Inhibitor

0

1

2

Negative

control

(Ovalbumin)

(Y59A) Positive control (MBL alone)

150 75 50

25 37

15

0

1

2

Excess of MASP-3/MAp19

MASP-3 MAp19

0

1

2

Positive control

(rMBL)

rMBL (K55E) rMBL (K55A) Negative control

(Ovalbumin)

150 75 50

25 37

Fig 6 (A) Inhibition of calreticulin binding to rMBL by wild-type and mutant (D105G) MASP-2 Wells were coated at 4 C for 24 h with

100 lL of rMBL (1 lgÆmL)1in carbonate buffer, pH 9.6) The wells were then washed for 3 · 1 min in TTN and incubated with 100 lL

of supernatants from either non-transfected cells (positive control), HEK293 cells containing wild-type MASP-2 or the D105G mutant, together with the addition of 1 lgÆmL)1of biotinylated calreticulin, thereby obtaining a 100-fold molar excess of the MASPs Control experiments with anti-MASP-2 and anti-MBL sera confirmed the presence of rMBL and MASP-2, respectively, in the wells (not shown) The results are presented as the mean ± SD of duplicate absorbance readings at 405 nm The presence and integrity of MASP-2 in the used supernatant were confirmed by immunoblot: lane 1, MASP-2; lane 2, rMASP-2 D105G; lane 3, control superna-tant (B) Inhibition of calreticulin binding to rMBL by preincubation with rMASP-2 in the presence of 5 m M Ca 2+ Immobilized rMBL was pre-incubated with rMASP-2 (90 l M ) for 24 h, and then calreti-culin was added in Tris buffer containing 5 m M of Ca2+ (C) Inhibi-tion of calreticulin binding to rMBL by purified rMASP-3 (20 l M ), wild-type rMAp19, and the Y59A MAp19 mutant (80 l M ) To the right, the purity of the recombinant proteins was verified by SDS/ PAGE, stained with GelCODE blue stain: lane 1, rMAp19; lane 2, rMASP-3; lane 3, rMASP3 Y59A (D) Concentration-dependent inhi-bition of rMASP-3 and rMAp19 inhiinhi-bition of calreticulin binding to rMBL Calreticulin and MASP-3 or Map19 were co-incubated at the indicated ratio (w : w) over calreticulin on microtitre plates coated with rMBL (E) Binding of calreticulin to rMBL and two mutant rMBL forms (K55A, K55E), each coated at 1 lgÆmL)1 To the right, MBL purity is shown by SDS/PAGE with silver-staining: lane 1, rMBL K55E; lane 2, rMBL K55A; lane 3, rMBL; lane 4, pMBL.

The results of the present study also suggest that a

conformational change may take place in calreticulin

upon binding to immobilized MBL, resulting in a

non-covalent biphasic binding in terms of salt sensitivity

Although, it cannot be ruled out that the

immobiliza-tion of MBL may simply increase the number of

bind-ing sites or that further conformational changes may

also occur in MBL, these characteristics are strikingly

similar to those reported for the interaction between

calreticulin and C1q [53] Conformational changes in

calreticulin have previously been reported to occur in

conjunction with Ca2+ deprivation or removal of the

C-domain, and these changes induced a

polypeptide-receptive state of calreticulin [59,60]

Calreticulin is a multi-functional chaperone which

has been shown to possess Ca2+ binding, lectin-like

and polypeptide binding properties [61–67]

Calreticu-lin has been reported to be a candidate co-receptor for

the collectins and C1q and to be present on cell

sur-faces in complex with CD91 [48–52] This implies that

calreticulin is capable of associating with CD91 using

one site, and interacting with the collectins or C1q

through another site To determine which part of

cal-reticulin participates in the interaction with MBL, we

performed preliminary inhibition studies with

recombi-nant calreticulin N- and P-domains, which both

showed some inhibitory activity (results not shown) Based on this observation, it may be anticipated that the C-domain could be involved in binding to CD91, whereas the N- and P-domains are responsible for interaction with the collectins and C1q Alternatively, calreticulin may bind to CD91 through a site of the N-domain not involved in the interaction with C1q and the collectins

The physiological relevance of the interaction of cal-reticulin with MBL and C1q cannot be deduced from the results obtained in the present study However, it

is generally accepted that the MASPs are activated upon binding of MBL to its targets and that this initi-ates activation of the complement cascade, leading to target lysis and/or opsonization Inactivation of the MASPs to control this reaction can be achieved by binding to serum protease inhibitors, notably C1-inhib-itor and a2-macroglobulin [68] In this process, the protease inhibitors themselves change conformation and, in the case of a2-macroglobulin, a binding site for CD91 is exposed Upon binding of the protease inhibi-tors, the MASPs may be released from MBL and the

a2-macroglobulin/MASP complex may still possess the ability to bind to CD91 However, the target-bound MBL may bind to CD91 as well, provided that it asso-ciates with calreticulin, which may occur on the cell surface in complex with CD91 [48–52] The obvious advantage of this process is that the target would be opsonized for binding to CD91, whether or not the

a2-macroglobulin/MASP complex remains bound to MBL or dissociates In general, it may be anticipated that the process of infectious target/apoptotic cell rec-ognition depends on multiple factors and ligands, and that it has an inherent redundancy, in order to achieve maximal specificity and safety in self/non-self discrimi-nation Thus, the MBL/MASP/a2-macroglobulin/ calreticulin/CD91 system only constitutes a part of the phagocytic scavenging system

In conclusion, the potential MBL co-receptor/chap-erone calreticulin interacts with MBL at its MASP-binding site The interaction of calreticulin with MBL

is similar to that observed for C1q, indicating that pathogenic targets, activating the lectin or classical complement pathways, might be eliminated through interaction with the calreticulin/CD91 complex

Experimental procedures

Reagents Amino acids, ovalbumin, p-nitrophenyl-phosphate (pNPP) substrate tablets, 5-Br-4-Cl-3-indolylphosphate/nitrobluetet-razolium substrate tablets, urea, dimethylsulfoxide, glycerol,

0

50

100

Coating agent:

Inhibitor: GLRGLQGPOGKLGPOG Fucoidan

2.layer:

3.layer

rMBL

b-calreticulin

AP-strep.

C1q

b-calreticulin AP-strep.

rMBL

b-calreticulin AP-strep.

C1q

b-calreticulin AP-strep.

Fig 7 Inhibition ELISA performed with a synthetic peptide and the

algal polysaccharide fucoidan The inhibitory effects of the synthetic

peptide GLRGLQGPOGKLGPOG-NH2 (where O = hydroxyproline)

and fucoidan added in a 300-fold (w : w) excess over calreticulin

were assessed rMBL and C1q were diluted to a final concentration

of 1 lgÆmL)1in carbonate buffer, pH 9.6, and wells were incubated

under shaking for 24 h with 100 lL, at 4 C, followed by washing

and blocking for 1 h in TTN The inhibitors dissolved in

dimethylsulf-oxide (10 mgÆmL)1) were diluted 1 : 100 in TTN and added together

with biotin-labelled calreticulin (0.33 lgÆmL)1) Subsequently, wells

were incubated with AP-labelled streptavidin and developed with

pNPP Results are presented as the mean ± SD of duplicate

absor-bance readings at 405 nm.

dithiothreitol, sodium carbonate, Tris, Tris-hydrochloride,

N-hydroxy-succinimidobiotin, BSA, hemoglobin, lysozyme,

C1q, rabbit C1q antiserum and fucoidan from Fucus

vesicu-losus were obtained from Sigma (St Louis, MO, USA)

Acetonitrile, N,N-dimethyl formamide, MgCl2 and

Tween 20 were obtained from Merck (Darmstadt,

Ger-many) NaCl was from Unikem (Copenhagen, Denmark)

Alkaline phosphatase (AP)-conjugated streptavidin was

from DakoCytomation (Glostrup, Denmark) MaxiSorp

microtitre plates were from Nunc (Roskilde, Denmark)

Q-Sepharose, Superose 6 and Sephacryl S-100 HR were

from Amersham Biosciences/GE Healthcare (Uppsala,

Swe-den) Recombinant MBL was from NatImmune

(Copen-hagen, Denmark) NaCl and Na2HPO4Æ2H2O were from

Unikem A/S (Copenhagen, Denmark) Monoclonal MBL

antibodies, pMBL, purified as described previously [69],

and antiserum against the calreticulin C-terminus (peptide

CEDVPGQAKDEL conjugated to ovalbumin [70]) were

from Statens Serum Institut (Copenhagen, Denmark)

Tris-glycine gels were from Invitrogen (Carlsbad, CA, USA)

Pre-stained molecular weight markers for SDS/PAGE were

from Bio-Rad Laboratories (Hercules, CA, USA)

Gel-CODE blue stain reagent was from Pierce (Rockford, IL,

USA) Cell culture supernatants containing recombinant

MASP-2 and the D105G MASP-2 mutant, as well as the

purified K55A and K55E rMBL variants [31], were

gener-ous gifts from S Thiel (University of Aarhus, Denmark)

Purified MASP-3, MAp19 and the Y59A MAp19 variant

were produced at the Institut de Biologie Structurale

Jean-Pierre Ebel, Grenoble, France, as described previously

[57,71]

Purification of human placenta calreticulin

Human placenta calreticulin was purified using a slight

modification of a well established procedure [72,73] In

brief, a placenta was homogenized in 20 mm bis-Tris, pH

7.2 and centrifuged, followed by homogenization of the

precipitate in the same buffer with the addition of 1%

Tri-ton X-114 A separation of water and detergent phases of

the last two supernatants was induced by addition of

Triton X-114 to 2% and incubation at 37C Ammonium

sulfate (337 gÆL)1) was added to the water phase and the

precipitated proteins removed by centrifugation The

super-natant was then ultradiafiltered and chromatographed on a

Q-Sepharose ion-exchange column Eluted calreticulin was

further purified by size-exclusion chromatography on a

Sephacryl S-100 HR column The purified protein showed a

single band of apparent molecular mass 60 kDa by SDS/

PAGE

Biotinylation of calreticulin

The purified calreticulin was dialysed against 0.1 m

NaHCO3, pH 9.0, at 4C, followed by addition of

N-hydroxysuccinimidobiotin in N,N-dimethyl formamide (10 mgÆmL)1) to a final concentration of 4 mgÆmg)1 calreti-culin The solution was incubated for 2 h at room tempera-ture with end-over-end agitation, and then dialysed against NaCl/Pi (0.15 m NaCl, 10 mm NaH2PO4/Na2HPO4, pH 7.3) at 4C The biotinylated calreticulin was mixed with

an equal volume of glycerol and stored at)20 C until use

Chromatography of MBL Recombinant or plasma-derived human MBL (0.3 mg, 1.5 mgÆmL)1) was applied on a column (diameter: 1.6 cm) packed with 70 mL of Superose 6 and equilibrated with NaCl/Pi, pH 7.3 The column was connected to an A¨kta Explorer system (Amersham Biosciences/GE Healthcare) and eluted at a flow rate of 0.5 mLÆmin)1using NaCl/Pi as the buffer The eluted peaks were collected as fractions of

1 mL for subsequent analysis by 4–12% SDS/PAGE

Synthetic peptides Peptides were synthesized as amides by solid-phase peptide synthesis as described by Atherton and Sheppard [74] The identity and purity of the peptides were ascertained by HPLC and mass spectrometry

SDS/PAGE SDS/PAGE was performed according to Laemmli [75] and Studier [76] using precast gels and following the manufac-turer’s instructions (Invitrogen) Samples from each fraction were boiled with an equal volume of sample buffer, and

10 lL was loaded onto wells of 4–12% or 4–20% Tris-gly-cine gels After running of the gels, the protein bands were stained with Coomassie Brilliant Blue (GelCODE blue stain reagent) and then with silver as described previously [77]

Immunoblotting Gels were electroblotted overnight to nitrocellulose membranes using a semidry apparatus (Bio-Rad) and a current of 200 mA for 1 h and 20 mA overnight The membrane was then washed in 50 mm Tris, pH 7.5, 0.3 m NaCl, 1% Tween 20 for 30 min All subsequent incubations and washings were in the same buffer The primary rabbit antiserum directed against the C-terminus

of MASP-2 [54] was diluted 1 : 1000 and the membrane was incubated with this for 1 h followed by three 5-min washes Next, the membrane was incubated with AP-conjugated goat immunoglobulins against rabbit immunoglobulins diluted 1 : 1000 After washing three times for 5 min, the bound antibodies were visualized by incubation in staining solution (5-Br-4-Cl-3-indolylphos-phate/nitrobluetetrazolium)

Binding assays

Binding assays were carried out in polystyrene microtitre

plates Unless otherwise stated, incubations and washings

were performed at room temperature on a shaking table by

adding 100 lL per well of TTN buffer (25 mm Tris, 0.15 m

NaCl, 0.5% Tween 20, pH 7.5) For blocking, 200 lL per

well of TTN was used Proteins (rMBL, pMBL, C1q,

cal-reticulin) were immobilized using 0.05 m sodium carbonate,

pH 9.6, as the coating buffer Control wells only received

coating buffer or an irrelevant protein (ovalbumin or BSA)

After coating overnight at 4C, plates were washed three

times for 1 min, followed by blocking for 1 h in TTN

Sub-sequently, wells were incubated with biotinylated

calreticu-lin diluted 1 : 1000 with or without other proteins/peptides

for 1 h, followed by another three washes Finally,

AP-con-jugated streptavidin diluted 1 : 1000 was added and the

wells incubated for 1 h Following another three washes,

bound calreticulin was quantified using pNPP (1 mgÆmL)1)

in 1 m diethanolamine, 0.5 mm MgCl2, pH 9.8 The

absor-bance was read at 405 nm with background subtraction at

650 nm on a VERSAmax microplate reader, using softmax

prosoftware (Molecular Devices, Sunnyvale, CA, USA) In

some experiments, calreticulin was used in combination

with specific antibodies instead of biotinylated calreticulin

and streptavidin and, in some cases, BSA was used for

blocking instead of Tween 20 As stated in the text, in some

experiments, Ca2+was added to the incubation buffer and,

in these cases, Tween 20 had to be omitted due to

precipita-tion In other experiments, calreticulin was immobilized

and wells were incubated with biotin-labelled MBL or C1q

in the presence of mannan or IgG, respectively

All binding experiments were carried out at least twice

with double determination in each experiment Data are

represented as the mean ± SD of single experiments

Acknowledgements

We thank Kirsten Beth Hansen, Dorthe Tange Olsen,

Inger Christiansen and Jette Petersen for their excellent

technical work and the Novo Nordisk Foundation for

a grant to P Højrup and a scholarship grant to

K Duus Steffen Thiel, Institute of Medical

Microbiol-ogy, University of Aarhus, Denmark is thanked for

providing recombinant proteins

References

1 Holmskov U, Thiel S & Jensenius JC (2003) Collections

and ficolins: humoral lectins of the innate immune

defense Annu Rev Immunol 21, 547–578

2 Kishore U, Greenhough TJ, Waters P, Shrive AK, Ghai

R, Kamran MF, Bernal AL, Reid KB, Madan T &

Chakraborty T (2006) Surfactant proteins SP-A and

SP-D: structure, function and receptors Mol Immunol

43, 1293–1315

3 Kojima M, Presanis JS & Sim RB (2003) The mannose-binding lectin (MBL) route for activation of comple-ment Adv Exp Med Biol 535, 229–250

4 Stuart LM, Henson PM & Vandivier RW (2006) Col-lectins: opsonins for apoptotic cells and regulators of inflammation Curr Dir Autoimmun 9, 143–161

5 Takahashi K, Ip WE, Michelow IC & Ezekowitz RA (2006) The mannose-binding lectin: a prototypic pattern recognition molecule Curr Opin Immunol 18, 16–23

6 Gadjeva M, Takahashi K & Thiel S (2004) Mannan-binding lectin – a soluble pattern recognition molecule Mol Immunol 41, 113–121

7 Drickamer K, Dordal MS & Reynolds L (1986) Man-nose-binding proteins isolated from rat liver contain carbohydrate-recognition domains linked to collagenous tails Complete primary structures and homology with pulmonary surfactant apoprotein J Biol Chem 261, 6878–6887

8 Jensen PH, Weilguny D, Matthiesen F, Mcguire KA, Shi L & Hojrup P (2005) Characterization of the oligomer structure of recombinant human mannan-binding lectin J Biol Chem 280, 11043–11051

9 Lu JH, Thiel S, Wiedemann H, Timpl R & Reid KB (1990) Binding of the pentamer/hexamer forms of man-nan-binding protein to zymosan activates the proen-zyme C1r2C1s2 complex, of the classical pathway of complement, without involvement of C1q J Immunol

144, 2287–2294

10 Sheriff S, Chang CY & Ezekowitz RA (1994) Human mannose-binding protein carbohydrate recognition domain trimerizes through a triple alpha-helical coiled-coil Nat Struct Biol 1, 789–794

11 Teillet F, Dublet B, Andrieu JP, Gaboriaud C, Arlaud

GJ & Thielens NM (2005) The two major oligomeric forms of human mannan-binding lectin: chemical char-acterization, carbohydrate-binding properties, and inter-action with MBL-associated serine proteases J Immunol

174, 2870–2877

12 Weis WI, Drickamer K & Hendrickson WA (1992) Structure of a C-type mannose-binding protein com-plexed with an oligosaccharide Nature 360, 127–134

13 Matsushita M & Fujita T (1992) Activation of the clas-sical complement pathway by mannose-binding protein

in association with a novel C1s-like serine protease

J Exp Med 176, 1497–1502

14 Sato T, Endo Y, Matsushita M & Fujita T (1994) Molecular characterization of a novel serine protease involved in activation of the complement system by mannose-binding protein Int Immunol 6, 665–669

15 Schwaeble W, Dahl MR, Thiel S, Stover C & Jensenius

JC (2002) The mannan-binding lectin-associated serine proteases (MASPs) and MAp19: four components of

the lectin pathway activation complex encoded by two

genes Immunobiology 205, 455–466

16 Sorensen R, Thiel S & Jensenius JC (2005)

Mannan-binding-lectin-associated serine proteases, characteristics

and disease associations Springer Semin Immunopathol

27, 299–319

17 Stover CM, Thiel S, Thelen M, Lynch NJ,

Vorup-Jensen T, Vorup-Jensenius JC & Schwaeble WJ (1999) Two

constituents of the initiation complex of the

mannan-binding lectin activation pathway of complement are

encoded by a single structural gene J Immunol 162,

3481–3490

18 Thiel S, Vorup-Jensen T, Stover CM, Schwaeble W,

Laursen SB, Poulsen K, Willis AC, Eggleton P, Hansen

S, Holmskov U et al (1997) A second serine protease

associated with mannan-binding lectin that activates

complement Nature 386, 506–510

19 Thiel S, Petersen SV, Vorup-Jensen T, Matsushita M,

Fujita T, Stover CM, Schwaeble WJ & Jensenius JC

(2000) Interaction of C1q and mannan-binding lectin

(MBL) with C1r C1s MBL-associated serine proteases 1

and 2, and the MBL-associated protein MAp19

J Immunol 165, 878–887

20 Thielens NM, Cseh S, Thiel S, Vorup-Jensen T, Rossi

V, Jensenius JC & Arlaud GJ (2001) Interaction

proper-ties of human mannan-binding lectin (MBL)-associated

serine proteases-1 and -2, MBL-associated protein 19,

and MBL J Immunol 166, 5068–5077

21 Vorup-Jensen T, Petersen SV, Hansen AG, Poulsen K,

Schwaeble W, Sim RB, Reid KB, Davis SJ, Thiel S &

Jensenius JC (2000) Distinct pathways of

mannan-bind-ing lectin (MBL)- and C1-complex autoactivation

revealed by reconstitution of MBL with recombinant

MBL-associated serine protease-2 J Immunol 165,

2093–2100

22 Vorup-Jensen T, Jensenius JC & Thiel S (1998)

MASP-2, the C3 convertase generating protease of the

MBLec-tin complement activaMBLec-ting pathway Immunobiology 199,

348–357

23 Chen CB & Wallis R (2001) Stoichiometry of complexes

between mannose-binding protein and its associated

ser-ine proteases Defining functional units for complement

activation J Biol Chem 276, 25894–25902

24 Arora M, Munoz E & Tenner AJ (2001) Identification

of a site on mannan-binding lectin critical for

enhancement of phagocytosis J Biol Chem 276,

43087–43094

25 Dean MM, Heatley S & Minchinton RM (2006)

Het-eroligomeric forms of codon 54 mannose binding lectin

(MBL) in circulation demonstrate reduced in vitro

func-tion Mol Immunol 43, 950–961

26 Larsen F, Madsen HO, Sim RB, Koch C & Garred P

(2004) Disease-associated mutations in human

man-nose-binding lectin compromise oligomerization and

activity of the final protein J Biol Chem 279, 21302– 21311

27 Matsushita M, Ezekowitz RA & Fujita T (1995) The Gly-54–>Asp allelic form of human mannose-binding protein (MBP) fails to bind MBP-associated serine pro-tease Biochem J 311, 1021–1023

28 Mohs A, Li Y, Doss-Pepe E, Baum J & Brodsky B (2005) Stability junction at a common mutation site in the collagenous domain of the mannose binding lectin Biochemistry 44, 1793–1799

29 Wallis R, Shaw JM, Uitdehaag J, Chen CB, Torgersen

D & Drickamer K (2004) Localization of the serine pro-tease-binding sites in the collagen-like domain of man-nose-binding protein: indirect effects of naturally occurring mutations on protease binding and activation

J Biol Chem 279, 14065–14073

30 Wallis R, Lynch NJ, Roscher S, Reid KB & Schwaeble

WJ (2005) Decoupling of carbohydrate binding and MASP-2 autoactivation in variant mannose-binding lec-tins associated with immunodeficiency J Immunol 175, 6846–6851

31 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 essen-tial role of Lys55 J Immunol 178, 5710–5716

32 Kishore U & Reid KB (2000) C1q: structure, function, and receptors Immunopharmacology 49, 159–170

33 Kishore U, Gaboriaud C, Waters P, Shrive AK, Greenhough TJ, Reid KB, Sim RB & Arlaud GJ (2004) C1q and tumor necrosis factor superfamily: modularity and versatility Trends Immunol 25, 551–561

34 Eisen DP & Minchinton RM (2003) Impact of man-nose-binding lectin on susceptibility to infectious dis-eases Clin Infect Dis 37, 1496–1505

35 Gasque P (2004) Complement: a unique innate immune sensor for danger signals Mol Immunol 41, 1089–1098

36 Ogden CA & Elkon KB (2006) Role of complement and other innate immune mechanisms in the removal of apoptotic cells Curr Dir Autoimmun 9, 120–142

37 Roos A, Xu W, Castellano G, Nauta AJ, Garred P, Daha MR & Van Kooten C (2004) Mini-review: A piv-otal role for innate immunity in the clearance of apop-totic cells Eur J Immunol 34, 921–929

38 Turner MW (2003) The role of mannose-binding lectin

in health and disease Mol Immunol 40, 423–429

39 Dong M, Xu S, Oliveira CL, Pedersen JS, Thiel S, Besenbacher F & Vorup-Jensen T (2007) Conforma-tional changes in mannan-binding lectin bound to ligand surfaces J Immunol 178, 3016–3022

40 Ghebrehiwet B & Peerschke EI (2004) cC1q-R (calreticulin) and gC1q-R/p33: ubiquitously expressed multi-ligand binding cellular proteins involved in